Compare commits
| Author | SHA1 | Date | |
|---|---|---|---|
 zjy
|
a206a43063 |
No files matched your search
@@ -4,8 +4,7 @@
|
||||
"PowerShell(uv --version)",
|
||||
"PowerShell(netstat -ano)",
|
||||
"PowerShell(Get-NetTCPConnection -LocalPort 8001 -ErrorAction SilentlyContinue)",
|
||||
"PowerShell(Get-Process -Id 20560 -ErrorAction SilentlyContinue | Select-Object Id,ProcessName,Path,StartTime; try { \\(Invoke-WebRequest -Uri \"http://127.0.0.1:8001/health\" -UseBasicParsing -TimeoutSec 5\\).Content } catch { \"health check failed: $_\" })",
|
||||
"Bash(curl -s -X POST -H \"Content-Type: application/json\" -d '{\"username\":\"demo\"}' http://localhost:8001/api/v1/auth/login/username)"
|
||||
"PowerShell(Get-Process -Id 20560 -ErrorAction SilentlyContinue | Select-Object Id,ProcessName,Path,StartTime; try { \\(Invoke-WebRequest -Uri \"http://127.0.0.1:8001/health\" -UseBasicParsing -TimeoutSec 5\\).Content } catch { \"health check failed: $_\" })"
|
||||
]
|
||||
}
|
||||
}
|
||||
+1
-13
@@ -1,7 +1,7 @@
|
||||
# Workspace/editor
|
||||
.vscode/
|
||||
frontend-web/.vscode/
|
||||
node_modules/
|
||||
|
||||
# Claude Code local-only settings (per-user permissions, not shared)
|
||||
.claude/settings.local.json
|
||||
|
||||
@@ -84,9 +84,6 @@ offline-backend-20260512/backend/tests/*
|
||||
deploy
|
||||
.history
|
||||
|
||||
# Cloned third-party repos (local only)
|
||||
git-clone/
|
||||
|
||||
# Repo-root offline dependency pack (local only)
|
||||
/dist/
|
||||
|
||||
@@ -104,12 +101,3 @@ docs/backups
|
||||
|
||||
# uv lockfile (not tracked per project preference)
|
||||
offline-backend-20260512/backend/uv.lock
|
||||
|
||||
# Local-only scratch directory (not committed)
|
||||
0622/
|
||||
|
||||
# Python bytecode cache
|
||||
__pycache__/
|
||||
|
||||
# misc
|
||||
.DS_Store
|
||||
@@ -1,103 +0,0 @@
|
||||
# AGENTS.md
|
||||
|
||||
This file provides guidance to Codex (Codex.ai/code) when working with code in this repository.
|
||||
|
||||
## Project Overview
|
||||
|
||||
This is a customized offline deployment of **DeerFlow** — a LangGraph-based AI super agent platform. The repo contains two sub-projects:
|
||||
|
||||
- `offline-backend-20260512/backend/` — Python backend (FastAPI Gateway + LangGraph agent runtime)
|
||||
- `frontend-web/` — Vite + React 19 frontend (TypeScript, pnpm)
|
||||
|
||||
Managed together via shell scripts in `scripts/`.
|
||||
|
||||
## Commands
|
||||
|
||||
### Full Application (from repo root)
|
||||
|
||||
```bash
|
||||
./scripts/start-all.sh # Start backend + frontend as background daemons
|
||||
./scripts/stop-all.sh # Stop both
|
||||
./scripts/restart-all.sh # Restart both
|
||||
./scripts/status.sh # Check running status
|
||||
```
|
||||
|
||||
Logs and PID files are written to `.runtime/logs/` and `.runtime/pids/`.
|
||||
|
||||
### Backend Only (from `offline-backend-20260512/backend/`)
|
||||
|
||||
```bash
|
||||
make install # Install Python dependencies (uv)
|
||||
make dev # Run Gateway API with hot-reload on port 8001
|
||||
make gateway # Run Gateway API without reload
|
||||
make test # Run all tests: PYTHONPATH=. uv run pytest tests/ -v
|
||||
make lint # Lint with ruff
|
||||
make format # Format with ruff
|
||||
```
|
||||
|
||||
Run a single test file:
|
||||
```bash
|
||||
PYTHONPATH=. uv run pytest tests/test_<feature>.py -v
|
||||
```
|
||||
|
||||
### Frontend Only (from `frontend-web/`)
|
||||
|
||||
```bash
|
||||
pnpm install # Install dependencies
|
||||
pnpm dev # Start Vite dev server on port 5174
|
||||
pnpm build # Production build
|
||||
pnpm typecheck # TypeScript type checking (tsc --noEmit)
|
||||
```
|
||||
|
||||
## Service Ports
|
||||
|
||||
| Service | Port | Notes |
|
||||
|---------|------|-------|
|
||||
| Backend (Gateway API) | 8001 | FastAPI + LangGraph runtime |
|
||||
| Frontend (Vite) | 5174 | React SPA |
|
||||
|
||||
Frontend connects to backend via env vars `VITE_BACKEND_BASE_URL` and `VITE_LANGGRAPH_BASE_URL` (injected by `scripts/start-frontend.sh`).
|
||||
|
||||
## Architecture
|
||||
|
||||
### Backend
|
||||
|
||||
The backend (`offline-backend-20260512/backend/`) is split into two layers:
|
||||
|
||||
- **`packages/harness/deerflow/`** — publishable agent framework (`deerflow.*` imports): LangGraph lead agent, middleware chain, sandbox, tools, MCP, memory, models, skills. **Never imports from `app.*`.**
|
||||
- **`app/`** — application layer (`app.*` imports): FastAPI Gateway routers, IM channel integrations (Feishu, Slack, Telegram, DingTalk).
|
||||
|
||||
The **Lead Agent** (`deerflow/agents/lead_agent/`) is the single LangGraph graph entry point. It wraps a middleware chain (~18 middlewares assembled in strict order) around LLM calls, handling: thread isolation, sandbox lifecycle, uploads, memory, plan mode, vision, subagent delegation, and loop detection.
|
||||
|
||||
**Middleware order matters** — middlewares are appended in a fixed sequence; `ClarificationMiddleware` must always be last. See `offline-backend-20260512/backend/AGENTS.md` for the full ordered list.
|
||||
|
||||
**Harness → App import firewall** is enforced by `tests/test_harness_boundary.py` (runs in CI).
|
||||
|
||||
### Frontend
|
||||
|
||||
The frontend (`frontend-web/`) is a React 19 + Vite SPA:
|
||||
|
||||
- **Routing**: React Router v7; top-level routes in `src/pages/PageRoutes.tsx`, workspace sub-routes in `src/pages/WorkspaceRoutes.tsx`
|
||||
- **Pages**: `ChatPage` (thread chat), `AgentChatPage` (custom agent chat), `AgentsPage`/`NewAgentPage` (agent management), `ScheduledTasksPage`/`ScheduledTaskRunDetailPage` (scheduled tasks), `LightAppManagementPage`/`LightAppListPage`/`LightAppIframeView` (轻应用中心 — register/list/embed mini-apps)
|
||||
- **AI 写作·对话版（对话驱动写作台）**: a **parallel new entry** beside the classic AI-writing page (classic page/routes untouched) — route `/page/canvas/ai-writing-chat`, sidebar 「问答管理 → 课题研究(对话版)」. Page `open-canvas/pages/AIWritingChatPage.tsx` reuses the engine (`AIWritingContext` + `useStream` + `AIWritingDraftPanel`); **idle keeps the classic setup layout** (`WritingSetupChat` + collapsible form) and switches to the chat workbench once writing starts; left panel `components/ai-writing-chat/WritingChatPanel.tsx` = **unified chronological feed** (`ChatFlowTimeline.tsx`, chat bubbles anchor-interleaved into the progress timeline so triggered events render below the user's message) + button-only intervention cards (`ChatInterventionCards.tsx`, inline textareas removed) + bottom unified smart composer. Intent resolution: frontend fast path + `POST /api/ai-writing/sessions/{id}/intent` → `deerflow/agents/ai_writing/intent_router.py` (per-pause action whitelist, high-cost confidence gates, payload cleaning); supports QA bypass, clarify follow-ups, control intents, and material-stage auto re-search + pending-intent resume. Tests `tests/test_ai_writing_intent.py`; plan doc `frontend-web/docs/ai-writing-对话驱动写作台-实现方案.md`.
|
||||
- **轻应用中心 (Light App Center)**: a sibling of 任务管理 in the sidebar with two pages — 应用管理 (admin registers `window_open` or `iframe` mini-apps with a route-param builder: login params resolved from `localStorage.userInfo`, theme params literal) and 应用列表 (card grid of `window_open` apps, `window.open` launch). `iframe` apps are mounted under a chosen parent menu (XK/PG/TY/综合管理) and **dynamically injected into the sidebar** by `core/page-layout/use-sidebar-menu.ts` (`useSidebarMenuItems`), opening in the `LightAppIframeView` host. Client + React Query hooks in `strategy-components/api/light-apps.ts`; URL builder/constants in `strategy-components/lib/light-app.ts`. Backend: `/api/light-apps` (global/shared, admin-gated writes).
|
||||
- **菜单管理 (Menu Management, admin-only)**: an **overlay** over the static sidebar menu — `MenuManagementPage` (`pages/MenuManagementPage.tsx`, route `/page/strategy/admin/menu-management`, opened from the admin-only 「菜单管理」 item in the **bottom-left settings dropdown** of `components/page-sidebar.tsx`). Admins rename / 停用·启用 / reorder (↑↓ or @dnd-kit drag) / re-parent (horizontal drag **or** the edit dialog's 「移动到」parent selector) any node across 一级/二级/三级, **including injected iframe light-app nodes** (`light-app-<appId>`). The static tree (`core/page-layout/sidebar-menu.ts`) stays the source of icons/routes; only per-node deltas (`label`/`disabled`/`parentId`/`sortOrder`) persist to the backend. Merge logic in `core/page-layout/menu-overrides.ts` (`applyMenuOverrides`, applied in `use-sidebar-menu.ts` after light-app injection, before role filter; disabling a node hides its whole subtree) — strict **3-level cap** (`wouldExceedDepth`). Client `strategy-components/api/menu-overrides.ts`. Backend: `/api/menu-overrides` (global/shared; open read, admin atomic replace-all).
|
||||
- **敏感词管理 (Sensitive Word Management)**: configures the **展示层脱敏映射** (原始词 → 替换词). **Gated on the username `lqq`** (not the admin role) — `getAccountDisplayName().toLowerCase() === "lqq"` (username = email `@` prefix, else `user_id`, same as the profile page); the dropdown entry, the route guard, and the page-level guard all use this check. `src/lib/desensitize.ts` ships a frozen seed map `SEED_DESENSITIZE_MAP` (the deployment seed) plus a live mutable `DESENSITIZE_MAP` (initialized from the seed so `desensitize()` works before any fetch). `DesensitizeWordsLoader` (mounted in `App.tsx` `AppProviders`) fetches `/api/sensitive-words` on startup and, if non-empty, calls `replaceDesensitizeMap(...)` to make the backend config the source of truth (empty/failed → keep seed). Page `SensitiveWordsPage.tsx` (route `/page/strategy/admin/sensitive-words`, opened from the 「敏感词管理」 item in the **bottom-left settings dropdown** of `components/page-sidebar.tsx`, present in **both** full + collapsed sidebar modes) lists/adds/edits/启停/deletes words and saves the **whole set atomically** (replace-all); a 「导入种子词」 button merges any missing `SEED_DESENSITIZE_MAP` entries into the editor. Client + hooks in `strategy-components/api/sensitive-words.ts` (`wordsToMap` folds enabled rows into the runtime map). Backend: `/api/sensitive-words` (global/shared; open read, write allowed for admin **or** the `lqq` account), mirroring 菜单管理's store shape. **侧边栏 + 导航栏 menu labels are desensitized** at the data source: `useSidebarMenuItems`/`useActiveThirdLevel` (`core/page-layout/use-sidebar-menu.ts`) map every node's `label` through a desensitizer built **directly from the `/api/sensitive-words` query** (`makeDesensitizer(wordsToMap(words))`, falling back to the global seed `desensitize` while empty) — this sidesteps the `replaceDesensitizeMap` effect-timing race. Only `label` is rewritten; `path`/`id`/`icon` are untouched so routing/highlight are unaffected. All nav surfaces (`page-sidebar.tsx`, `page-sidebar-dark.tsx`, `page-sidebar-v1.tsx`, `strategy-components/Header.tsx`, `PageRoutes.tsx`) consume this hook and are covered; `MenuManagementPage` uses the raw tree so admins still edit real labels.
|
||||
- **技能地址管理 (Skill Address Management, admin-only)**: 扫描运行中的技能（`skills/{public,custom}` 下所有 `.md`/`.py`）里的 **http(s) URL 与 IP/IP:端口**，按唯一地址聚合全部出现位置（跨 .md/.py），支持**批量替换**（如把 UAT 地址一次性改成生产）。主体抽成可复用组件 `SkillAddressesPanel`（`pages/SkillAddressesPage.tsx` 导出，默认导出 `SkillAddressesPage` 仍保留旧路由 `/page/strategy/admin/skill-addresses` 作直链回退），**作为「技能地址管理」标签页嵌入管理员「技能管理」页 `pages/AdminSkillsPage.tsx`**（该页加 Tabs：「技能列表」+「技能地址管理」；入口在「设置和更多」下拉的「技能管理」`workspace-nav-menu.tsx` → `/page/workspace/admin/skills`，不再单列于 `components/page-sidebar.tsx` 左下角设置下拉）：打开自动扫描并带**进度条**（扫描逐技能 / 应用逐文件 X/N，应对 150+ 技能），URL/IP 类型过滤 + 搜索 + 「片段批量套用」（如所有 `uat.4.cn` → `prod.4.cn`）、每地址「替换为」输入框（可展开看 `file:line` 片段）、内联**预览**、**应用**、**多关键词检索**（空格分隔全部命中），以及**修改记录**面板——每次应用生成一条记录（显示时间/改了哪些地址 from→to/文件数/处数），可**针对单次记录回退**（二次确认）。client + NDJSON 流式解析在 `strategy-components/api/skill-addresses.ts`。后端 `/api/skill-addresses`（**仅管理员**）：所有文件 IO 走 `asyncio.to_thread` 不阻塞事件循环；`GET /scan`(+`/scan/stream`)、`POST /preview`、`POST /apply`(+`/apply/stream`)、`GET /backups`、`POST /rollback/{id}`；**子串安全替换**（基于 token span，不会误伤子串）、`expected_total` 乐观锁（409）、写前自动磁盘备份到 `{skills_root}/.address-edits/`（同时写一份 `{id}.manifest.json` 记录本次替换明细，供修改记录展示）、应用后刷新技能系统提示缓存。纯引擎 `deerflow/skills/address_scan.py`，router `app/gateway/routers/skill_addresses.py`，tests `tests/test_skill_addresses.py`。范围仅运行中技能，不动 `deploy/minimal/skills/`。
|
||||
- **按钮管理 (Task Button Management, admin-only)**: configures the **可配置跳转按钮** rendered in the left task panel of the task deep-link workspaces (rwfx / 3qfx / xdfx, see `TaskFxWorkspace`). Each button = `{business, label, linkType, target, appendTaskId, idParam, idKind, enabled, sortOrder}`; `linkType` is `url` (外链 `window.open`, optional append of `?{idParam}={id}` — **`idParam` is the editable query-param name**, e.g. `id`/`task_id`), `business` (内部切换到另一业务, reuses `resolveTaskDeeplinkTarget`, always task id), or `purpose` (弹出当前任务的「任务目的」表格 via the shared `TaskPurposeDialog`, no target — configurable for any page incl. xdfx). **`idKind` (`task`/`action`) only matters for xdfx**: its deep-link `?taskId=` is an 行动id and the 任务id is resolved at login (`fetchTaskIdFromAction`) — `TaskContext` now also stashes the raw `actionId`, so a url button can append either (default `task`). Page `TaskButtonsPage.tsx` (route `/page/strategy/admin/task-buttons`, **admin-gated**, entry「按钮管理」in the bottom-left settings dropdown of `components/page-sidebar.tsx`, both full + collapsed modes) has per-business tabs, edits name/type/target/启停/排序/增删, saves the **whole set atomically** (replace-all), plus 「导入默认」 to merge code-seed defaults. **Seed/fallback** (like 敏感词管理): only when the whole table is empty (fresh deploy, never saved) do the workspace + page fall back to `defaultTaskButtons(cfg)` — rwfx's url buttons (编辑/下一步/3q详情) built from the **live** `/api/public/config/task-deeplink` plus cross-business switch buttons for **all three** businesses (each switches to the other two — rwfx→三情/行动, 3qfx→任务/行动, xdfx→任务/三情). Once saved, the store is the sole source of truth. The old hardcoded jump buttons were migrated out: `TaskFxWorkspace` no longer hardcodes 切换到三情分析, `TaskActionsPanel` keeps only the non-jump 清除记录, and `TaskPurposePanel` dropped its 3q详情 button (now configurable) **and defaults the 目的树 table to collapsed** (chevron toggle). Helpers `core/auth/task-buttons.ts` (`TaskButton`, `defaultTaskButtons`, `resolveTaskButtons`, `buttonsForBusiness`, `runTaskButton`), client+hooks `strategy-components/api/task-buttons.ts`. Backend: `/api/task-buttons` (global/shared; open read, admin replace-all), store `deerflow.persistence.task_buttons` (table `task_buttons`, auto-created by `create_all`), wired `app.state.task_button_store`. Tests: `tests/test_task_buttons.py`.
|
||||
- **State**: TanStack Query for server state, context providers (`SubtasksProvider`, `ArtifactsProvider`, `PromptInputProvider`) scoped to `WorkspaceRoutes`
|
||||
- **LLM streaming**: `@langchain/langgraph-sdk` streams messages from the backend via SSE
|
||||
- **UI**: Radix UI primitives + TailwindCSS v4 + shadcn-style components in `src/components/ui/`
|
||||
- **Shims**: `src/shims/` provides `next/link` and `next/navigation` polyfills (original code was Next.js)
|
||||
- **Login URL params** (`LoginPage.tsx`, frontend-only — the backend never sees these): besides the auth params (`ticket`/`userId`/`env`/`theme`/`yUserId`/`token`/`access_token`/`authToken`/`password`), the page supports a **task deep-link** via `?taskId=&goPath=`. After login it calls the consumer API `GET {VITE_AIPI_CONSUMER_URL}/taskAnalyseSearch/cop-task-detail?taskId=` (header `Authorization: Bearer admin`, auto-injected by `utils/request.ts`), stashes the full `data` (`id/overview/taskDirection/taskName/taskContent`) into `sessionStorage["login.copTaskDetail"]`, then routes by `goPath`: base codes `3qfx`/`rwfx`/`xdfx` → that business's configured **single agent** Q&A inside a **任务工作区** (resolved via `/api/business-mapping`; `resolveTaskDeeplinkTarget` calls `persistTaskContext({taskId, agentId, openingText})` and navigates to the `…/chats` **index**, which opens the latest task conversation or 新建 — a 1:N task↔对话 workspace, NOT `pendingAgentMessage`); the `-A` suffix (`3qfx-A`/…) → **multi-agent roundtable** with `roundtablePrefill.businessCode` so the page auto-selects the configured 业务链条. **xdfx's deep-link `?taskId=` is actually an 行动id, not a 任务id** — `resolveDeeplinkTaskId(goPath, rawId)` (called in `LoginPage` before everything) detects `xdfx`/`xdfx-A` and resolves it via `GET {action_detail_url}/{actionId}` → `data.taskId` (`fetchTaskIdFromAction`; `action_test`+`mock.action_detail` for offline testing); the resolved 任务id then drives `cop-task-detail`, the purpose tree, and the 1:N binding (other businesses pass the id through unchanged). **xdfx additionally partitions its chat records by 行动id (action_id), NOT by task_id**: `resolveTaskDeeplinkTarget` sets `partitionByAction:true` + keeps the raw action_id as `TaskContext.actionId`, so new threads are tagged `metadata.{taskId,actionId}` (taskId kept only to preserve the backend「任务深链对话全局共享」semantics, which keys off `metadata.taskId`) and the conversation list filters by `actionId` (+`agent_id`) — different actions of the same task get separate lists. xdfx's first-question opening text also appends `行动id（action_id）为…` (see `buildOpeningText` / `taskThreadMetadata`). goPath→businessCode: `3qfx`→`3Q`, `rwfx`→`6BF`, `xdfx`→`7BF`. The roundtable target also threads the URL `taskId` into `roundtablePrefill.taskId` (`resolveTaskDeeplinkTarget(goPath, detail, taskId)`), which turns on two taskId-scoped behaviors: (1) **separate, unpartitioned per-task storage** — taskId 会商聊天记录 live in their **own** store/table (`roundtable_task_drafts`), keyed **only by `task_id`** (one taskId = many records, shared across users, **no user 分权** at all; `created_by` is audit-only), kept entirely out of the per-user personal store (`roundtable_drafts`, now strictly per-user) so the two never mix. Backend router `/api/roundtable-task-drafts` (`GET /by-task/{id}` + `/by-task/{id}/latest`, `POST`, `GET/PUT/DELETE /{id}` — all user-agnostic); frontend client `roundtable-planning/api/task-drafts.ts`, and `useDraftPersistence`/`useRoundtableDrafts` route **all** CRUD to it when a `taskId` is present (else the personal `drafts.ts`). The matching **background jobs** are shared by task too: `roundtable_jobs.task_id` marks a task job; reads/dedup/stream/resume/cancel pass `?task_id=` so any user opening the task sees the same running 研讨 (progress/flowchart), and the executor writes the final report back to the task-draft store. Existing rows are moved over by `scripts/migrate_roundtable_task_drafts.py` (migration `20260618_01`). The history dropdown filters to the task and 新建 keeps the binding; (2) **Step-3 auto structured output** — after the `roundtable-summary` agent writes the Markdown report it automatically runs a second round emitting a `flow-json` flowchart (`useStep3Summary` `autoFlow`+`chainLevels`), whose hierarchy follows the business chain's stage goals (任务 → 目的 → 行为体 → …; coordinatorPrompt-specified hierarchy wins) and whose nodes never carry 阶段编号/阶段名 (the prior bug). **Both targets render embedded under the 任务管理 layout** — nested routes under `task-management/<business-path>/{roundtable,agent/:agent_id/chats}` in `StrategyRoutes.tsx` (`<business-path>` = `plan-task-sentiment` / `plan-task/plan-task-fx` / `plan-task/plan-task-action`, matching the COH中任务 menu paths in `sidebar-menu.ts`) so the global sidebar keeps 任务管理 → COH中任务 → the matching business highlighted. **All three single-agent businesses share `TaskFxWorkspace`** (`task-components/TaskFxWorkspace.tsx`): left = a「当前任务」card (`readCopTaskDetail()` 任务名称 + 简介, desensitized, 留存 above 新建对话) + per-task 1:N conversation list (`useTaskThreads(taskId, agentId, actionId?)` — filtered by **both** `metadata.taskId` AND `metadata.agent_id` so the three businesses' conversations staLine truncated
|
||||
- **iframe 嵌入模式 (外部系统用 iframe 套用深链问答/会商)** — 三个**按标签页**开关，全部由深链 URL 参数开启、存进同一份 `sessionStorage` 会话（`embed.session`，键名常量在 `core/embed/embed-session.ts`），跨登录跳转/刷新/页内导航存活，**绝不跨标签污染**同浏览器另开的正常系统页。核心隔离手法：用 `sessionStorage`（按标签页/浏览上下文隔离，iframe 即便同源也有独立一份）而非 `localStorage`（跨标签共享）。开关：(1) **`?embed=1`（可带 `&theme=`）** — 去全局 chrome（`PageLayout` 的 `pageSidebarAllowed && !isEmbedActive()` → 隐藏 Header + PageSidebar）+ 按外部主题色**强制**明/暗：走 next-themes 的 `forcedTheme`（`App.tsx` `AppProviders` 调 `resolveForcedEmbedTheme`，**不写** localStorage `strategy-theme`），离散映射 `dark-blue`/`dark`→深、其余→浅；嵌入侧**绝不**调 `setTheme`/`persistSidebarEnabled`（那俩落 localStorage 跨标签污染，`LoginPage` 据 `readEmbedParam` 跳过）。(2) **`?hideJump=1`** — 隐藏深链底部「快捷跳转」栏 `TaskJumpBar` 的两部分（底部「下一步」等按钮 + 右侧垂直居中悬浮菜单），单智能体问答 + 会商共用该栏，`isJumpBarHidden()` 为真时整条 `return null`。(3) **登录用户隔离** — 嵌入态下鉴权/身份键（`deerflow.auth` token、`userInfo`、`login.*`、`tools_token`、maxkey/userDetail token）全部经 `core/auth/scoped-storage.ts` 的 `authStorage()`（嵌入→sessionStorage / 正常→localStorage）读写，使 iframe 里登录的用户与正常登录用户**完全隔离**、互不覆盖；所有 token 读取都过 `getStoredAuth()`/`getAuthorizationHeaderValue()`（API 客户端 `core/api/*` 全走这俩），改中心函数即让全体调用方自动隔离；跨标签账号变更监听 `AuthSyncWatcher`（`core/auth/sync.tsx`）在嵌入态禁用。嵌入会话在 `main.tsx` 渲染前 `syncEmbedSessionFromCurrentUrl()` 同步启用（早于任何鉴权写入，避免首帧 effect 顺序漏写 localStorage）；`AppProviders` 的 effect 再按 SPA 导航同步。`embed=0`/`hideJump=0` 可显式关闭。正常（非嵌入）标签页所有 `isXxx()` 恒 false，行为**零变化**。第三个开关 `?isolate=1`（**独立登录会话**）把同一套**登录用户隔离**单独提供给**普通（带 chrome）标签页**——只切换鉴权存储、不去 chrome / 不强制主题 / 不动跳转栏：`authStorage()` 的条件从「仅嵌入」放宽为 `isEmbedActive() || isIsolatedSession()`，`AuthSyncWatcher` 的跨标签监听对嵌入态**或**独立态都即时忽略（handler 内实时判，运行时开关也即刻生效）。除 URL 参数外，**设置 → 外观 →「登录会话」** 有一个 Switch「独立登录会话（本标签页）」运行时切换（`setIsolatedSession`，`appearance-settings-page.tsx`）：**开启时先 `migrateAuthToSession()`**（`scoped-storage.ts`，把 localStorage 的受保护键 `deerflow.auth`/`userInfo`/`userId`/`tools_token` + `login.*` 前缀快照搬进 sessionStorage）再翻标志，故无感不登出；此后本标签页登录态独立——别处登录/换号影响不到它，它换号也影响不到别人，但关掉该标签即结束（需重登）。开启给 toast 提示。
|
||||
|
||||
### Configuration
|
||||
|
||||
Backend config lives in `offline-backend-20260512/backend/config.yaml`. Values starting with `$` resolve as environment variables (e.g., `$OPENAI_API_KEY`). MCP servers and skills are configured in `extensions_config.json` in the same directory.
|
||||
|
||||
## Detailed Backend Reference
|
||||
|
||||
The backend has its own comprehensive documentation at:
|
||||
- `offline-backend-20260512/backend/AGENTS.md` — architecture, middleware chain, all API routes, config schema, sandbox/memory/subagent systems
|
||||
- `offline-backend-20260512/backend/README.md` — setup guide, technology stack
|
||||
@@ -78,24 +78,11 @@ The **Lead Agent** (`deerflow/agents/lead_agent/`) is the single LangGraph graph
|
||||
The frontend (`frontend-web/`) is a React 19 + Vite SPA:
|
||||
|
||||
- **Routing**: React Router v7; top-level routes in `src/pages/PageRoutes.tsx`, workspace sub-routes in `src/pages/WorkspaceRoutes.tsx`
|
||||
- **Pages**: `ChatPage` (thread chat), `AgentChatPage` (custom agent chat), `AgentsPage`/`NewAgentPage` (agent management), `ScheduledTasksPage`/`ScheduledTaskRunDetailPage` (scheduled tasks), `LightAppManagementPage`/`LightAppListPage`/`LightAppIframeView` (轻应用中心 — register/list/embed mini-apps)
|
||||
- **定时任务「JSON → HTML 模板填充」(模板数据转换器)**: a third scheduled-task execution mode. In the create/edit dialog (`ScheduledTasksPage.tsx`), selecting the built-in agent **「模板数据转换器」** (`execution_agent_name === "template-json-builder"`, `TEMPLATE_BUILDER_AGENT_ID`) reveals a 「HTML模板」 select (`useHtmlTemplates` → `GET /api/scheduled-tasks/html-templates`) + a **.json 上传** button (reads the file's text into the prompt box) and **hides** the 必须生成 HTML/Markdown switches (this mode always outputs HTML). The chosen template persists as `execution_context.template_config.template_id` (`TemplateFillConfig` in `core/scheduled-tasks/types.ts`). At run time the backend converts the source JSON (the task prompt) into that template's `DATA` structure, validates (retry-on-fail), fills the template, and stores an HTML page previewable via the existing run-detail HTML preview. The source can be **pure JSON, plain text, or text-with-embedded-JSON** — the backend pre-extracts any embedded JSON and the agent's `SOUL.md` maps it (or the plain text's semantics) to the template's schema skeleton. **Chat-page debug entry (only this agent's `AgentChatPage`)**: when `agent_id === "template-json-builder"`, the chat header shows a 「页面模板」 select + a 「渲染预览」 button — after a normal Q&A run, the button takes the **latest AI reply**, POSTs it to `POST /api/scheduled-tasks/template-fill/render` (`renderTemplateFill`), and pops a dialog rendering the filled HTML in a `sandbox="allow-scripts"` iframe plus the structure-validation status/errors/gaps — for debugging the agent's conversion effect without creating a scheduled task. Gated strictly on the agent id, so other agents' chat pages are untouched. See the backend CLAUDE.md「Template-fill mode」for the engine (`deerflow/runtime/scheduler/template_fill.py`), agent seed, dispatch, and the render API.
|
||||
- **AI 写作·对话版（对话驱动写作台）**: a **parallel new entry** beside the classic AI-writing page (classic page/routes untouched) — route `/page/canvas/ai-writing-chat` (`OpenCanvasRoutes.tsx`), sidebar 「问答管理 → 课题研究(对话版)」. Page `open-canvas/pages/AIWritingChatPage.tsx` fully **reuses the engine** (`AIWritingContext` state machine + LangGraph `useStream`, `AIWritingHistoryProvider`, right-side `AIWritingDraftPanel`) but replaces the presentation: left = `components/ai-writing-chat/WritingChatPanel.tsx` (timeline + **button-only intervention cards** `ChatInterventionCards.tsx` — the five pause cards copied from `InterventionCard.tsx` with all inline textareas removed, confirm/action buttons kept — plus a bottom **unified smart composer**). Free text goes through two-level intent resolution (`open-canvas/api/ai-writing-intent.ts`): deterministic short-phrase fast path (确认/继续/重新搜索/重写大纲…; high-cost actions never fast-pathed) then `POST /api/ai-writing/sessions/{id}/intent` — the backend reads the graph checkpoint (authoritative pause point from the pending interrupt) and routes via `deerflow/agents/ai_writing/intent_router.py` (LLM + strict per-pause action whitelist aligned to the graph's real resume actions, confidence gates: finalize/force_finalize/覆盖式 re_search need ≥0.85 else clarify, payload key-whitelist cleaning). Results map to `submitIntervention`/control (暂停/继续/后台挂起)/QA bypass (`answer` rendered as a chat bubble, never touches graph state)/clarify (amber follow-up bubble with multi-turn context). **Two-state layout**: idle (未开始写作) renders the **classic setup experience unchanged** — `WritingSetupChat` + collapsible `AIWritingPanel` form under `WritingFormProvider` (same as `AIWritingPage`'s idle branch); once writing starts (or history view opens) it switches to the chat workbench. **Unified chronological feed** (`ChatFlowTimeline.tsx`): intent-layer chat bubbles are interleaved INTO the progress-event timeline by anchor (= `progressEvents.length` at message creation), so events triggered by a user message render **below** it (Google-AI-Studio-style single stream, not timeline-above/chat-below); composer-driven interventions suppress the duplicate 「我的要求」card (`suppressedRequirementIndices`). **素材不足自动续跑** only at `material_confirm` (the only pause the graph supports re-search from): needResearch caches the payload as a pending intent, submits `re_search + userQuery`, and auto-submits the pending intent when the new material-confirm pause arrives (cancellable pill). Stage chips + per-stage placeholder + 「当前交互对象」status badge. Backend tests `tests/test_ai_writing_intent.py`. Plan doc: `frontend-web/docs/ai-writing-对话驱动写作台-实现方案.md`.
|
||||
- **轻应用中心 (Light App Center)**: a sibling of 任务管理 in the sidebar with two pages — 应用管理 (admin registers `window_open` or `iframe` mini-apps with a route-param builder: login params resolved from `localStorage.userInfo`, theme params literal) and 应用列表 (card grid of `window_open` apps, `window.open` launch). `iframe` apps are mounted under a chosen parent menu (XK/PG/TY/综合管理) and **dynamically injected into the sidebar** by `core/page-layout/use-sidebar-menu.ts` (`useSidebarMenuItems`), opening in the `LightAppIframeView` host. Client + React Query hooks in `strategy-components/api/light-apps.ts`; URL builder/constants in `strategy-components/lib/light-app.ts`. Backend: `/api/light-apps` (global/shared, admin-gated writes).
|
||||
- **菜单管理 (Menu Management, admin-only)**: an **overlay** over the static sidebar menu — `MenuManagementPage` (`pages/MenuManagementPage.tsx`, route `/page/strategy/admin/menu-management`, opened from the admin-only 「菜单管理」 item in the **bottom-left settings dropdown** of `components/page-sidebar.tsx`). Admins rename / 停用·启用 / reorder (↑↓ or @dnd-kit drag) / re-parent (horizontal drag **or** the edit dialog's 「移动到」parent selector) any node across 一级/二级/三级, **including injected iframe light-app nodes** (`light-app-<appId>`). The static tree (`core/page-layout/sidebar-menu.ts`) stays the source of icons/routes; only per-node deltas (`label`/`disabled`/`parentId`/`sortOrder`) persist to the backend. Merge logic in `core/page-layout/menu-overrides.ts` (`applyMenuOverrides`, applied in `use-sidebar-menu.ts` after light-app injection, before role filter; disabling a node hides its whole subtree) — strict **3-level cap** (`wouldExceedDepth`). Client `strategy-components/api/menu-overrides.ts`. Backend: `/api/menu-overrides` (global/shared; open read, admin atomic replace-all).
|
||||
- **白名单管理 (Login Whitelist, admin-only)**: gates **who may log in**, run as a **blocklist (黑名单模式)**. Two pieces: a global master switch (`system_settings.json → login_whitelist.enabled`, surfaced by **admin-only** `GET|PUT /api/system-settings/login-whitelist`) + a per-user **放行** flag (`users.approved` column, migration `20260625_02`). Users **self-register** as usual and **every new account defaults to `approved=True`** — so a user who has never logged into this system before is allowed straight through; the switch only blocks users an admin has explicitly **取消放行** (`approved=False`). (`create_user`/`User`/`UserRow` all default `approved=True`; admins are always approved; existing rows are backfilled `True`.) Login interception is the single helper `enforce_user_approved(user)` in `app/gateway/routers/auth.py`, called in all three login endpoints (`/login/{local,username,token}`) **after** the account is resolved/created and **before** the JWT is issued: switch off → everyone passes; `system_role=="admin"` → always passes (anti-lockout); else `approved=False` (admin-revoked) → 403 `{code:"PENDING_APPROVAL"}` (the frontend's `extractErrorMessage` surfaces the message and login throws before writing any auth state, so no half-login). Admin approves on `pages/LoginWhitelistPage.tsx` (route `/page/strategy/admin/login-whitelist`, entry「白名单管理」in the bottom-left settings dropdown of `components/page-sidebar.tsx`, both sidebar modes): a top master-switch + a list of all registered users (`GET /api/v1/auth/users`, now carrying `approved`) with a per-row 放行 Switch (`PUT /api/v1/auth/users/{id}/approval`, admin-only; revoking only blocks the **next** login — `token_version` is left untouched, existing sessions are not force-killed). Client/hooks: switch in `strategy-components/api/login-whitelist.ts`, user list + approval in `core/auth/{api,hooks}.ts` (`useSetUserApproval`). Backend tests: `tests/test_login_approval.py`. **账号级「登录口令」(per-account password — 临时开通)**: the same page also lets an admin assign one specific account an individual **plaintext** password (`users.login_password` column, migration `20260625_04`, nullable; auto-added on startup by `_ensure_orm_columns_sync`). Once set, **that account's** username login (`POST /api/v1/auth/login/username`) **requires** `?password=` to exactly match it and that personal password **takes precedence over** the global shared-password gate (`username_login.require_password`); accounts without one keep today's shared-gate / passwordless behavior. The user then logs in via the **existing** frontend URL `#/login/<用户名>?password=xxxx` (LoginPage already forwards `?password=`; `<用户名>` = the account's email `@`-prefix). **Plaintext is intentional** — the password rides in the URL query anyway, so storing it readably lets the admin re-copy the full login link from the page; it is exposed only on the **admin-only** `GET /api/v1/auth/users` (`UserListItem.login_password`) and the new admin-only `PUT /api/v1/auth/users/{id}/login-password` (set non-empty / clear with null·empty; doesn't touch `token_version`). The whitelist page renders a per-user 「登录口令」 column with a set/修改 editor + a live 「复制登录链接」 button (`useSetUserLoginPassword` in `core/auth/hooks.ts`, `setUserLoginPassword` in `core/auth/api.ts`). Backend gate helper `_lookup_user_by_username` (lookup-only, no auto-register) in `routers/auth.py`; tests `tests/test_login_personal_password.py`.
|
||||
- **敏感词管理 (Sensitive Word Management)**: configures the **展示层脱敏映射** (原始词 → 替换词). **Gated on the username `lqq`** (not the admin role) — `getAccountDisplayName().toLowerCase() === "lqq"` (username = email `@` prefix, else `user_id`, same as the profile page); the dropdown entry, the route guard, and the page-level guard all use this check. `src/lib/desensitize.ts` ships a frozen seed map `SEED_DESENSITIZE_MAP` (the deployment seed) plus a live mutable `DESENSITIZE_MAP` (initialized from the seed so `desensitize()` works before any fetch). `DesensitizeWordsLoader` (mounted in `App.tsx` `AppProviders`) fetches `/api/sensitive-words` on startup and, if non-empty, calls `replaceDesensitizeMap(...)` to make the backend config the source of truth (empty/failed → keep seed). Page `SensitiveWordsPage.tsx` (route `/page/strategy/admin/sensitive-words`, opened from the 「敏感词管理」 item in the **bottom-left settings dropdown** of `components/page-sidebar.tsx`, present in **both** full + collapsed sidebar modes) lists/adds/edits/启停/deletes words and saves the **whole set atomically** (replace-all); a 「导入种子词」 button merges any missing `SEED_DESENSITIZE_MAP` entries into the editor. Client + hooks in `strategy-components/api/sensitive-words.ts` (`wordsToMap` folds enabled rows into the runtime map). Backend: `/api/sensitive-words` (global/shared; open read, write allowed for admin **or** the `lqq` account), mirroring 菜单管理's store shape. **侧边栏 + 导航栏 menu labels are desensitized** at the data source: `useSidebarMenuItems`/`useActiveThirdLevel` (`core/page-layout/use-sidebar-menu.ts`) map every node's `label` through a desensitizer built **directly from the `/api/sensitive-words` query** (`makeDesensitizer(wordsToMap(words))`, falling back to the global seed `desensitize` while empty) — this sidesteps the `replaceDesensitizeMap` effect-timing race. Only `label` is rewritten; `path`/`id`/`icon` are untouched so routing/highlight are unaffected. All nav surfaces (`page-sidebar.tsx`, `page-sidebar-dark.tsx`, `page-sidebar-v1.tsx`, `strategy-components/Header.tsx`, `PageRoutes.tsx`) consume this hook and are covered; `MenuManagementPage` uses the raw tree so admins still edit real labels.
|
||||
- **技能地址管理 (Skill Address Management, admin-only)**: 扫描运行中的技能（`skills/{public,custom}` 下所有 `.md`/`.py`）里的 **http(s) URL 与 IP/IP:端口**，按唯一地址聚合全部出现位置（跨 .md/.py），支持**批量替换**（如把 UAT 地址一次性改成生产）。主体抽成可复用组件 `SkillAddressesPanel`（`pages/SkillAddressesPage.tsx` 导出，默认导出 `SkillAddressesPage` 仍保留旧路由 `/page/strategy/admin/skill-addresses` 作直链回退），**作为「技能地址管理」标签页嵌入管理员「技能管理」页 `pages/AdminSkillsPage.tsx`**（该页加 Tabs：「技能列表」+「技能地址管理」；入口在「设置和更多」下拉的「技能管理」`workspace-nav-menu.tsx` → `/page/workspace/admin/skills`，不再单列于 `components/page-sidebar.tsx` 左下角设置下拉）：打开自动扫描并带**进度条**（扫描逐技能 / 应用逐文件 X/N，应对 150+ 技能），URL/IP 类型过滤 + 搜索 + 「片段批量套用」（如所有 `uat.4.cn` → `prod.4.cn`）、每地址「替换为」输入框（可展开看 `file:line` 片段）、内联**预览**、**应用**、**多关键词检索**（空格分隔全部命中），以及**修改记录**面板——每次应用生成一条记录（显示时间/改了哪些地址 from→to/文件数/处数），可**针对单次记录回退**（二次确认）。client + NDJSON 流式解析在 `strategy-components/api/skill-addresses.ts`。后端 `/api/skill-addresses`（**仅管理员**）：所有文件 IO 走 `asyncio.to_thread` 不阻塞事件循环；`GET /scan`(+`/scan/stream`)、`POST /preview`、`POST /apply`(+`/apply/stream`)、`GET /backups`、`POST /rollback/{id}`；**子串安全替换**（基于 token span，不会误伤子串）、`expected_total` 乐观锁（409）、写前自动磁盘备份到 `{skills_root}/.address-edits/`（同时写一份 `{id}.manifest.json` 记录本次替换明细，供修改记录展示）、应用后刷新技能系统提示缓存。纯引擎 `deerflow/skills/address_scan.py`，router `app/gateway/routers/skill_addresses.py`，tests `tests/test_skill_addresses.py`。范围仅运行中技能，不动 `deploy/minimal/skills/`。
|
||||
- **按钮管理 (Task Button Management, admin-only)**: configures the **可配置跳转按钮** rendered in the left task panel of the task deep-link workspaces (rwfx / 3qfx / xdfx, see `TaskFxWorkspace`). Each button = `{business, label, linkType, target, appendTaskId, idParam, idKind, openMode, loginParams, appendAuth, authTokenParam, authNameParam, enabled, sortOrder}`; `linkType` is `url` (外链 `window.open`, optional append of `?{idParam}={id}` — **`idParam` is the editable query-param name**, e.g. `id`/`task_id`), `business` (内部切换到另一业务, reuses `resolveTaskDeeplinkTarget`, always task id), or `purpose` (弹出当前任务的「任务目的」表格 via the shared `TaskPurposeDialog`, no target — configurable for any page incl. xdfx). **`url` buttons support two independent, coexisting login-append mechanisms**: (1) **`loginParams` (追加登录参数, like 应用管理/Light App Center)** — a list of `{key, source, customValue}` resolved at click time: `runTaskButton` reads `userInfo` (via `readUserInfo` from `strategy-components/lib/light-app`, embed-isolated through `authStorage`) and appends `?{key}={value}` for each (`source` = a userInfo field such as `accessToken`/`yUserId`, or `other` → literal `customValue`); empty list = nothing appended. Reuses `LOGIN_SOURCE_OPTIONS` from light-app; stored as JSON column `task_buttons.login_params_json` (`PortableJSON`, migration `20260625_01`). (2) **`appendAuth` (追加登录态)** — appends `?{authTokenParam}={token1}&{authNameParam}={username}` (default param names `token`/`username`) so the跳转目标外部系统 can identify the current user — e.g. `…/action/sentiment?token=<JWT>&username=lizhaoxing`. **Gated on token login**: `token1` = the `?authToken=` used by token-exchange login; `username` = the raw login name the backend resolved from `getTokenInfo`. Both are stashed at login by `persistLoginCredentials` into `login.token1`/`login.username` (`core/auth/login-credentials.ts`, scoped via [[scoped-storage]] so嵌入/独立会话隔离 + 随独立会话迁移). Only token login writes them → a non-token-login session has no `token1`, so `runTaskButton` silently skips the auth append (the jump still works, just without those two params). The frontend reads `username` straight from the new `UsernameLoginResponse.username` field (the stored email local-part is lossy — lowercased/normalized — so not used), **not** by decoding the JWT. Stored as columns `task_buttons.{append_auth, auth_token_param, auth_name_param}` (migration `20260625_02`). The admin page renders both editors for `url` buttons (loginParams 虚线框 + 追加登录态 switch). **`idKind` (`task`/`action`) only matters for xdfx**: its deep-link `?taskId=` is an 行动id and the 任务id is resolved at login (`fetchTaskIdFromAction`) — `TaskContext` now also stashes the raw `actionId`, so a url button can append either (default `task`). Page `TaskButtonsPage.tsx` (route `/page/strategy/admin/task-buttons`, **admin-gated**, entry「按钮管理」in the bottom-left settings dropdown of `components/page-sidebar.tsx`, both full + collapsed modes) has per-business tabs, edits name/type/target/启停/排序/增删, saves the **whole set atomically** (replace-all), plus 「导入默认」 to merge code-seed defaults. **Seed/fallback** (like 敏感词管理): only when the whole table is empty (fresh deploy, never saved) do the workspace + page fall back to `defaultTaskButtons(cfg)` — rwfx's url buttons (编辑/上一步/下一步/3q详情) built from the **live** `/api/public/config/task-deeplink` plus cross-business switch buttons for **all three** businesses (each switches to the other two — rwfx→三情/行动, 3qfx→任务/行动, xdfx→任务/三情). Once saved, the store is the sole source of truth. **Bottom 快捷跳转 bar groups by prev/next**: `TaskJumpBar` renders only「上一步/下一步」-type buttons in the bottom bar — a `prev` group (id 含 `prev` 或文案「上一步」, 描边次要样式 + ←图标) rendered **左侧**, then a `next` group (id 含 `next` 或文案「下一步」, 实心主色) on the **右侧** (container `justify-end`); everything else goes to the right-edge floating menu. The default 上一步 button (`rwfx-prev`, `url`, target = `DASHBOARD_AGENT_ROUTE` `/page/strategy/dashboard-agent`, `appendTaskId` + `idParam=taskId` → `?taskId=任务id`, `openMode=self`) 默认指向**应用内**「大屏绘制」页，在 **both** 单智能体 (`TaskFxWorkspace`) 与 多智能体会商 (`RoundtablePlanningPage` 深链 rwfx-A) 都显示（共用 `TaskJumpBar`）。**`runTaskButton` 的 `url` 类型支持应用内路由**：target 以单个 `/` 开头时按内部路由处理——`self` 走 react-router 同窗导航（不刷新），`blank` 拼成当前源的 hash 路由新标签打开；外链行为不变。 The old hardcoded jump buttons were migrated out: `TaskFxWorkspace` no longer hardcodes 切换到三Line truncated
|
||||
- **实时并发监控 (Concurrency / System-Pressure Monitor, admin-only)**: a live panel mounted at the **top of the admin 排行榜 page** (`components/workspace/admin/concurrency-monitor.tsx` `ConcurrencyMonitorPanel`, rendered in `pages/AdminUserLeaderboardPage.tsx`) showing **当前并发的大模型调用数 + 列表（用户/智能体/类型/已运行时长）**, each row with a 「停止」button (and a 「全部停止」) that **truly aborts** that run's in-flight model stream via `RunManager.cancel` (`POST /api/admin/active-runs/{run_id}/cancel`) — to relieve pressure when 使用人数过多. Two line charts (minute/hour toggle): **并发量折线图** (peak/avg, `GET /api/admin/active-runs/concurrency`, fed by a lightweight background sampler into `concurrency_samples`) and **大模型出字速度折线图** (tokens/sec, `GET /api/admin/active-runs/token-speed`, aggregated from `llm_call_metrics`) so a dip pinpoints where the bottleneck is. A **「后台采样」Switch** (`system_settings.json → concurrency_monitor.enabled`, **default off / opt-in**, `GET|PUT /api/admin/active-runs/settings`) turns the background sampling on/off **at runtime, no restart** (the live concurrency count + 「停止」 work even while sampling is off — only the 并发量 history chart needs it on); everything is designed to **never impact production** (one tiny INSERT/15s, retention-purged, best-effort). Client/hooks: `core/admin/active-runs.ts` + `core/admin/hooks.ts` (`useActiveRuns`/`useConcurrencySeries`/`useTokenSpeedSeries`/`useCancel*`/`useMonitorSettings`). See the backend CLAUDE.md「Admin Concurrency Monitor」.
|
||||
- **会话冲突 409 中文化**: when a user submits a new question while the previous turn is still streaming, the backend's `RunManager.create_or_reject` rejects with HTTP 409 (`services.py` now returns a **Chinese** detail「当前对话还在生成回答…」). `core/threads/hooks.ts` `onError` detects the conflict (`isActiveRunConflictError`: 409 / "already has an active run" / the Chinese text) and shows a friendly `toast.info` instead of a raw error.
|
||||
- **Pages**: `ChatPage` (thread chat), `AgentChatPage` (custom agent chat), `AgentsPage`/`NewAgentPage` (agent management), `ScheduledTasksPage`/`ScheduledTaskRunDetailPage` (scheduled tasks)
|
||||
- **State**: TanStack Query for server state, context providers (`SubtasksProvider`, `ArtifactsProvider`, `PromptInputProvider`) scoped to `WorkspaceRoutes`
|
||||
- **LLM streaming**: `@langchain/langgraph-sdk` streams messages from the backend via SSE
|
||||
- **UI**: Radix UI primitives + TailwindCSS v4 + shadcn-style components in `src/components/ui/`
|
||||
- **Shims**: `src/shims/` provides `next/link` and `next/navigation` polyfills (original code was Next.js)
|
||||
- **Login URL params** (`LoginPage.tsx`, frontend-only — the backend never sees these): besides the auth params (`ticket`/`userId`/`env`/`theme`/`yUserId`/`token`/`access_token`/`authToken`/`password`), the page supports a **task deep-link** via `?taskId=&goPath=`. After login it calls the consumer API `GET {VITE_AIPI_CONSUMER_URL}/taskAnalyseSearch/cop-task-detail?taskId=` (header `Authorization: Bearer admin`, auto-injected by `utils/request.ts`), stashes the full `data` (`id/overview/taskDirection/taskName/taskContent`) into `sessionStorage["login.copTaskDetail"]`, then routes by `goPath`: base codes `3qfx`/`rwfx`/`xdfx` → that business's configured **single agent** Q&A inside a **任务工作区** (resolved via `/api/business-mapping`; `resolveTaskDeeplinkTarget` calls `persistTaskContext({taskId, agentId, openingText})` and navigates to the `…/chats` **index**, which opens the latest task conversation or 新建 — a 1:N task↔对话 workspace, NOT `pendingAgentMessage`); the `-A` suffix (`3qfx-A`/…) → **multi-agent roundtable** with `roundtablePrefill.businessCode` so the page auto-selects the configured 业务链条. **xdfx's deep-link `?taskId=` is actually an 行动id, not a 任务id** — `resolveDeeplinkTaskId(goPath, rawId)` (called in `LoginPage` before everything) detects `xdfx`/`xdfx-A` and resolves it via `GET {action_detail_url}/{actionId}` → `data.taskId` (`fetchTaskIdFromAction`; `action_test`+`mock.action_detail` for offline testing); the resolved 任务id then drives `cop-task-detail`, the purpose tree, and the 1:N binding (other businesses pass the id through unchanged). **xdfx additionally partitions its chat records by 行动id (action_id), NOT by task_id**: `resolveTaskDeeplinkTarget` sets `partitionByAction:true` + keeps the raw action_id as `TaskContext.actionId`, so new threads are tagged `metadata.{taskId,actionId}` (taskId kept only to preserve the backend「任务深链对话全局共享」semantics, which keys off `metadata.taskId`) and the conversation list filters by `actionId` (+`agent_id`) — different actions of the same task get separate lists. xdfx's first-question opening text also appends `行动id（action_id）为…` (see `buildOpeningText` / `taskThreadMetadata`). goPath→businessCode: `3qfx`→`3Q`, `rwfx`→`6BF`, `xdfx`→`7BF`. The roundtable target also threads the URL `taskId` into `roundtablePrefill.taskId` (`resolveTaskDeeplinkTarget(goPath, detail, taskId)`), which turns on two taskId-scoped behaviors: (1) **separate, unpartitioned per-task storage** — taskId 会商聊天记录 live in their **own** store/table (`roundtable_task_drafts`), keyed **only by `task_id`** (one taskId = many records, shared across users, **no user 分权** at all; `created_by` is audit-only), kept entirely out of the per-user personal store (`roundtable_drafts`, now strictly per-user) so the two never mix. Backend router `/api/roundtable-task-drafts` (`GET /by-task/{id}` + `/by-task/{id}/latest`, `POST`, `GET/PUT/DELETE /{id}` — all user-agnostic); frontend client `roundtable-planning/api/task-drafts.ts`, and `useDraftPersistence`/`useRoundtableDrafts` route **all** CRUD to it when a `taskId` is present (else the personal `drafts.ts`). The matching **background jobs** are shared by task too: `roundtable_jobs.task_id` marks a task job; reads/dedup/stream/resume/cancel pass `?task_id=` so any user opening the task sees the same running 研讨 (progress/flowchart), and the executor writes the final report back to the task-draft store. Existing rows are moved over by `scripts/migrate_roundtable_task_drafts.py` (migration `20260618_01`). The history dropdown filters to the task and 新建 keeps the binding; (2) **Step-3 auto structured output** — after the `roundtable-summary` agent writes the Markdown report it automatically runs a second round emitting a `flow-json` flowchart (`useStep3Summary` `autoFlow`+`chainLevels`), whose hierarchy follows the business chain's stage goals (任务 → 目的 → 行为体 → …; coordinatorPrompt-specified hierarchy wins) and whose nodes never carry 阶段编号/阶段名 (the prior bug). **Both targets render embedded under the 任务管理 layout** — nested routes under `task-management/<business-path>/{roundtable,agent/:agent_id/chats}` in `StrategyRoutes.tsx` (`<business-path>` = `plan-task-sentiment` / `plan-task/plan-task-fx` / `plan-task/plan-task-action`, matching the COH中任务 menu paths in `sidebar-menu.ts`) so the global sidebar keeps 任务管理 → COH中任务 → the matching business highlighted. **All three single-agent businesses share `TaskFxWorkspace`** (`task-components/TaskFxWorkspace.tsx`): left = a「当前任务」card (`readCopTaskDetail()` 任务名称 + 简介, desensitized, 留存 above 新建对话) + per-task 1:N conversation list (`useTaskThreads(taskId, agentId, actionId?)` — filtered by **both** `metadata.taskId` AND `metadata.agent_id` so the three businesses' conversations staLine truncated
|
||||
- **iframe 嵌入模式 (外部系统用 iframe 套用深链问答/会商)** — 三个**按标签页**开关，全部由深链 URL 参数开启、存进同一份 `sessionStorage` 会话（`embed.session`，键名常量在 `core/embed/embed-session.ts`），跨登录跳转/刷新/页内导航存活，**绝不跨标签污染**同浏览器另开的正常系统页。核心隔离手法：用 `sessionStorage`（按标签页/浏览上下文隔离，iframe 即便同源也有独立一份）而非 `localStorage`（跨标签共享）。开关：(1) **`?embed=1`（可带 `&theme=`）** — 去全局 chrome（`PageLayout` 的 `pageSidebarAllowed && !isEmbedActive()` → 隐藏 Header + PageSidebar）+ 按外部主题色**强制**明/暗：走 next-themes 的 `forcedTheme`（`App.tsx` `AppProviders` 调 `resolveForcedEmbedTheme`，**不写** localStorage `strategy-theme`），离散映射 `dark-blue`/`dark`→深、其余→浅；嵌入侧**绝不**调 `setTheme`/`persistSidebarEnabled`（那俩落 localStorage 跨标签污染，`LoginPage` 据 `readEmbedParam` 跳过）。(2) **`?hideJump=1`** — 隐藏深链底部「快捷跳转」栏 `TaskJumpBar` 的两部分（底部「下一步」等按钮 + 右侧垂直居中悬浮菜单），单智能体问答 + 会商共用该栏，`isJumpBarHidden()` 为真时整条 `return null`。(2.5) **`?hideChrome=1`** — **仅**去全局 chrome（Header 导航栏 + PageSidebar 侧边栏），**不**强制主题、**不**隔离鉴权（区别于 `embed`：只想要一个干净全屏页面，如「大屏绘制」`/page/strategy/dashboard-agent?taskId=…&hideChrome=1`）。新增 `EmbedSession.hideChrome` 字段 + `readHideChromeParam()` + `isChromeHidden()`（取 `embed || hideChrome` 之并），与 `hideJump` 完全对称；`PageRoutes.tsx` 的 `PageLayout` 把 `chromeHidden`（`isChromeHidden() || readHideChromeParam(location.search)==="on"`）并入 `pageSidebarAllowed`，且其**优先级高于** dashboard-agent 路由原本的强制全局 chrome（`forceChromeForDashboardAgent`）。`hideChrome=0` 可显式关闭。(3) **登录用户隔离** — 嵌入态下鉴权/身份键（`deerflow.auth` token、`userInfo`、`login.*`、`tools_token`、maxkey/userDetail token）全部经 `core/auth/scoped-storage.ts` 的 `authStorage()`（嵌入→sessionStorage / 正常→localStorage）读写，使 iframe 里登录的用户与正常登录用户**完全隔离**、互不覆盖；所有 token 读取都过 `getStoredAuth()`/`getAuthorizationHeaderValue()`（API 客户端 `core/api/*` 全走这俩），改中心函数即让全体调用方自动隔离；跨标签账号变更监听 `AuthSyncWatcher`（`core/auth/sync.tsx`）在嵌入态禁用。嵌入会话在 `main.tsx` 渲染前 `syncEmbedSessionFromCurrentUrl()` 同步启用（早于任何鉴权写入，避免首帧 effect 顺序漏写 localStorage）；`AppProviders` 的 effect 再按 SPA 导航同步。`embed=0`/`hideJump=0` 可显式关闭。正常（非嵌入）标签页所有 `isXxx()` 恒 false，行为**零变化**。第三个开关 `?isolate=1`（**独立登录会话**）把同一套**登录用户隔离**单独提供给**普通（带 chrome）标签页**——只切换鉴权存储、不去 chrome / 不强制主题 / 不动跳转栏：`authStorage()` 的条件从「仅嵌入」放宽为 `isEmbedActive() || isIsolatedSession()`，`AuthSyncWatcher` 的跨标签监听对嵌入态**或**独立态都即时忽略（handler 内实时判，运行时开关也即刻生效）。除 URL 参数外，**设置 → 外观 →「登录会话」** 有一个 Switch「独立登录会话（本标签页）」运行时切换（`setIsolatedSession`，`appearance-settings-page.tsx`）：**开启时先 `migrateAuthToSession()`**（`scoped-storage.ts`，把 localStorage 的受保护键 `deerflow.auth`/`userInfo`/`userId`/`tools_token` + `login.*` 前缀快照搬进 sessionStorage）再翻标志，故无感不登出；此后本标签页登录态独立——别处登录/换号影响不到它，它换号也影响不到别人，但关掉该标签即结束（需重登）。开启给 toast 提示。
|
||||
- **操作日志（Y-Log）上报** — 用户的关键操作（登录登出/查询/查看/增删改/启停/导入导出/问答等）成功后，**前端直连**外部接口 `POST {write_url}`（默认 `https://ch1.b.uat.4.cn/api/y-log/write`）记一条审计日志。核心模块 `src/core/log/`：`y-log.ts`（`writeYLog(params)` + `OptType` 枚举 0/1/3..12；**异步即发即忘、永不 await、永不抛错、失败静默**，自动把完整 url 转相对路径去 ip/host、对象入参/返回自动 `JSON.stringify` + 50k 裁剪、`optResult` 恒 0 只报成功），`y-log-config.ts`（从公开只读 `GET /api/public/config/y-log` 拉 `{enabled, writeUrl}`，模块级缓存 + 去重 + 默认兜底；`App.tsx` `AppProviders` 启动预热 `fetchYLogConfig()`），`y-log-token.ts`（token1 = token 登录的 `?authToken=`，经 [[scoped-storage]] 的 `authStorage*` 存取，嵌入态自动隔离）。**双门控**：仅当①存在 token1（即用 token 登录，`LoginPage` 成功后 `persistYLogToken(authToken)`，登出 `account-settings-page.tsx` 先报后清 token1）且②配置 `enabled && writeUrl` 才上报；否则一律不发。`optDetail` 统一「CM助手-全量-…」（**禁内部框架名**）。埋点**广覆盖**各 API client 的成功路径——原则是「一个用户操作（查看/查询/新增/编辑/删除/启停/导入导出）= 一条日志」，每个 API client 的**单条详情 GET（查看→view 4）/ 列表查询 GET（search 3）/ 新增 POST（create 9）/ 改 PUT·PATCH（edit 5）/ 删 DELETE（delete 6）/ 启停（enabled 11 / disabled 10）/ 导入导出（import 7 / export 8）**都各落一条，仅跳过高频 typeahead/状态轮询/纯 bootstrap config 读。覆盖文件：`strategy-components/api/{sensitive-words,menu-overrides,task-buttons,skill-addresses,light-apps,business-mapping}.ts`、`roundtable-planning/api/{drafts,task-drafts,roundtable-jobs,recommend,chains,multi-agent,dashboard-sessions}.ts`、`core/{agents,scheduled-tasks,knowledge,positions,tags,studio,system-settings,memory,mcp,models,custom-prompts,user-preferences,notifications,recommended-questions,user-qa,curator,admin,thread-shares,roundtable-draft-shares,tool-metrics,llm-metrics,note,artifacts}/api.ts`、`core/api/markdown-docx-export.ts`。**问答（Q&A）改为「一次大模型调用 = 一条日志」**：在共享的 `core/threads/hooks.ts` `useThreadStream` 的 `onLangChainEvent` 里监听 `on_chat_model_end`，每次模型调用结束就上报一条（optType 12，`reqParam` 取本次 input 的最后一条 user 文本 + 模型名，`returnParam` 取本次输出文本）——一个问答轮次里 lead agent / 子代理的**每次** LLM 调用都各记一条，覆盖所有走 `useThreadStream` 的页面（`ChatPage`/`AgentChatPage`/创建智能体/iframe/会商等）；`ChatPage`/`AgentChatPage` 的 `onFinish` 仍额外保留一条轮次汇总。后端仅新增公开只读配置：`deerflow/config/y_log_config.py`（`YLogConfig{enabled,write_url}`）挂到 `AppConfig.y_log`，`app/gateway/routers/public_config.py` 加 `GET /api/public/config/y-log`，地址写在 `config.yaml` 的 `y_log` 段（mtime 热重载，前端无需重建）；**不做后端代理**。Tests：`tests/test_y_log_config.py`。
|
||||
|
||||
### Configuration
|
||||
|
||||
|
||||
Submodule _skills_src/ai-skills deleted from cde1c8385e.
Submodule _skills_src/anthropics-skills deleted from 5754626092.
Submodule _skills_src/awesome-claude-skills deleted from 92568c1eda.
Submodule _skills_src/ce-kit deleted from d4973c4827.
Submodule _skills_src/csm deleted from 3fa16a94e0.
Submodule _skills_src/superpowers deleted from 6fd4507659.
Submodule _skills_src/tapestry-skills deleted from 80e1dc56df.
@@ -1,33 +0,0 @@
|
||||
# 批量人物蒸馏 → llmwiki
|
||||
|
||||
一个**完全独立的单文件脚本 + 内置定时器**：读 Excel 名单，串行调用线上 DeerFlow 服务问答接口，
|
||||
复用「女娲 / huashu-nuwa」技能逐个蒸馏人物，**由脚本抓取回复并保存**成 llmwiki 风格 markdown。
|
||||
|
||||
- **不联网**：只用已配置的知识库查询技能（`knowledge-search-v2`）检索资料。
|
||||
- **定时执行**：每天 **23:00 ~ 次日 08:00** 自动跑，到点暂停、次日继续。
|
||||
- **md 由 py 保存**：agent 只产出内容，脚本抓最终回复写成 `.md`（不让项目自己写文件）。
|
||||
|
||||
## 你只需要做两件事
|
||||
|
||||
1. **改一行地址**：打开 `distill_people.py`，把顶部 `BASE_URL = "..."` 改成你的线上服务地址。
|
||||
2. **放名单**：同目录放 `people.xlsx`（或 `people.csv`），**第一列 = 人名**，其余列=可选提示。
|
||||
|
||||
其余（用户名 `distiller`、口令 `123ewq`、技能名、时间窗）都已内置，不用填。
|
||||
|
||||
## 运行
|
||||
|
||||
```bash
|
||||
python distill_people.py # 常驻，按 23:00~08:00 窗口自动执行（推荐挂后台）
|
||||
python distill_people.py --now # 忽略时间窗，立刻把名单跑完（测试用）
|
||||
python distill_people.py --redo # 忽略已完成，全部重做
|
||||
```
|
||||
|
||||
读 `.xlsx` 需要 `openpyxl`；用 `.csv` 则零额外依赖。
|
||||
|
||||
## 输出（都在 output/ 一个文件夹）
|
||||
|
||||
- `<人名>.md` —— 每个人的 llmwiki 档案
|
||||
- `manifest.json` —— 执行清单（状态/耗时/run_id）
|
||||
- `index.md` —— 可读清单表（谁完成/失败，点开看 md）
|
||||
|
||||
**断点续跑**：已完成的人自动跳过；中途往名单里加人，下个执行窗会带上。
|
||||
@@ -1,563 +0,0 @@
|
||||
#!/usr/bin/env python3
|
||||
# -*- coding: utf-8 -*-
|
||||
"""
|
||||
批量人物蒸馏 → llmwiki markdown（完全独立的单文件客户端 + 内置定时器）
|
||||
|
||||
特点
|
||||
----
|
||||
- 串行（一个一个来）调用正在运行的 DeerFlow 服务问答接口
|
||||
- 复用线上「女娲 / huashu-nuwa」技能做提炼；**不联网**，只用已配置的
|
||||
知识库查询技能（knowledge-search-v2）检索资料
|
||||
- agent 只负责产出内容，**markdown 由本脚本抓取最终回复后自己保存**
|
||||
- 内置时间窗：每天 23:00 ~ 次日 08:00 才执行，到点自动暂停、次日继续
|
||||
- 断点续跑：已完成的人自动跳过；中途加名单也会被带上
|
||||
|
||||
你只需要改一处：下面的 BASE_URL（线上服务地址）。其余都不用填。
|
||||
|
||||
名单
|
||||
----
|
||||
脚本同目录放 people.xlsx 或 people.csv，**第一列 = 人名**，其余列=可选提示。
|
||||
|
||||
运行
|
||||
----
|
||||
python distill_people.py # 常驻，按 23:00~08:00 窗口自动执行
|
||||
python distill_people.py --now # 忽略时间窗，立刻把名单跑完（测试用）
|
||||
python distill_people.py --redo # 忽略已完成，全部重做
|
||||
"""
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import csv
|
||||
import json
|
||||
import re
|
||||
import sys
|
||||
import time
|
||||
from datetime import datetime, time as dtime
|
||||
from pathlib import Path
|
||||
|
||||
try:
|
||||
import requests
|
||||
except ImportError:
|
||||
sys.exit("缺少依赖 requests：pip install requests")
|
||||
|
||||
|
||||
# ════════════════════════════════════════════════════════════════════════
|
||||
# 只改这一行：你的线上服务地址（不带结尾斜杠）
|
||||
# ════════════════════════════════════════════════════════════════════════
|
||||
BASE_URL = "http://47.94.209.59:2026/"
|
||||
# ════════════════════════════════════════════════════════════════════════
|
||||
|
||||
|
||||
# ---- 以下都已内置，正常无需改动 ----
|
||||
USERNAME = "distiller" # 免密自动注册
|
||||
PASSWORD = "123ewq" # 共享口令（你们开了口令门）
|
||||
NUWA_SKILL = "huashu-nuwa" # 女娲技能
|
||||
KB_SKILL = "knowledge-search-v2" # 知识库查询技能（不联网，用它检索）
|
||||
DEPTH = "full" # full 完整 / fast 精简
|
||||
|
||||
HERE = Path(__file__).resolve().parent
|
||||
OUTPUT_DIR = HERE / "output"
|
||||
PEOPLE_FILE_DEFAULT = HERE / "people.example.csv"
|
||||
|
||||
# 执行时间窗：23:00 ~ 次日 08:00（跨午夜）
|
||||
WIN_START = dtime(23, 0)
|
||||
WIN_END = dtime(8, 0)
|
||||
|
||||
PER_PERSON_TIMEOUT = 1800 # 单人最长等待（秒）
|
||||
POLL_INTERVAL = 5 # 轮询间隔（秒）
|
||||
HTTP_TIMEOUT = 60
|
||||
TERMINAL_STATUS = {"success", "error", "timeout", "interrupted"}
|
||||
|
||||
|
||||
# ================================ llmwiki 模板 ================================
|
||||
LLMWIKI_TEMPLATE = """\
|
||||
---
|
||||
name: <中文名/常用名>
|
||||
aliases: [<别名/英文名...>]
|
||||
type: person
|
||||
era: <在世 | 历史人物>
|
||||
domains: [<领域1>, <领域2>]
|
||||
distilled_at: <YYYY-MM-DD>
|
||||
sources_count: <检索到的来源条数>
|
||||
primary_source_ratio: <一手来源占比, 0~1 的小数>
|
||||
confidence: <high | medium | low>
|
||||
---
|
||||
|
||||
# <人物名>
|
||||
|
||||
> **一句话定位**：<此人看世界的独特方式，一句话>
|
||||
|
||||
## 身份卡
|
||||
- **活跃年代**：
|
||||
- **核心身份**：
|
||||
- **自述(50字, 以其语气)**：
|
||||
|
||||
## 核心观点（真信念 · 反复出现的主张）
|
||||
1. ...
|
||||
|
||||
## 心智模型（3–7 个；每个含：一句话 / 证据场景≥2 / 应用 / 局限 / 来源）
|
||||
### 1. <模型名>
|
||||
- **一句话**：
|
||||
- **证据场景**：
|
||||
- **应用方式**：
|
||||
- **局限**：
|
||||
- **来源**：
|
||||
|
||||
## 决策启发式（5–10 条 · 如果 X 则 Y，带案例）
|
||||
- 若 <X> → <Y>
|
||||
|
||||
## 表达 DNA
|
||||
| 维度 | 特征 |
|
||||
|------|------|
|
||||
| 句式 | |
|
||||
| 高频词 | |
|
||||
| 幽默 | |
|
||||
| 确定性 | |
|
||||
| 引用习惯 | |
|
||||
|
||||
## 价值观与反模式
|
||||
- **价值观排序**：
|
||||
- **明确反对**：
|
||||
|
||||
## 内在张力（≥2 对矛盾）
|
||||
- <A> ↔ <B>
|
||||
|
||||
## 经历时间线
|
||||
| 年份 | 事件 | 思想意义 |
|
||||
|------|------|---------|
|
||||
| | | |
|
||||
|
||||
## 代表语录
|
||||
- “...” —（出处/年份，标一手/二手）
|
||||
|
||||
## 智识谱系
|
||||
- **受影响于**：
|
||||
- **影响了**：
|
||||
|
||||
## 外部评价与争议
|
||||
-
|
||||
|
||||
## 诚实边界
|
||||
- <至少 3 条具体局限，含信息截止日期>
|
||||
|
||||
## 调研来源
|
||||
**一手(N)**：
|
||||
**二手(N)**：
|
||||
"""
|
||||
|
||||
|
||||
def build_prompt(name: str, hints: str, depth: str) -> str:
|
||||
depth_note = (
|
||||
"本次走【完整模式】：尽量覆盖著作/对话/表达/他者评价/决策/时间线 6 个维度。"
|
||||
if depth == "full"
|
||||
else "本次走【精简模式】：重点覆盖核心观点、心智模型、时间线三块即可。"
|
||||
)
|
||||
hint_block = f"\n补充提示（来自名单其它列，可作为检索关键词）：{hints}\n" if hints.strip() else ""
|
||||
return f"""\
|
||||
你的任务：把人物【{name}】蒸馏成一份 llmwiki 风格的人物档案 markdown。
|
||||
|
||||
请复用本服务中已启用的「女娲 / {NUWA_SKILL}」技能的提炼方法论
|
||||
（心智模型三重验证、表达DNA分析、诚实边界）。
|
||||
|
||||
⚠️ 本环境**无法联网，严禁使用 WebSearch 等任何联网工具**。
|
||||
所有资料**只能通过已配置的知识库查询技能「{KB_SKILL}」检索获取**。
|
||||
请充分调用该知识库技能，围绕该人物多角度检索（生平、观点、著作、语录、决策、评价等），
|
||||
基于检索到的内容做提炼。知识库里查不到的部分，如实在「诚实边界」标注，不要编造。
|
||||
|
||||
{depth_note}
|
||||
{hint_block}
|
||||
输出要求：
|
||||
1. **不要写任何文件，不要调用 write_file / present_files**。
|
||||
2. 直接在你的回复正文里输出**完整中文 markdown 全文**，严格按下面模板结构
|
||||
（保留各级标题，逐项填实，不要保留尖括号占位符），**不要任何额外解说、前言或结语**：
|
||||
|
||||
------------------ 模板开始 ------------------
|
||||
{LLMWIKI_TEMPLATE}
|
||||
------------------ 模板结束 ------------------
|
||||
|
||||
现在开始，独立完成，不要向我提问。"""
|
||||
|
||||
|
||||
# ================================ HTTP 客户端 ================================
|
||||
class DeerFlowClient:
|
||||
def __init__(self, base_url: str):
|
||||
self.base = base_url.rstrip("/")
|
||||
self.s = requests.Session()
|
||||
self.token: str | None = None
|
||||
|
||||
def _headers(self) -> dict:
|
||||
h = {"Content-Type": "application/json"}
|
||||
if self.token:
|
||||
h["Authorization"] = f"Bearer {self.token}" # 只用 Bearer，不带 cookie → 跳过 CSRF
|
||||
return h
|
||||
|
||||
def login(self) -> bool:
|
||||
url = f"{self.base}/api/v1/auth/login/username"
|
||||
params = {"password": PASSWORD} if PASSWORD else None
|
||||
try:
|
||||
r = self.s.post(url, params=params, json={"username": USERNAME}, timeout=HTTP_TIMEOUT)
|
||||
except requests.RequestException as e:
|
||||
print(f" [登录] 连接失败：{e}")
|
||||
return False
|
||||
if r.status_code == 200:
|
||||
self.token = r.json().get("access_token")
|
||||
self.s.cookies.clear()
|
||||
print(f" [登录] 成功，用户={USERNAME}")
|
||||
return True
|
||||
print(f" [登录] 返回 {r.status_code}：{r.text[:200]}")
|
||||
return False
|
||||
|
||||
def ensure_skill_enabled(self, name: str) -> None:
|
||||
try:
|
||||
r = self.s.get(f"{self.base}/api/skills", headers=self._headers(), timeout=HTTP_TIMEOUT)
|
||||
if r.status_code == 200:
|
||||
skills = r.json().get("skills", [])
|
||||
hit = next((x for x in skills if x.get("name") == name), None)
|
||||
if hit is None:
|
||||
print(f" [技能] 警告：服务里没找到 '{name}'")
|
||||
return
|
||||
if hit.get("enabled"):
|
||||
print(f" [技能] '{name}' 已启用")
|
||||
return
|
||||
pr = self.s.put(f"{self.base}/api/skills/{name}", headers=self._headers(),
|
||||
json={"enabled": True}, timeout=HTTP_TIMEOUT)
|
||||
print(f" [技能] 启用 '{name}' → {pr.status_code}")
|
||||
except requests.RequestException as e:
|
||||
print(f" [技能] 检查 '{name}' 失败（忽略）：{e}")
|
||||
|
||||
def create_thread(self) -> str:
|
||||
r = self.s.post(f"{self.base}/api/threads", headers=self._headers(), json={}, timeout=HTTP_TIMEOUT)
|
||||
r.raise_for_status()
|
||||
return r.json()["thread_id"]
|
||||
|
||||
def start_run(self, thread_id: str, prompt: str) -> str:
|
||||
body = {
|
||||
"input": {"messages": [{"role": "user", "content": prompt}]},
|
||||
"excluded_tools": ["ask_clarification"], # 禁澄清，无人值守不卡住
|
||||
}
|
||||
r = self.s.post(f"{self.base}/api/threads/{thread_id}/runs",
|
||||
headers=self._headers(), json=body, timeout=HTTP_TIMEOUT)
|
||||
r.raise_for_status()
|
||||
return r.json()["run_id"]
|
||||
|
||||
def poll_run(self, thread_id: str, run_id: str) -> str:
|
||||
deadline = time.monotonic() + PER_PERSON_TIMEOUT
|
||||
last = ""
|
||||
while time.monotonic() < deadline:
|
||||
try:
|
||||
r = self.s.get(f"{self.base}/api/threads/{thread_id}/runs/{run_id}",
|
||||
headers=self._headers(), timeout=HTTP_TIMEOUT)
|
||||
if r.status_code == 200:
|
||||
st = r.json().get("status", "")
|
||||
if st != last:
|
||||
print(f" ...状态：{st}")
|
||||
last = st
|
||||
if st in TERMINAL_STATUS:
|
||||
return st
|
||||
except requests.RequestException as e:
|
||||
print(f" 轮询出错（重试）：{e}")
|
||||
time.sleep(POLL_INTERVAL)
|
||||
return "timeout"
|
||||
|
||||
def fetch_result_text(self, thread_id: str, run_id: str) -> str | None:
|
||||
"""抓取该 thread 的最终 AI 回复文本（由 py 保存为 md）。
|
||||
|
||||
注意：本部署的 `/runs/{rid}/messages`、`/messages`、`/events` 端点都返回空
|
||||
（事件库按登录用户过滤，后台 worker 写入时未带 user_id → 读不到）。
|
||||
可靠来源是 LangGraph 检查点状态 `/api/threads/{tid}/state` → values.messages，
|
||||
其中每条消息形如 {"type": "human"|"ai"|"tool", "content": <str | 内容块列表>, ...}。
|
||||
取最后一条有正文的 ai 消息即可。
|
||||
"""
|
||||
try:
|
||||
r = self.s.get(f"{self.base}/api/threads/{thread_id}/state",
|
||||
headers=self._headers(), timeout=HTTP_TIMEOUT)
|
||||
if r.status_code != 200:
|
||||
print(f" 取回复返回 {r.status_code}：{r.text[:200]}")
|
||||
return None
|
||||
messages = (r.json().get("values") or {}).get("messages") or []
|
||||
# 从后往前找最后一条「有正文」的 AI 回复（跳过只含工具调用的中间消息）
|
||||
for msg in reversed(messages):
|
||||
if not isinstance(msg, dict) or msg.get("type") != "ai":
|
||||
continue
|
||||
txt = _extract_text(msg)
|
||||
if txt and txt.strip():
|
||||
return txt
|
||||
except requests.RequestException as e:
|
||||
print(f" 取回复出错：{e}")
|
||||
return None
|
||||
|
||||
|
||||
def _extract_text(payload) -> str:
|
||||
"""从一条消息的 content 里提取纯文本。
|
||||
|
||||
payload 可能是：
|
||||
- LangChain 消息的 model_dump 字典（正文在 payload["content"]）
|
||||
- 直接的字符串
|
||||
- 内容块列表（[{"type":"text","text":"..."}, ...]）
|
||||
"""
|
||||
if payload is None:
|
||||
return ""
|
||||
# 若是整条消息的 model_dump，正文在其 content 字段
|
||||
if isinstance(payload, dict):
|
||||
c = payload.get("content", payload.get("text", ""))
|
||||
else:
|
||||
c = payload
|
||||
if isinstance(c, str):
|
||||
return c
|
||||
if isinstance(c, list):
|
||||
parts = []
|
||||
for b in c:
|
||||
if isinstance(b, dict):
|
||||
# 只取文本块，跳过 tool_use / thinking 等非文本块
|
||||
if b.get("type") in (None, "text", "text_block"):
|
||||
parts.append(b.get("text") or b.get("content") or "")
|
||||
elif b.get("text"):
|
||||
parts.append(b["text"])
|
||||
elif isinstance(b, str):
|
||||
parts.append(b)
|
||||
return "\n".join(p for p in parts if p)
|
||||
return str(c) if c else ""
|
||||
|
||||
|
||||
def _clean_markdown(text: str) -> str:
|
||||
"""去掉 agent 可能套上的 ```markdown ... ``` 围栏，并规整 YAML frontmatter。"""
|
||||
t = text.strip()
|
||||
m = re.match(r"^```(?:markdown|md)?\s*\n(.*)\n```$", t, re.S)
|
||||
if m:
|
||||
t = m.group(1).strip()
|
||||
return _normalize_frontmatter(t)
|
||||
|
||||
|
||||
def _normalize_frontmatter(text: str) -> str:
|
||||
"""修复模型常见的 frontmatter 跑偏，让 YAML 能正常解析。
|
||||
|
||||
模型经常把字段名写成 markdown 加粗（``**name:**``）、加空行、留行尾双空格，
|
||||
这些都会让 frontmatter 解析失败。这里把开头 ``---`` 到下一个 ``---`` 之间的内容
|
||||
规整为干净的 ``key: value`` 形式。正文（第二个 ``---`` 之后）保持原样。
|
||||
"""
|
||||
if not text.startswith("---"):
|
||||
return text
|
||||
# 拆出 frontmatter 块：开头 --- 与下一行单独的 --- 之间
|
||||
m = re.match(r"^---[ \t]*\n(.*?)\n---[ \t]*\n?(.*)$", text, re.S)
|
||||
if not m:
|
||||
return text
|
||||
body_fm, rest = m.group(1), m.group(2)
|
||||
out_lines = []
|
||||
for line in body_fm.splitlines():
|
||||
s = line.rstrip() # 去行尾双空格
|
||||
s = s.replace("**", "") # 去 markdown 加粗标记
|
||||
if not s.strip():
|
||||
continue # 丢掉空行
|
||||
out_lines.append(s)
|
||||
return "---\n" + "\n".join(out_lines) + "\n---\n\n" + rest.lstrip("\n")
|
||||
|
||||
|
||||
# ================================ 名单读取 ================================
|
||||
def slugify(name: str) -> str:
|
||||
s = re.sub(r'[\\/:*?"<>|]+', "", name).strip()
|
||||
s = re.sub(r"\s+", "_", s)
|
||||
return s or "unnamed"
|
||||
|
||||
|
||||
def read_people(path: Path) -> list[dict]:
|
||||
if not path.exists():
|
||||
return []
|
||||
rows: list[list[str]] = []
|
||||
if path.suffix.lower() in (".xlsx", ".xlsm"):
|
||||
try:
|
||||
import openpyxl
|
||||
except ImportError:
|
||||
sys.exit("读取 .xlsx 需要 openpyxl：pip install openpyxl（或把名单存成 .csv）")
|
||||
wb = openpyxl.load_workbook(path, read_only=True, data_only=True)
|
||||
for row in wb.active.iter_rows(values_only=True):
|
||||
rows.append(["" if c is None else str(c).strip() for c in row])
|
||||
elif path.suffix.lower() == ".csv":
|
||||
with open(path, encoding="utf-8-sig", newline="") as f:
|
||||
rows = [[c.strip() for c in r] for r in csv.reader(f)]
|
||||
else:
|
||||
sys.exit(f"不支持的名单格式：{path.suffix}（用 .xlsx 或 .csv）")
|
||||
|
||||
if rows and re.search(r"(姓名|名字|人物|name)", rows[0][0] if rows[0] else "", re.I):
|
||||
rows = rows[1:]
|
||||
people, seen = [], set()
|
||||
for r in rows:
|
||||
if not r or not r[0].strip():
|
||||
continue
|
||||
name = r[0].strip()
|
||||
if name in seen:
|
||||
continue
|
||||
seen.add(name)
|
||||
hints = " | ".join(x for x in r[1:] if x.strip())
|
||||
people.append({"name": name, "hints": hints})
|
||||
return people
|
||||
|
||||
|
||||
# ================================ 时间窗 ================================
|
||||
def in_window(now: datetime | None = None) -> bool:
|
||||
now = now or datetime.now()
|
||||
t = now.time()
|
||||
return t >= WIN_START or t < WIN_END # 跨午夜
|
||||
|
||||
def seconds_until_window(now: datetime | None = None) -> int:
|
||||
now = now or datetime.now()
|
||||
if in_window(now):
|
||||
return 0
|
||||
target = now.replace(hour=WIN_START.hour, minute=WIN_START.minute, second=0, microsecond=0)
|
||||
return max(1, int((target - now).total_seconds()))
|
||||
|
||||
def sleep_interruptible(total: int) -> None:
|
||||
end = time.monotonic() + total
|
||||
while time.monotonic() < end:
|
||||
time.sleep(min(30, max(1, int(end - time.monotonic()))))
|
||||
|
||||
|
||||
# ================================ 蒸馏单人 ================================
|
||||
def distill_one(client: DeerFlowClient, person: dict, out_dir: Path,
|
||||
manifest: dict, save_manifest, depth: str) -> None:
|
||||
name, hints = person["name"], person["hints"]
|
||||
slug = slugify(name)
|
||||
md_path = out_dir / f"{slug}.md"
|
||||
t0 = time.time()
|
||||
|
||||
rec = {"name": name, "slug": slug, "status": "running",
|
||||
"started_at": _now(), "file": f"{slug}.md"}
|
||||
manifest[name] = rec
|
||||
save_manifest()
|
||||
|
||||
try:
|
||||
tid = client.create_thread()
|
||||
run_id = client.start_run(tid, build_prompt(name, hints, depth))
|
||||
status = client.poll_run(tid, run_id)
|
||||
rec.update(thread_id=tid, run_id=run_id, run_status=status)
|
||||
|
||||
text = client.fetch_result_text(tid, run_id) if status == "success" else None
|
||||
if status == "success" and text:
|
||||
md = _clean_markdown(text)
|
||||
md_path.write_text(md, encoding="utf-8")
|
||||
rec["status"] = "succeeded"
|
||||
rec["chars"] = len(md)
|
||||
print(f" ✅ 完成 → {md_path.name}（{len(md)} 字，耗时 {int(time.time()-t0)}s）")
|
||||
else:
|
||||
rec["status"] = "failed"
|
||||
rec["error"] = f"run={status}, text={'有' if text else '无'}"
|
||||
if text:
|
||||
(out_dir / f"{slug}.partial.md").write_text(text, encoding="utf-8")
|
||||
print(f" ❌ 失败（run={status}）")
|
||||
except Exception as e:
|
||||
rec["status"] = "failed"
|
||||
rec["error"] = str(e)
|
||||
print(f" ❌ 异常：{e}")
|
||||
|
||||
rec["finished_at"] = _now()
|
||||
rec["seconds"] = int(time.time() - t0)
|
||||
manifest[name] = rec
|
||||
save_manifest()
|
||||
|
||||
|
||||
# ================================ 主流程 ================================
|
||||
def main():
|
||||
ap = argparse.ArgumentParser(description="批量人物蒸馏 → llmwiki（定时窗 23:00~08:00）")
|
||||
ap.add_argument("excel", nargs="?", default=str(PEOPLE_FILE_DEFAULT),
|
||||
help="名单文件（.xlsx 或 .csv），默认同目录 people.xlsx")
|
||||
ap.add_argument("--now", action="store_true", help="忽略时间窗，立刻跑完整批（测试用）")
|
||||
ap.add_argument("--redo", action="store_true", help="忽略已完成，全部重做")
|
||||
args = ap.parse_args()
|
||||
|
||||
excel = Path(args.excel)
|
||||
OUTPUT_DIR.mkdir(parents=True, exist_ok=True)
|
||||
manifest_path = OUTPUT_DIR / "manifest.json"
|
||||
|
||||
manifest: dict = {}
|
||||
if manifest_path.exists():
|
||||
try:
|
||||
manifest = json.loads(manifest_path.read_text(encoding="utf-8"))
|
||||
except Exception:
|
||||
manifest = {}
|
||||
|
||||
# 校验名单存在
|
||||
if not read_people(excel):
|
||||
sys.exit(f"名单为空或不存在：{excel}\n请放 people.xlsx（第一列写人名）。")
|
||||
|
||||
# --redo：一次性清掉名单里这些人的旧记录，让他们重做一遍；
|
||||
# 之后本次会话内跑成功的会正常计为「已完成」，循环才会往后推进
|
||||
# （否则每轮 pending() 都把已完成的人当未完成，永远卡在第 1 个人）
|
||||
if args.redo:
|
||||
for p in read_people(excel):
|
||||
manifest.pop(p["name"], None)
|
||||
|
||||
client = DeerFlowClient(BASE_URL)
|
||||
if not client.login():
|
||||
sys.exit("登录失败。请检查 BASE_URL 是否指向正在运行的服务、口令是否正确。")
|
||||
client.ensure_skill_enabled(NUWA_SKILL)
|
||||
client.ensure_skill_enabled(KB_SKILL)
|
||||
|
||||
def save_manifest():
|
||||
manifest_path.write_text(json.dumps(manifest, ensure_ascii=False, indent=2), encoding="utf-8")
|
||||
_write_index(OUTPUT_DIR, read_people(excel), manifest)
|
||||
|
||||
def pending() -> list[dict]:
|
||||
people = read_people(excel)
|
||||
out = []
|
||||
for p in people:
|
||||
r = manifest.get(p["name"], {})
|
||||
done = r.get("status") == "succeeded" and (OUTPUT_DIR / f"{slugify(p['name'])}.md").exists()
|
||||
if not done:
|
||||
out.append(p)
|
||||
return out
|
||||
|
||||
print(f"启动 | 服务={BASE_URL} | 输出={OUTPUT_DIR}")
|
||||
print(f"模式={'立即跑完' if args.now else '定时窗 23:00~08:00'} | 名单={excel}")
|
||||
|
||||
try:
|
||||
while True:
|
||||
todo = pending()
|
||||
if not todo:
|
||||
print(f"[{_now()}] 全部已完成，待命中（每小时复查名单）...")
|
||||
if args.now:
|
||||
break
|
||||
sleep_interruptible(3600)
|
||||
continue
|
||||
|
||||
if not args.now and not in_window():
|
||||
wait = seconds_until_window()
|
||||
nxt = datetime.now().replace(hour=WIN_START.hour, minute=0, second=0, microsecond=0)
|
||||
print(f"[{_now()}] 不在执行窗口，剩 {len(todo)} 人待蒸；"
|
||||
f"休眠到 {nxt.strftime('%H:%M')}（约 {wait//60} 分钟）...")
|
||||
sleep_interruptible(wait)
|
||||
continue
|
||||
|
||||
# 在窗口内（或 --now）：处理一个人，然后回到循环重新判断窗口/名单
|
||||
p = todo[0]
|
||||
idx_total = len(read_people(excel))
|
||||
done_n = sum(1 for v in manifest.values() if v.get("status") == "succeeded")
|
||||
print(f"[{_now()}] ({done_n}/{idx_total} 已完成) 蒸馏：{p['name']}")
|
||||
distill_one(client, p, OUTPUT_DIR, manifest, save_manifest, DEPTH)
|
||||
except KeyboardInterrupt:
|
||||
print("\n已手动停止。重跑脚本会从未完成处继续。")
|
||||
|
||||
ok = sum(1 for r in manifest.values() if r.get("status") == "succeeded")
|
||||
fail = sum(1 for r in manifest.values() if r.get("status") == "failed")
|
||||
print(f"\n结束：成功 {ok} / 失败 {fail}")
|
||||
print(f"结果目录：{OUTPUT_DIR} | 清单：{manifest_path} | 索引：{OUTPUT_DIR/'index.md'}")
|
||||
|
||||
|
||||
def _now() -> str:
|
||||
return time.strftime("%Y-%m-%d %H:%M:%S")
|
||||
|
||||
|
||||
def _write_index(out_dir: Path, people: list[dict], manifest: dict) -> None:
|
||||
lines = ["# 蒸馏执行清单", "", f"更新时间：{_now()}", "",
|
||||
"| # | 人物 | 状态 | 字数 | 耗时 | 文件 |",
|
||||
"|---|------|------|------|------|------|"]
|
||||
badge = {"succeeded": "✅完成", "failed": "❌失败", "running": "⏳进行中"}
|
||||
for i, p in enumerate(people, 1):
|
||||
r = manifest.get(p["name"], {})
|
||||
st = badge.get(r.get("status", ""), "⌛排队")
|
||||
f = r.get("file", "")
|
||||
link = f"[{f}]({f})" if r.get("status") == "succeeded" and f else "-"
|
||||
lines.append(f"| {i} | {p['name']} | {st} | {r.get('chars','-')} | {r.get('seconds','-')} | {link} |")
|
||||
(out_dir / "index.md").write_text("\n".join(lines) + "\n", encoding="utf-8")
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
main()
|
||||
@@ -1,4 +0,0 @@
|
||||
姓名,领域(可选提示),资料链接或备注(可选)
|
||||
查理·芒格,投资/多元思维,
|
||||
理查德·费曼,物理/教学,
|
||||
纳瓦尔·拉维坎特,创业/杠杆,https://nav.al/
|
||||
|
@@ -1,136 +0,0 @@
|
||||
# zncm Web Clipper
|
||||
|
||||
这是给 zncm 使用的 Chrome 插件，安装后默认连接 `config.js` 中配置好的服务地址。
|
||||
|
||||
默认地址配置在 `browser-extension/deer-web-clipper/config.js`：
|
||||
|
||||
```js
|
||||
globalThis.ZNCM_WEB_CLIPPER_CONFIG = Object.freeze({
|
||||
defaultWebBaseUrl: "http://47.94.209.59:2026",
|
||||
defaultApiBaseUrl: "http://47.94.209.59:2026",
|
||||
});
|
||||
```
|
||||
|
||||
如果部署地址变化，只需要修改 `config.js` 里的这两个字段，然后重新加载插件。
|
||||
|
||||
正常情况下不需要手动配置 URL，也不需要手动复制 token。
|
||||
|
||||
## 能做什么
|
||||
|
||||
1. 把当前浏览器标签页发送到 zncm，作为 Markdown 附件进入聊天。
|
||||
2. 接收 zncm 聊天中由 `browser_fetch_page` 工具创建的网页抓取任务。
|
||||
3. 插件在 Chrome 中打开目标网页，读取可见文本和表格，再把结果回传给 zncm。
|
||||
4. 在普通网页右下角显示“zncm 页面问答”浮窗，直接把当前页面内容交给 zncm 问答。
|
||||
5. 接收 zncm 的 `browser_act` 交互式操作任务：在 Chrome 中打开页面并按 zncm 的指令**操作页面**——点击、输入、滚动、等待、页内跳转，每一步把页面上的可交互元素清单和正文回传给 zncm，由 zncm 决定下一步（observe→act 循环）。
|
||||
|
||||
## 页面操作（browser_act）
|
||||
|
||||
除了只读抓取，zncm 还能驱动插件**实际操作网页**，完成登录、填表、翻页找数据等多步任务：
|
||||
|
||||
1. zncm 调用 `browser_act` 工具，先 `navigate` 打开页面。
|
||||
2. 插件打开标签页（同一会话会复用这个标签页），扫描页面上的可交互元素，给每个元素编号后连同正文回传给 zncm。
|
||||
3. zncm 根据元素编号决定下一步（点第 3 个按钮、在第 1 个输入框填字……），再次调用 `browser_act`。
|
||||
4. 插件执行动作、重新扫描页面、回传新状态，如此循环，直到任务完成。
|
||||
5. 任务结束时 zncm 用 `action="close"` 关闭该会话标签页。
|
||||
|
||||
支持的动作：`navigate`（打开/跳转 URL）、`click`（点击元素）、`type`（输入文字，可选回车提交）、`scroll`（滚动到元素或下翻一屏）、`wait`（等待某 CSS 选择器出现）、`extract`（重新读取正文）、`observe`（仅重新扫描）、`close`（关闭会话标签页）。
|
||||
|
||||
> 提示：页面操作能力较强（可点击任意按钮、提交表单），可在后端 `config.yaml` 用 `browser_context.actions_enabled: false` 关闭，只保留只读抓取。
|
||||
|
||||
## 后端启用
|
||||
|
||||
在 `offline-backend-20260512/backend/config.yaml` 中保持：
|
||||
|
||||
```yaml
|
||||
browser_context:
|
||||
enabled: true # 只读抓取 + 页面问答
|
||||
actions_enabled: true # 交互式页面操作 browser_act（不需要可设为 false）
|
||||
```
|
||||
|
||||
修改配置后需要重启后端。
|
||||
|
||||
## 安装插件
|
||||
|
||||
1. 打开 Chrome 的 `chrome://extensions`。
|
||||
2. 开启右上角的“开发者模式”。
|
||||
3. 点击“加载已解压的扩展程序”。
|
||||
4. 选择目录：`browser-extension/deer-web-clipper`。
|
||||
5. 安装后插件名会显示为 `zncm Web Clipper`。
|
||||
|
||||
## 安装即用
|
||||
|
||||
插件默认已经开启：
|
||||
|
||||
- 自动识别已打开 zncm 页面的服务地址。
|
||||
- 自动同步已登录 zncm 的 token。
|
||||
- 自动轮询 zncm 聊天触发的浏览器抓取任务。
|
||||
- 页面问答框默认开启。
|
||||
|
||||
如果你打开的是 `config.js` 里配置的默认服务器，通常无需进入设置页。
|
||||
|
||||
如果部署地址变了，保持一个已登录的 zncm 页面打开，然后在插件设置里点击：
|
||||
|
||||
- `重新识别服务地址`
|
||||
- `同步 token`
|
||||
|
||||
## 自动同步 Token
|
||||
|
||||
插件会自动从两个位置获取登录 token：
|
||||
|
||||
1. zncm API 域名下的 `access_token` Cookie，包括 HttpOnly Cookie。
|
||||
2. 已登录 zncm 前端页面里的浏览器本地登录态。
|
||||
|
||||
切换登录账号后，插件后台每次轮询任务前都会重新同步 token；点击插件里的“同步 token”也可以立即刷新。
|
||||
|
||||
如果当前浏览器已经登出 zncm，且打开的 zncm 页面里没有 token，插件会清空旧 token，避免继续用上一个账号请求。
|
||||
|
||||
## 页面问答框
|
||||
|
||||
安装插件后，普通 `http` / `https` 页面右下角会出现 `AI` 按钮。点击后可以在当前网页上直接提问，例如：
|
||||
|
||||
```text
|
||||
总结这页的核心结论
|
||||
提取页面里的价格和时间
|
||||
这张表里哪个指标最高？
|
||||
```
|
||||
|
||||
插件会读取当前页面可见文本和表格，发送到 zncm，并把 zncm 的回答显示回浮窗里。多轮追问会复用同一个 zncm 会话；如果页面 URL 变化，会自动重新开始当前页会话。
|
||||
|
||||
是否显示问答框可以自己控制：
|
||||
|
||||
- 插件弹窗里勾选或取消“启用页面问答框”。
|
||||
- 插件设置页里勾选或取消“启用页面问答框”。
|
||||
|
||||
关闭问答框只会隐藏页面浮窗，不会影响“发送当前页”和 zncm 聊天中的 `browser_fetch_page` 抓取任务。
|
||||
|
||||
## 手动发送当前网页
|
||||
|
||||
点击插件图标，选择“发送当前页”。
|
||||
|
||||
插件会读取当前页面的可见文本和表格，发送到 zncm，并打开一个新的 zncm 聊天页面。zncm 会把网页内容作为上传文件注入到首条问题里。
|
||||
|
||||
## 在 zncm 聊天中触发浏览器抓取
|
||||
|
||||
在 zncm 中输入带有明确 URL 的问题，例如：
|
||||
|
||||
```text
|
||||
使用 browser_fetch_page 打开 https://example.com/report，提取页面里的关键表格数据并总结。
|
||||
```
|
||||
|
||||
执行流程：
|
||||
|
||||
1. zncm 调用 `browser_fetch_page` 工具创建浏览器任务。
|
||||
2. 插件后台轮询到任务。
|
||||
3. 插件在 Chrome 中打开目标网页。
|
||||
4. 插件提取页面文本和表格。
|
||||
5. 插件把结果回传给 zncm。
|
||||
6. zncm 基于回传内容回答你的问题。
|
||||
|
||||
这个方式适合读取需要浏览器登录态、内网权限或当前 Chrome 环境才能访问的页面。
|
||||
|
||||
## 注意事项
|
||||
|
||||
- 目标 URL 必须是明确的 `http` 或 `https` 地址。
|
||||
- 插件不能读取 `chrome://` 等浏览器内部页面。
|
||||
- 如果目标网站阻止扩展脚本注入，抓取可能失败。
|
||||
- 后台任务默认会打开新标签页；是否抓取完成后关闭标签页，可在插件设置里配置。
|
||||
@@ -1,905 +0,0 @@
|
||||
importScripts("config.js", "shared.js");
|
||||
|
||||
const ALARM_NAME = "deer-browser-task-poll";
|
||||
const LEGACY_LOCAL_API_BASE_URL = "http://localhost:8001";
|
||||
const LEGACY_LOCAL_WEB_BASE_URL = "http://localhost:5174";
|
||||
let pollTimer = null;
|
||||
let isPolling = false;
|
||||
|
||||
function delay(ms) {
|
||||
return new Promise((resolve) => setTimeout(resolve, ms));
|
||||
}
|
||||
|
||||
async function getSettings() {
|
||||
const settings = await chrome.storage.sync.get(DEER_CLIPPER_DEFAULTS);
|
||||
const updates = {};
|
||||
if (!settings.apiBaseUrl || settings.apiBaseUrl === LEGACY_LOCAL_API_BASE_URL) {
|
||||
updates.apiBaseUrl = DEER_CLIPPER_DEFAULTS.apiBaseUrl;
|
||||
}
|
||||
if (!settings.webBaseUrl || settings.webBaseUrl === LEGACY_LOCAL_WEB_BASE_URL) {
|
||||
updates.webBaseUrl = DEER_CLIPPER_DEFAULTS.webBaseUrl;
|
||||
}
|
||||
if (!settings.workerId) {
|
||||
updates.workerId = deerMakeWorkerId();
|
||||
}
|
||||
if (Object.keys(updates).length > 0) {
|
||||
await chrome.storage.sync.set(updates);
|
||||
Object.assign(settings, updates);
|
||||
}
|
||||
return settings;
|
||||
}
|
||||
|
||||
function pollIntervalMs(settings) {
|
||||
const seconds = Number(settings.taskPollIntervalSeconds || DEER_CLIPPER_DEFAULTS.taskPollIntervalSeconds);
|
||||
return Math.max(3, seconds) * 1000;
|
||||
}
|
||||
|
||||
function originMatchPattern(value) {
|
||||
try {
|
||||
const url = new URL(deerNormalizeBaseUrl(value));
|
||||
return `${url.protocol}//${url.host}/*`;
|
||||
} catch {
|
||||
return "";
|
||||
}
|
||||
}
|
||||
|
||||
function normalizeServiceBase(value) {
|
||||
return deerNormalizeBaseUrl(value || "");
|
||||
}
|
||||
|
||||
function serviceBaseFromTabUrl(value) {
|
||||
try {
|
||||
const url = new URL(value);
|
||||
const path = url.pathname.replace(/\/+$/, "");
|
||||
if (path === "/deerflow" || path.startsWith("/deerflow/")) {
|
||||
return `${url.origin}/deerflow`;
|
||||
}
|
||||
return url.origin;
|
||||
} catch {
|
||||
return "";
|
||||
}
|
||||
}
|
||||
|
||||
function apiCandidatesForWebBase(webBase, currentApiBase) {
|
||||
const candidates = [];
|
||||
const add = (value) => {
|
||||
const normalized = normalizeServiceBase(value);
|
||||
if (normalized && !candidates.includes(normalized)) {
|
||||
candidates.push(normalized);
|
||||
}
|
||||
};
|
||||
|
||||
add(currentApiBase);
|
||||
add(webBase);
|
||||
add(DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
add(LEGACY_LOCAL_API_BASE_URL);
|
||||
|
||||
try {
|
||||
const url = new URL(normalizeServiceBase(webBase));
|
||||
if (["5173", "5174", "3000"].includes(url.port)) {
|
||||
const apiUrl = new URL(url.href);
|
||||
apiUrl.port = "8001";
|
||||
add(apiUrl.href);
|
||||
}
|
||||
} catch {
|
||||
// Ignore invalid candidates.
|
||||
}
|
||||
|
||||
return candidates;
|
||||
}
|
||||
|
||||
async function browserContextStatus(apiBaseUrl) {
|
||||
try {
|
||||
const response = await fetch(`${apiBaseUrl}/api/browser-context/status`, {
|
||||
credentials: "include",
|
||||
});
|
||||
if (!response.ok) {
|
||||
return null;
|
||||
}
|
||||
const data = await response.json().catch(() => null);
|
||||
if (data && typeof data.enabled === "boolean") {
|
||||
return data;
|
||||
}
|
||||
} catch {
|
||||
return null;
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
function tabLooksLikeZncm(tab) {
|
||||
const value = `${tab.url || ""} ${tab.title || ""}`.toLowerCase();
|
||||
return (
|
||||
value.includes("deerflow") ||
|
||||
value.includes("deer") ||
|
||||
value.includes("zncm") ||
|
||||
value.includes("/page/workspace") ||
|
||||
value.includes("/chats/") ||
|
||||
value.includes("#/login")
|
||||
);
|
||||
}
|
||||
|
||||
async function findOpenZncmTabs(settings) {
|
||||
const tabs = [];
|
||||
const seen = new Set();
|
||||
const addTabs = (items) => {
|
||||
for (const tab of items || []) {
|
||||
if (tab?.id && !seen.has(tab.id)) {
|
||||
seen.add(tab.id);
|
||||
tabs.push(tab);
|
||||
}
|
||||
}
|
||||
};
|
||||
|
||||
for (const pattern of [
|
||||
originMatchPattern(settings.webBaseUrl || DEER_CLIPPER_DEFAULTS.webBaseUrl),
|
||||
originMatchPattern(DEER_CLIPPER_DEFAULTS.webBaseUrl),
|
||||
]) {
|
||||
if (!pattern) {
|
||||
continue;
|
||||
}
|
||||
try {
|
||||
addTabs(await chrome.tabs.query({ url: pattern }));
|
||||
} catch {
|
||||
// Continue with broader discovery below.
|
||||
}
|
||||
}
|
||||
|
||||
try {
|
||||
const allTabs = await chrome.tabs.query({ url: ["http://*/*", "https://*/*"] });
|
||||
addTabs(allTabs.filter(tabLooksLikeZncm));
|
||||
} catch {
|
||||
// Ignore tab discovery failures.
|
||||
}
|
||||
|
||||
return tabs;
|
||||
}
|
||||
|
||||
async function discoverServiceBases(settings, force = false) {
|
||||
if (!settings.autoDiscoverServiceUrl && !force) {
|
||||
return settings;
|
||||
}
|
||||
|
||||
let webBaseUrl = normalizeServiceBase(settings.webBaseUrl || DEER_CLIPPER_DEFAULTS.webBaseUrl);
|
||||
let discoveredWebBaseUrl = webBaseUrl;
|
||||
let source = "configured";
|
||||
const tabs = await findOpenZncmTabs(settings);
|
||||
if (tabs.length > 0) {
|
||||
const tabBase = serviceBaseFromTabUrl(tabs[0].url);
|
||||
if (tabBase) {
|
||||
discoveredWebBaseUrl = tabBase;
|
||||
source = `tab:${tabs[0].url || tabBase}`;
|
||||
}
|
||||
}
|
||||
|
||||
let apiBaseUrl = normalizeServiceBase(settings.apiBaseUrl || DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
let verified = false;
|
||||
for (const candidate of apiCandidatesForWebBase(discoveredWebBaseUrl, apiBaseUrl)) {
|
||||
const status = await browserContextStatus(candidate);
|
||||
if (status) {
|
||||
apiBaseUrl = candidate;
|
||||
webBaseUrl = discoveredWebBaseUrl;
|
||||
verified = true;
|
||||
source = status.enabled ? `${source}:browser_context_enabled` : `${source}:browser_context_disabled`;
|
||||
break;
|
||||
}
|
||||
}
|
||||
if (!verified && force) {
|
||||
source = `${source}:api_not_verified`;
|
||||
}
|
||||
|
||||
const next = {
|
||||
...settings,
|
||||
apiBaseUrl,
|
||||
webBaseUrl,
|
||||
lastServiceSyncAt: new Date().toISOString(),
|
||||
lastServiceSyncSource: source,
|
||||
};
|
||||
|
||||
if (
|
||||
force ||
|
||||
settings.apiBaseUrl !== next.apiBaseUrl ||
|
||||
settings.webBaseUrl !== next.webBaseUrl ||
|
||||
settings.lastServiceSyncSource !== next.lastServiceSyncSource
|
||||
) {
|
||||
await chrome.storage.sync.set({
|
||||
apiBaseUrl: next.apiBaseUrl,
|
||||
webBaseUrl: next.webBaseUrl,
|
||||
lastServiceSyncAt: next.lastServiceSyncAt,
|
||||
lastServiceSyncSource: next.lastServiceSyncSource,
|
||||
});
|
||||
}
|
||||
return next;
|
||||
}
|
||||
|
||||
async function getCookieToken(settings) {
|
||||
if (!chrome.cookies?.get) {
|
||||
return null;
|
||||
}
|
||||
const apiBaseUrl = deerNormalizeBaseUrl(settings.apiBaseUrl || DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
try {
|
||||
const cookie = await chrome.cookies.get({ url: apiBaseUrl, name: "access_token" });
|
||||
if (cookie?.value) {
|
||||
return { token: cookie.value, source: `cookie:${new URL(apiBaseUrl).host}:access_token` };
|
||||
}
|
||||
} catch (error) {
|
||||
console.warn("Failed to read zncm access_token cookie", error);
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
async function getTokenFromOpenZncmTab(settings) {
|
||||
const tabs = await findOpenZncmTabs(settings);
|
||||
let sawZncmTab = false;
|
||||
|
||||
for (const tab of tabs) {
|
||||
if (!tab.id) {
|
||||
continue;
|
||||
}
|
||||
sawZncmTab = true;
|
||||
try {
|
||||
const [result] = await chrome.scripting.executeScript({
|
||||
target: { tabId: tab.id },
|
||||
func: deerExtractAuthTokenFromPage,
|
||||
});
|
||||
const tokenInfo = result?.result;
|
||||
if (tokenInfo?.token) {
|
||||
const tabBase = serviceBaseFromTabUrl(tab.url);
|
||||
if (tabBase && settings.autoDiscoverServiceUrl && settings.webBaseUrl !== tabBase) {
|
||||
await chrome.storage.sync.set({
|
||||
webBaseUrl: tabBase,
|
||||
lastServiceSyncAt: new Date().toISOString(),
|
||||
lastServiceSyncSource: `token-tab:${tab.url || tabBase}`,
|
||||
});
|
||||
}
|
||||
return {
|
||||
token: tokenInfo.token,
|
||||
source: tokenInfo.source || `tab:${tab.url || "open-deer-tab"}`,
|
||||
};
|
||||
}
|
||||
} catch (error) {
|
||||
console.warn("Failed to read zncm token from tab", tab.id, error);
|
||||
}
|
||||
}
|
||||
|
||||
if (sawZncmTab) {
|
||||
return { token: "", source: "open-deer-tab", clearStale: true };
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
async function syncAuthToken(settings = null, forceStatus = false) {
|
||||
let current = settings || (await getSettings());
|
||||
current = await discoverServiceBases(current);
|
||||
if (!current.tokenSyncEnabled && !forceStatus) {
|
||||
return current;
|
||||
}
|
||||
|
||||
const tokenInfo = (await getTokenFromOpenZncmTab(current)) || (await getCookieToken(current));
|
||||
if (!tokenInfo?.token) {
|
||||
if (tokenInfo?.clearStale && current.accessToken) {
|
||||
const next = {
|
||||
...current,
|
||||
accessToken: "",
|
||||
lastTokenSyncSource: "cleared:open-deer-tab-no-token",
|
||||
lastTokenSyncAt: new Date().toISOString(),
|
||||
};
|
||||
await chrome.storage.sync.set({
|
||||
accessToken: "",
|
||||
lastTokenSyncSource: next.lastTokenSyncSource,
|
||||
lastTokenSyncAt: next.lastTokenSyncAt,
|
||||
});
|
||||
return next;
|
||||
}
|
||||
return current;
|
||||
}
|
||||
|
||||
const next = {
|
||||
...current,
|
||||
accessToken: tokenInfo.token,
|
||||
lastTokenSyncSource: tokenInfo.source,
|
||||
lastTokenSyncAt: new Date().toISOString(),
|
||||
};
|
||||
if (
|
||||
forceStatus ||
|
||||
current.accessToken !== next.accessToken ||
|
||||
current.lastTokenSyncSource !== next.lastTokenSyncSource
|
||||
) {
|
||||
await chrome.storage.sync.set({
|
||||
accessToken: next.accessToken,
|
||||
lastTokenSyncSource: next.lastTokenSyncSource,
|
||||
lastTokenSyncAt: next.lastTokenSyncAt,
|
||||
});
|
||||
}
|
||||
return next;
|
||||
}
|
||||
|
||||
async function refreshAlarm(settings) {
|
||||
await chrome.alarms.clear(ALARM_NAME);
|
||||
if (!settings.taskPollingEnabled) {
|
||||
return;
|
||||
}
|
||||
const minutes = Math.max(0.5, pollIntervalMs(settings) / 60000);
|
||||
await chrome.alarms.create(ALARM_NAME, { delayInMinutes: minutes, periodInMinutes: minutes });
|
||||
}
|
||||
|
||||
async function scheduleNextPoll(delayMs = null) {
|
||||
if (pollTimer) {
|
||||
clearTimeout(pollTimer);
|
||||
pollTimer = null;
|
||||
}
|
||||
const settings = await getSettings();
|
||||
if (!settings.taskPollingEnabled) {
|
||||
return;
|
||||
}
|
||||
const ms = delayMs ?? pollIntervalMs(settings);
|
||||
pollTimer = setTimeout(() => {
|
||||
void pollOnce().catch((error) => console.warn("zncm task poll failed", error));
|
||||
}, ms);
|
||||
}
|
||||
|
||||
async function waitForTabLoad(tabId, timeoutSeconds) {
|
||||
const timeoutMs = Math.max(5, Number(timeoutSeconds || 30)) * 1000;
|
||||
try {
|
||||
const tab = await chrome.tabs.get(tabId);
|
||||
if (tab.status === "complete") {
|
||||
return true;
|
||||
}
|
||||
} catch {
|
||||
return false;
|
||||
}
|
||||
|
||||
return new Promise((resolve) => {
|
||||
const timer = setTimeout(() => {
|
||||
chrome.tabs.onUpdated.removeListener(listener);
|
||||
resolve(false);
|
||||
}, timeoutMs);
|
||||
|
||||
function listener(updatedTabId, changeInfo) {
|
||||
if (updatedTabId === tabId && changeInfo.status === "complete") {
|
||||
clearTimeout(timer);
|
||||
chrome.tabs.onUpdated.removeListener(listener);
|
||||
resolve(true);
|
||||
}
|
||||
}
|
||||
|
||||
chrome.tabs.onUpdated.addListener(listener);
|
||||
});
|
||||
}
|
||||
|
||||
async function captureFromTab(tabId, maxContentChars) {
|
||||
const [result] = await chrome.scripting.executeScript({
|
||||
target: { tabId },
|
||||
func: deerExtractCurrentPage,
|
||||
});
|
||||
const page = result?.result;
|
||||
if (!page?.content) {
|
||||
throw new Error("当前页面没有可读取的文本。");
|
||||
}
|
||||
const limit = Math.max(1000, Number(maxContentChars || 200000));
|
||||
if (page.content.length > limit) {
|
||||
page.content = `${page.content.slice(0, limit)}\n\n[Truncated by zncm browser task limit]`;
|
||||
}
|
||||
return page;
|
||||
}
|
||||
|
||||
async function postTaskResult(taskId, settings, payload) {
|
||||
const apiBaseUrl = deerNormalizeBaseUrl(settings.apiBaseUrl || DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
const response = await fetch(`${apiBaseUrl}/api/browser-context/tasks/${encodeURIComponent(taskId)}/result`, {
|
||||
method: "POST",
|
||||
headers: {
|
||||
"Content-Type": "application/json",
|
||||
...deerAuthHeaders(settings),
|
||||
},
|
||||
body: JSON.stringify(payload),
|
||||
});
|
||||
const data = await response.json().catch(() => ({}));
|
||||
if (!response.ok) {
|
||||
const detail = typeof data.detail === "string" ? data.detail : JSON.stringify(data.detail || data);
|
||||
throw new Error(detail || `HTTP ${response.status}`);
|
||||
}
|
||||
return data;
|
||||
}
|
||||
|
||||
function zncmApiBaseCandidates(settings) {
|
||||
const candidates = [];
|
||||
const add = (value) => {
|
||||
const normalized = deerNormalizeBaseUrl(value || "");
|
||||
if (normalized && !candidates.includes(normalized)) {
|
||||
candidates.push(normalized);
|
||||
}
|
||||
};
|
||||
|
||||
add(settings.apiBaseUrl || DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
try {
|
||||
const url = new URL(deerNormalizeBaseUrl(settings.apiBaseUrl || ""));
|
||||
if (url.pathname && url.pathname !== "/") {
|
||||
add(url.origin);
|
||||
}
|
||||
} catch {
|
||||
// Ignore invalid configured API URL.
|
||||
}
|
||||
add(DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
return candidates;
|
||||
}
|
||||
|
||||
async function zncmFetchJson(settings, path, options) {
|
||||
let lastError = null;
|
||||
for (const apiBaseUrl of zncmApiBaseCandidates(settings)) {
|
||||
try {
|
||||
const response = await fetch(`${apiBaseUrl}${path}`, options);
|
||||
const contentType = response.headers.get("content-type") || "";
|
||||
const data = contentType.includes("application/json")
|
||||
? await response.json().catch(() => ({}))
|
||||
: { detail: await response.text().catch(() => "") };
|
||||
if (response.ok) {
|
||||
return { data, apiBaseUrl, status: response.status };
|
||||
}
|
||||
const detail = typeof data.detail === "string" ? data.detail : JSON.stringify(data.detail || data);
|
||||
lastError = new Error(detail || `HTTP ${response.status}`);
|
||||
lastError.status = response.status;
|
||||
lastError.detail = detail;
|
||||
lastError.apiBaseUrl = apiBaseUrl;
|
||||
if (response.status !== 404) {
|
||||
throw lastError;
|
||||
}
|
||||
} catch (error) {
|
||||
lastError = error;
|
||||
if (error?.status && error.status !== 404) {
|
||||
throw error;
|
||||
}
|
||||
}
|
||||
}
|
||||
throw lastError || new Error("zncm 服务请求失败。");
|
||||
}
|
||||
|
||||
function zncmMakeThreadId() {
|
||||
if (globalThis.crypto?.randomUUID) {
|
||||
return globalThis.crypto.randomUUID();
|
||||
}
|
||||
return `zncm-${Date.now().toString(36)}-${Math.random().toString(36).slice(2, 10)}`;
|
||||
}
|
||||
|
||||
function zncmBuildPagePrompt(page, content, truncated, followUp) {
|
||||
const title = page.title || "未知标题";
|
||||
const url = page.url || "未知 URL";
|
||||
const capturedAt = new Date().toISOString();
|
||||
if (followUp) {
|
||||
return [
|
||||
"你是 zncm 页面问答助手。请继续基于本会话里已经读取的网页内容回答用户问题。",
|
||||
"如果已有网页内容不足以回答，请说明缺少哪些信息；不要编造，也不要再尝试抓取其它网页。",
|
||||
`当前页面标题：${title}`,
|
||||
`当前页面 URL：${url}`,
|
||||
`本轮提问时间：${capturedAt}`,
|
||||
].join("\n\n");
|
||||
}
|
||||
|
||||
return [
|
||||
"你是 zncm 页面问答助手。请直接基于当前网页内容回答用户问题。",
|
||||
"如果网页内容不足以回答，请说明缺少哪些信息；不要编造，也不要再尝试抓取其它网页。",
|
||||
`页面标题：${title}`,
|
||||
`页面 URL：${url}`,
|
||||
`读取时间：${capturedAt}`,
|
||||
truncated ? "说明：页面内容较长，本次已按插件限制截断。" : "说明：页面内容如下。",
|
||||
"当前网页内容：",
|
||||
page.contentType === "text" ? `\`\`\`text\n${content}\n\`\`\`` : content,
|
||||
].join("\n\n");
|
||||
}
|
||||
|
||||
function zncmExtractLatestAnswer(values) {
|
||||
const messages = Array.isArray(values?.messages) ? values.messages : [];
|
||||
for (let index = messages.length - 1; index >= 0; index -= 1) {
|
||||
const message = messages[index] || {};
|
||||
const type = String(message.type || message.role || "").toLowerCase();
|
||||
if (type !== "ai" && type !== "assistant") {
|
||||
continue;
|
||||
}
|
||||
const content = message.content;
|
||||
if (typeof content === "string" && content.trim()) {
|
||||
return content.trim();
|
||||
}
|
||||
if (Array.isArray(content)) {
|
||||
const text = content
|
||||
.map((part) => {
|
||||
if (typeof part === "string") return part;
|
||||
if (part && typeof part.text === "string") return part.text;
|
||||
return "";
|
||||
})
|
||||
.filter(Boolean)
|
||||
.join("\n")
|
||||
.trim();
|
||||
if (text) {
|
||||
return text;
|
||||
}
|
||||
}
|
||||
}
|
||||
return "";
|
||||
}
|
||||
|
||||
async function askZncmViaRunsWait(settings, payload, page, question) {
|
||||
const contentLimit = 60000;
|
||||
const rawContent = String(page.content || "");
|
||||
const truncated = rawContent.length > contentLimit;
|
||||
const content = truncated ? `${rawContent.slice(0, contentLimit)}\n\n[Truncated locally by zncm Web Clipper]` : rawContent;
|
||||
const threadId = payload.threadId || zncmMakeThreadId();
|
||||
const promptPrefix = zncmBuildPagePrompt(page, content, truncated, Boolean(payload.threadId));
|
||||
const { data } = await zncmFetchJson(settings, "/api/runs/wait", {
|
||||
method: "POST",
|
||||
headers: {
|
||||
"Content-Type": "application/json",
|
||||
...deerAuthHeaders(settings),
|
||||
},
|
||||
body: JSON.stringify({
|
||||
assistant_id: "lead_agent",
|
||||
input: {
|
||||
messages: [
|
||||
{
|
||||
type: "human",
|
||||
content: question,
|
||||
additional_kwargs: {
|
||||
prompt_prefix: promptPrefix,
|
||||
},
|
||||
},
|
||||
],
|
||||
},
|
||||
config: {
|
||||
configurable: {
|
||||
thread_id: threadId,
|
||||
},
|
||||
},
|
||||
metadata: {
|
||||
source: "zncm_page_assistant",
|
||||
browser_context_url: page.url || "",
|
||||
browser_context_title: page.title || "",
|
||||
browser_context_fallback: true,
|
||||
},
|
||||
stream_mode: ["values"],
|
||||
on_disconnect: "continue",
|
||||
multitask_strategy: "reject",
|
||||
}),
|
||||
});
|
||||
return {
|
||||
success: true,
|
||||
thread_id: threadId,
|
||||
run_id: "",
|
||||
answer: zncmExtractLatestAnswer(data) || "zncm 没有返回可显示的答案。",
|
||||
truncated,
|
||||
message: "zncm 页面问答已完成。",
|
||||
};
|
||||
}
|
||||
|
||||
async function askZncmAboutPage(payload) {
|
||||
let settings = await getSettings();
|
||||
settings = await discoverServiceBases(settings);
|
||||
settings = await syncAuthToken(settings);
|
||||
if (!settings.pageAssistantEnabled) {
|
||||
throw new Error("页面问答框已关闭。");
|
||||
}
|
||||
|
||||
const page = payload?.page || {};
|
||||
const question = String(payload?.question || "").trim();
|
||||
if (!question) {
|
||||
throw new Error("请输入要询问的问题。");
|
||||
}
|
||||
if (!page.content) {
|
||||
throw new Error("当前页面没有可读取的正文。");
|
||||
}
|
||||
|
||||
try {
|
||||
const { data } = await zncmFetchJson(settings, "/api/browser-context/ask", {
|
||||
method: "POST",
|
||||
headers: {
|
||||
"Content-Type": "application/json",
|
||||
...deerAuthHeaders(settings),
|
||||
},
|
||||
body: JSON.stringify({
|
||||
thread_id: payload.threadId || null,
|
||||
title: page.title || "",
|
||||
url: page.url || "",
|
||||
content: page.content || "",
|
||||
content_type: page.contentType || "markdown",
|
||||
question,
|
||||
}),
|
||||
});
|
||||
return data;
|
||||
} catch (error) {
|
||||
const detail = String(error?.detail || error?.message || "");
|
||||
if (error?.status === 404 && !/browser context integration is disabled/i.test(detail)) {
|
||||
return askZncmViaRunsWait(settings, payload, page, question);
|
||||
}
|
||||
throw error;
|
||||
}
|
||||
}
|
||||
|
||||
async function getActSessionMap() {
|
||||
try {
|
||||
const stored = await chrome.storage.session.get({ actSessions: {} });
|
||||
return stored.actSessions && typeof stored.actSessions === "object" ? stored.actSessions : {};
|
||||
} catch {
|
||||
return {};
|
||||
}
|
||||
}
|
||||
|
||||
async function setActSessionTab(sessionId, tabId) {
|
||||
if (!sessionId) return;
|
||||
const map = await getActSessionMap();
|
||||
map[sessionId] = tabId;
|
||||
await chrome.storage.session.set({ actSessions: map }).catch(() => {});
|
||||
}
|
||||
|
||||
async function removeActSession(sessionId) {
|
||||
if (!sessionId) return;
|
||||
const map = await getActSessionMap();
|
||||
if (sessionId in map) {
|
||||
delete map[sessionId];
|
||||
await chrome.storage.session.set({ actSessions: map }).catch(() => {});
|
||||
}
|
||||
}
|
||||
|
||||
async function resolveSessionTabId(sessionId) {
|
||||
const map = await getActSessionMap();
|
||||
const tabId = map[sessionId];
|
||||
if (!tabId) return null;
|
||||
try {
|
||||
await chrome.tabs.get(tabId);
|
||||
return tabId;
|
||||
} catch {
|
||||
await removeActSession(sessionId);
|
||||
return null;
|
||||
}
|
||||
}
|
||||
|
||||
async function performActionInTab(tabId, action) {
|
||||
const [result] = await chrome.scripting.executeScript({
|
||||
target: { tabId },
|
||||
func: deerPerformAction,
|
||||
args: [action],
|
||||
});
|
||||
return result?.result || { status: "ok", message: "" };
|
||||
}
|
||||
|
||||
async function observeTab(tabId, maxElements) {
|
||||
const [result] = await chrome.scripting.executeScript({
|
||||
target: { tabId },
|
||||
func: deerObservePage,
|
||||
args: [maxElements],
|
||||
});
|
||||
const observation = result?.result;
|
||||
if (!observation) {
|
||||
throw new Error("当前页面无法读取交互元素。");
|
||||
}
|
||||
return observation;
|
||||
}
|
||||
|
||||
async function runActTask(task, settings) {
|
||||
const sessionId = task.session_id || "";
|
||||
const action = task.action || {};
|
||||
const actionType = String(action.type || "").toLowerCase();
|
||||
const maxElements = 150;
|
||||
|
||||
try {
|
||||
if (!sessionId) {
|
||||
throw new Error("act 任务缺少 session_id。");
|
||||
}
|
||||
|
||||
if (actionType === "close") {
|
||||
const tabId = await resolveSessionTabId(sessionId);
|
||||
if (tabId) {
|
||||
await chrome.tabs.remove(tabId).catch(() => {});
|
||||
}
|
||||
await removeActSession(sessionId);
|
||||
await postTaskResult(task.id, settings, {
|
||||
kind: "act",
|
||||
observation: { action_status: "ok", message: "Tab closed.", url: "", title: "", elements: [], text: "" },
|
||||
metadata: { source: "zncm-web-clipper", worker_id: settings.workerId, session_id: sessionId },
|
||||
});
|
||||
return;
|
||||
}
|
||||
|
||||
let tabId = await resolveSessionTabId(sessionId);
|
||||
|
||||
if (actionType === "navigate") {
|
||||
const url = String(action.url || task.url || "");
|
||||
if (!url) {
|
||||
throw new Error("navigate 动作缺少 url。");
|
||||
}
|
||||
if (tabId) {
|
||||
await chrome.tabs.update(tabId, { url });
|
||||
} else {
|
||||
const tab = await chrome.tabs.create({ url, active: false });
|
||||
tabId = tab.id;
|
||||
await setActSessionTab(sessionId, tabId);
|
||||
}
|
||||
await waitForTabLoad(tabId, settings.taskPageLoadTimeoutSeconds);
|
||||
await delay(800);
|
||||
} else {
|
||||
if (!tabId) {
|
||||
throw new Error("会话标签页不存在，请先用 navigate 打开页面。");
|
||||
}
|
||||
const outcome = await performActionInTab(tabId, action);
|
||||
// Give the page a moment to react / navigate, then settle.
|
||||
await delay(400);
|
||||
await waitForTabLoad(tabId, settings.taskPageLoadTimeoutSeconds);
|
||||
await delay(500);
|
||||
if (outcome.status === "error") {
|
||||
const observation = await observeTab(tabId, maxElements).catch(() => ({ elements: [], text: "", url: "", title: "" }));
|
||||
await postTaskResult(task.id, settings, {
|
||||
kind: "act",
|
||||
observation: { ...observation, action_status: outcome.status, message: outcome.message },
|
||||
metadata: { source: "zncm-web-clipper", worker_id: settings.workerId, session_id: sessionId },
|
||||
});
|
||||
return;
|
||||
}
|
||||
}
|
||||
|
||||
const observation = await observeTab(tabId, maxElements);
|
||||
await postTaskResult(task.id, settings, {
|
||||
kind: "act",
|
||||
title: observation.title || "",
|
||||
url: observation.url || "",
|
||||
observation: { ...observation, action_status: "ok", message: "" },
|
||||
metadata: { source: "zncm-web-clipper", worker_id: settings.workerId, session_id: sessionId },
|
||||
});
|
||||
} catch (error) {
|
||||
await postTaskResult(task.id, settings, {
|
||||
kind: "act",
|
||||
error: error instanceof Error ? error.message : String(error),
|
||||
metadata: { source: "zncm-web-clipper", worker_id: settings.workerId, session_id: sessionId },
|
||||
}).catch((submitError) => console.warn("Failed to submit zncm act task error", submitError));
|
||||
}
|
||||
}
|
||||
|
||||
async function runFetchTask(task, settings) {
|
||||
let tabId = null;
|
||||
try {
|
||||
const tab = await chrome.tabs.create({ url: task.url, active: false });
|
||||
tabId = tab.id;
|
||||
await waitForTabLoad(tabId, settings.taskPageLoadTimeoutSeconds);
|
||||
await delay(750);
|
||||
const page = await captureFromTab(tabId, task.max_content_chars);
|
||||
await postTaskResult(task.id, settings, {
|
||||
kind: "fetch",
|
||||
title: page.title,
|
||||
url: page.url,
|
||||
content: page.content,
|
||||
content_type: page.contentType || "markdown",
|
||||
metadata: {
|
||||
source: "zncm-web-clipper",
|
||||
worker_id: settings.workerId,
|
||||
instruction: task.instruction || "",
|
||||
},
|
||||
});
|
||||
} catch (error) {
|
||||
await postTaskResult(task.id, settings, {
|
||||
kind: "fetch",
|
||||
error: error instanceof Error ? error.message : String(error),
|
||||
metadata: {
|
||||
source: "zncm-web-clipper",
|
||||
worker_id: settings.workerId,
|
||||
},
|
||||
}).catch((submitError) => console.warn("Failed to submit zncm task error", submitError));
|
||||
} finally {
|
||||
if (tabId && settings.closeTaskTabs) {
|
||||
await chrome.tabs.remove(tabId).catch(() => {});
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
async function runTask(task, settings) {
|
||||
if ((task.kind || "fetch") === "act") {
|
||||
await runActTask(task, settings);
|
||||
return;
|
||||
}
|
||||
await runFetchTask(task, settings);
|
||||
}
|
||||
|
||||
async function pollOnce() {
|
||||
if (isPolling) {
|
||||
return { ok: true, message: "任务轮询正在运行。" };
|
||||
}
|
||||
|
||||
isPolling = true;
|
||||
try {
|
||||
let settings = await getSettings();
|
||||
settings = await discoverServiceBases(settings);
|
||||
settings = await syncAuthToken(settings);
|
||||
if (!settings.taskPollingEnabled) {
|
||||
return { ok: false, message: "任务轮询已关闭。" };
|
||||
}
|
||||
|
||||
const apiBaseUrl = deerNormalizeBaseUrl(settings.apiBaseUrl || DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
const url = `${apiBaseUrl}/api/browser-context/tasks/next?worker_id=${encodeURIComponent(settings.workerId)}`;
|
||||
const response = await fetch(url, { headers: deerAuthHeaders(settings) });
|
||||
const data = await response.json().catch(() => ({}));
|
||||
if (!response.ok) {
|
||||
const detail = typeof data.detail === "string" ? data.detail : JSON.stringify(data.detail || data);
|
||||
throw new Error(detail || `HTTP ${response.status}`);
|
||||
}
|
||||
if (!data.task) {
|
||||
return { ok: true, message: "暂无待处理的 zncm 浏览器任务。" };
|
||||
}
|
||||
|
||||
await runTask(data.task, settings);
|
||||
return { ok: true, message: `已完成 zncm 浏览器任务 ${data.task.id}。` };
|
||||
} finally {
|
||||
isPolling = false;
|
||||
await scheduleNextPoll();
|
||||
}
|
||||
}
|
||||
|
||||
async function startup() {
|
||||
let settings = await getSettings();
|
||||
settings = await discoverServiceBases(settings);
|
||||
settings = await syncAuthToken(settings);
|
||||
await refreshAlarm(settings);
|
||||
await scheduleNextPoll(1000);
|
||||
}
|
||||
|
||||
chrome.runtime.onInstalled.addListener(() => void startup());
|
||||
chrome.runtime.onStartup.addListener(() => void startup());
|
||||
chrome.storage.onChanged.addListener((changes, areaName) => {
|
||||
if (areaName === "sync") {
|
||||
void startup();
|
||||
}
|
||||
});
|
||||
chrome.alarms.onAlarm.addListener((alarm) => {
|
||||
if (alarm.name === ALARM_NAME) {
|
||||
void pollOnce().catch((error) => console.warn("zncm alarm poll failed", error));
|
||||
}
|
||||
});
|
||||
chrome.runtime.onMessage.addListener((message, sender, sendResponse) => {
|
||||
if (message?.type === "zncm-discover-service") {
|
||||
getSettings()
|
||||
.then((settings) => discoverServiceBases(settings, true))
|
||||
.then((settings) =>
|
||||
sendResponse({
|
||||
ok: Boolean(settings.apiBaseUrl && settings.webBaseUrl),
|
||||
message: `服务地址：${settings.webBaseUrl}；API：${settings.apiBaseUrl}`,
|
||||
apiBaseUrl: settings.apiBaseUrl,
|
||||
webBaseUrl: settings.webBaseUrl,
|
||||
source: settings.lastServiceSyncSource || "",
|
||||
}),
|
||||
)
|
||||
.catch((error) => sendResponse({ ok: false, message: error instanceof Error ? error.message : String(error) }));
|
||||
return true;
|
||||
}
|
||||
|
||||
if (message?.type === "zncm-sync-token") {
|
||||
syncAuthToken(null, true)
|
||||
.then((settings) => {
|
||||
const hasToken = Boolean(settings.accessToken);
|
||||
sendResponse({
|
||||
ok: hasToken,
|
||||
message: hasToken
|
||||
? `Token 已从 ${settings.lastTokenSyncSource || "zncm"} 同步。`
|
||||
: "没有找到 zncm token。请打开已登录的 zncm 页面后重试。",
|
||||
source: settings.lastTokenSyncSource || "",
|
||||
});
|
||||
})
|
||||
.catch((error) => sendResponse({ ok: false, message: error instanceof Error ? error.message : String(error) }));
|
||||
return true;
|
||||
}
|
||||
|
||||
if (message?.type === "zncm-page-agent-ask") {
|
||||
askZncmAboutPage(message)
|
||||
.then((result) => sendResponse({ ok: true, ...result }))
|
||||
.catch((error) => sendResponse({ ok: false, message: error instanceof Error ? error.message : String(error) }));
|
||||
return true;
|
||||
}
|
||||
|
||||
if (message?.type === "zncm-open-options") {
|
||||
chrome.runtime.openOptionsPage()
|
||||
.then(() => sendResponse({ ok: true }))
|
||||
.catch((error) => sendResponse({ ok: false, message: error instanceof Error ? error.message : String(error) }));
|
||||
return true;
|
||||
}
|
||||
|
||||
if (message?.type !== "zncm-poll-now") {
|
||||
return false;
|
||||
}
|
||||
pollOnce()
|
||||
.then((result) => sendResponse(result))
|
||||
.catch((error) => sendResponse({ ok: false, message: error instanceof Error ? error.message : String(error) }));
|
||||
return true;
|
||||
});
|
||||
|
||||
void startup();
|
||||
@@ -1,9 +0,0 @@
|
||||
// Default deployment settings for zncm Web Clipper.
|
||||
//
|
||||
// Change these two values before loading or packaging the extension when the
|
||||
// zncm service address changes. Users can still override them in the extension
|
||||
// settings page after installation.
|
||||
globalThis.ZNCM_WEB_CLIPPER_CONFIG = Object.freeze({
|
||||
defaultWebBaseUrl: "http://47.94.209.59:2026",
|
||||
defaultApiBaseUrl: "http://47.94.209.59:2026",
|
||||
});
|
||||
@@ -1,737 +0,0 @@
|
||||
(() => {
|
||||
if (window.top !== window || window.__zncmPageAssistantLoaded) {
|
||||
return;
|
||||
}
|
||||
window.__zncmPageAssistantLoaded = true;
|
||||
|
||||
const HOST_ID = "zncm-page-assistant-host";
|
||||
const state = {
|
||||
enabled: true,
|
||||
open: false,
|
||||
busy: false,
|
||||
threadId: "",
|
||||
pageUrl: location.href,
|
||||
messages: [],
|
||||
};
|
||||
|
||||
let host = null;
|
||||
let shadow = null;
|
||||
let ui = {};
|
||||
|
||||
function sendRuntimeMessage(message) {
|
||||
return new Promise((resolve, reject) => {
|
||||
chrome.runtime.sendMessage(message, (response) => {
|
||||
const error = chrome.runtime.lastError;
|
||||
if (error) {
|
||||
reject(new Error(error.message));
|
||||
return;
|
||||
}
|
||||
resolve(response || {});
|
||||
});
|
||||
});
|
||||
}
|
||||
|
||||
function makeElement(tag, className = "", text = "") {
|
||||
const element = document.createElement(tag);
|
||||
if (className) {
|
||||
element.className = className;
|
||||
}
|
||||
if (text) {
|
||||
element.textContent = text;
|
||||
}
|
||||
return element;
|
||||
}
|
||||
|
||||
function setStatus(message) {
|
||||
if (ui.status) {
|
||||
ui.status.textContent = message || "";
|
||||
}
|
||||
}
|
||||
|
||||
function setBusy(busy) {
|
||||
state.busy = busy;
|
||||
if (ui.send) ui.send.disabled = busy;
|
||||
if (ui.input) ui.input.disabled = busy;
|
||||
if (ui.panel) ui.panel.classList.toggle("is-busy", busy);
|
||||
}
|
||||
|
||||
function resetThreadIfPageChanged() {
|
||||
if (state.pageUrl === location.href) {
|
||||
return;
|
||||
}
|
||||
state.pageUrl = location.href;
|
||||
state.threadId = "";
|
||||
state.messages = [];
|
||||
renderMessages();
|
||||
setStatus("页面已变化，下一次提问会重新读取当前页面。");
|
||||
}
|
||||
|
||||
function destroyWidget() {
|
||||
if (host) {
|
||||
host.remove();
|
||||
}
|
||||
host = null;
|
||||
shadow = null;
|
||||
ui = {};
|
||||
}
|
||||
|
||||
function renderMessages() {
|
||||
if (!ui.messages) {
|
||||
return;
|
||||
}
|
||||
ui.messages.replaceChildren();
|
||||
if (state.messages.length === 0) {
|
||||
const empty = makeElement("div", "empty");
|
||||
const emptyTitle = makeElement("strong", "", "问问当前页面");
|
||||
const emptyText = makeElement("span", "", "可以提问页面里的数据、表格、段落和结论。");
|
||||
empty.append(emptyTitle, emptyText);
|
||||
ui.messages.appendChild(empty);
|
||||
return;
|
||||
}
|
||||
|
||||
for (const message of state.messages) {
|
||||
const item = makeElement("div", `message ${message.role}`);
|
||||
const avatar = makeElement("div", "avatar", message.role === "user" ? "你" : "zn");
|
||||
const bubble = makeElement("div", "bubble");
|
||||
const label = makeElement("div", "message-label", message.role === "user" ? "你" : "zncm");
|
||||
const body = makeElement("div", "message-body", message.text);
|
||||
if (message.pending) {
|
||||
body.classList.add("pending");
|
||||
}
|
||||
bubble.append(label, body);
|
||||
item.append(avatar, bubble);
|
||||
ui.messages.appendChild(item);
|
||||
}
|
||||
ui.messages.scrollTop = ui.messages.scrollHeight;
|
||||
}
|
||||
|
||||
function setOpen(open) {
|
||||
state.open = open;
|
||||
if (!ui.panel || !ui.launcher) {
|
||||
return;
|
||||
}
|
||||
ui.panel.classList.toggle("open", open);
|
||||
ui.launcher.classList.toggle("open", open);
|
||||
ui.launcher.setAttribute("aria-expanded", String(open));
|
||||
if (open) {
|
||||
resetThreadIfPageChanged();
|
||||
setTimeout(() => ui.input?.focus(), 50);
|
||||
}
|
||||
}
|
||||
|
||||
async function ask(question) {
|
||||
const normalized = String(question || "").trim();
|
||||
if (!normalized || state.busy) {
|
||||
return;
|
||||
}
|
||||
|
||||
resetThreadIfPageChanged();
|
||||
setBusy(true);
|
||||
setStatus("正在读取页面并发送给 zncm...");
|
||||
state.messages.push({ role: "user", text: normalized });
|
||||
const pending = { role: "assistant", text: "zncm 正在理解当前页面...", pending: true };
|
||||
state.messages.push(pending);
|
||||
renderMessages();
|
||||
ui.input.value = "";
|
||||
|
||||
try {
|
||||
if (typeof deerExtractCurrentPage !== "function") {
|
||||
throw new Error("页面读取脚本未加载，请刷新当前页面后重试。");
|
||||
}
|
||||
const page = deerExtractCurrentPage();
|
||||
const response = await sendRuntimeMessage({
|
||||
type: "zncm-page-agent-ask",
|
||||
page,
|
||||
question: normalized,
|
||||
threadId: state.threadId || "",
|
||||
});
|
||||
if (!response.ok) {
|
||||
throw new Error(response.message || "zncm 没有返回答案。");
|
||||
}
|
||||
state.threadId = response.thread_id || state.threadId;
|
||||
pending.text = response.answer || "zncm 没有返回答案。";
|
||||
pending.pending = false;
|
||||
setStatus(response.truncated ? "已回答。页面内容较长，本次按配置做了截断。" : "已回答。");
|
||||
} catch (error) {
|
||||
pending.text = error instanceof Error ? error.message : "页面问答失败。";
|
||||
pending.pending = false;
|
||||
setStatus("本次问答失败。");
|
||||
} finally {
|
||||
setBusy(false);
|
||||
renderMessages();
|
||||
ui.input.focus();
|
||||
}
|
||||
}
|
||||
|
||||
function buildWidget() {
|
||||
host = document.createElement("div");
|
||||
host.id = HOST_ID;
|
||||
document.documentElement.appendChild(host);
|
||||
shadow = host.attachShadow({ mode: "open" });
|
||||
|
||||
const style = document.createElement("style");
|
||||
style.textContent = `
|
||||
:host {
|
||||
all: initial;
|
||||
color-scheme: light;
|
||||
font-family: Inter, ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif;
|
||||
}
|
||||
|
||||
.launcher,
|
||||
.panel {
|
||||
position: fixed;
|
||||
z-index: 2147483647;
|
||||
font: 14px/1.45 Inter, ui-sans-serif, system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif;
|
||||
}
|
||||
|
||||
.launcher {
|
||||
right: 20px;
|
||||
bottom: 20px;
|
||||
display: grid;
|
||||
grid-template-columns: auto 1fr;
|
||||
align-items: center;
|
||||
gap: 8px;
|
||||
height: 44px;
|
||||
min-width: 112px;
|
||||
border: 1px solid rgba(15, 118, 110, 0.28);
|
||||
border-radius: 8px;
|
||||
padding: 0 12px 0 10px;
|
||||
background: #0f766e;
|
||||
color: #fff;
|
||||
box-shadow: 0 12px 34px rgba(15, 118, 110, 0.28), 0 2px 8px rgba(17, 24, 39, 0.12);
|
||||
cursor: pointer;
|
||||
font-weight: 750;
|
||||
transition: transform 160ms ease, box-shadow 160ms ease, background 160ms ease;
|
||||
}
|
||||
|
||||
.launcher:hover {
|
||||
transform: translateY(-2px);
|
||||
box-shadow: 0 16px 40px rgba(15, 118, 110, 0.32), 0 4px 12px rgba(17, 24, 39, 0.14);
|
||||
}
|
||||
|
||||
.launcher-mark {
|
||||
display: grid;
|
||||
place-items: center;
|
||||
width: 24px;
|
||||
height: 24px;
|
||||
border-radius: 7px;
|
||||
background: rgba(255, 255, 255, 0.18);
|
||||
font-size: 12px;
|
||||
letter-spacing: 0;
|
||||
box-shadow: inset 0 0 0 1px rgba(255, 255, 255, 0.2);
|
||||
}
|
||||
|
||||
.launcher.open {
|
||||
background: #115e59;
|
||||
}
|
||||
|
||||
.panel {
|
||||
right: 20px;
|
||||
bottom: 72px;
|
||||
width: min(430px, calc(100vw - 24px));
|
||||
max-height: min(640px, calc(100vh - 96px));
|
||||
display: none;
|
||||
grid-template-rows: auto auto minmax(180px, 1fr) auto;
|
||||
overflow: hidden;
|
||||
border: 1px solid #d6d8dc;
|
||||
border-radius: 8px;
|
||||
background: #ffffff;
|
||||
color: #171717;
|
||||
box-shadow: 0 24px 70px rgba(17, 24, 39, 0.2), 0 1px 2px rgba(17, 24, 39, 0.08);
|
||||
}
|
||||
|
||||
.panel.open {
|
||||
display: grid;
|
||||
animation: zncm-panel-in 160ms ease-out;
|
||||
}
|
||||
|
||||
@keyframes zncm-panel-in {
|
||||
from {
|
||||
opacity: 0;
|
||||
transform: translateY(8px) scale(0.985);
|
||||
}
|
||||
to {
|
||||
opacity: 1;
|
||||
transform: translateY(0) scale(1);
|
||||
}
|
||||
}
|
||||
|
||||
.header {
|
||||
display: grid;
|
||||
gap: 10px;
|
||||
padding: 14px;
|
||||
background: linear-gradient(180deg, #f4fffc 0%, #ffffff 84%);
|
||||
border-bottom: 1px solid #e5e7eb;
|
||||
}
|
||||
|
||||
.header-top {
|
||||
display: flex;
|
||||
align-items: center;
|
||||
justify-content: space-between;
|
||||
gap: 12px;
|
||||
}
|
||||
|
||||
.brand {
|
||||
display: flex;
|
||||
align-items: center;
|
||||
min-width: 0;
|
||||
gap: 10px;
|
||||
}
|
||||
|
||||
.brand-mark {
|
||||
display: grid;
|
||||
place-items: center;
|
||||
width: 34px;
|
||||
height: 34px;
|
||||
flex: 0 0 auto;
|
||||
border-radius: 8px;
|
||||
background: #0f766e;
|
||||
color: #fff;
|
||||
font-weight: 800;
|
||||
box-shadow: 0 8px 20px rgba(15, 118, 110, 0.18);
|
||||
}
|
||||
|
||||
.brand-copy {
|
||||
display: grid;
|
||||
min-width: 0;
|
||||
gap: 1px;
|
||||
}
|
||||
|
||||
.brand-copy strong {
|
||||
color: #111827;
|
||||
font-size: 15px;
|
||||
line-height: 1.25;
|
||||
}
|
||||
|
||||
.brand-copy span {
|
||||
overflow: hidden;
|
||||
color: #6b7280;
|
||||
font-size: 12px;
|
||||
text-overflow: ellipsis;
|
||||
white-space: nowrap;
|
||||
}
|
||||
|
||||
.header-actions {
|
||||
display: flex;
|
||||
flex: 0 0 auto;
|
||||
gap: 6px;
|
||||
}
|
||||
|
||||
.icon-button {
|
||||
display: grid;
|
||||
place-items: center;
|
||||
width: auto;
|
||||
min-width: 34px;
|
||||
height: 30px;
|
||||
padding: 0 8px;
|
||||
border: 1px solid #dfe3e8;
|
||||
border-radius: 7px;
|
||||
background: #fff;
|
||||
color: #374151;
|
||||
cursor: pointer;
|
||||
font: inherit;
|
||||
font-size: 12px;
|
||||
white-space: nowrap;
|
||||
transition: background 140ms ease, border-color 140ms ease, color 140ms ease;
|
||||
}
|
||||
|
||||
.icon-button:hover {
|
||||
border-color: #0f766e;
|
||||
color: #0f766e;
|
||||
background: #f0fdfa;
|
||||
}
|
||||
|
||||
.page-chip {
|
||||
display: grid;
|
||||
grid-template-columns: auto 1fr;
|
||||
align-items: center;
|
||||
gap: 8px;
|
||||
min-width: 0;
|
||||
border: 1px solid #dbeafe;
|
||||
border-radius: 8px;
|
||||
padding: 8px 10px;
|
||||
background: #f8fbff;
|
||||
color: #334155;
|
||||
box-shadow: inset 0 1px 0 rgba(255, 255, 255, 0.8);
|
||||
}
|
||||
|
||||
.page-chip-dot {
|
||||
width: 8px;
|
||||
height: 8px;
|
||||
border-radius: 999px;
|
||||
background: #0f766e;
|
||||
}
|
||||
|
||||
.page-chip-text {
|
||||
overflow: hidden;
|
||||
font-size: 12px;
|
||||
text-overflow: ellipsis;
|
||||
white-space: nowrap;
|
||||
}
|
||||
|
||||
.status {
|
||||
min-height: 20px;
|
||||
padding: 9px 14px;
|
||||
border-bottom: 1px solid #f0f2f5;
|
||||
color: #475569;
|
||||
background: #fbfcfd;
|
||||
font-size: 12px;
|
||||
}
|
||||
|
||||
.messages {
|
||||
display: grid;
|
||||
align-content: start;
|
||||
gap: 12px;
|
||||
overflow: auto;
|
||||
padding: 14px;
|
||||
background: linear-gradient(180deg, #f7f8fa 0%, #f3f6f8 100%);
|
||||
}
|
||||
|
||||
.empty {
|
||||
display: grid;
|
||||
gap: 5px;
|
||||
padding: 16px;
|
||||
border: 1px solid #dfe7ed;
|
||||
border-radius: 8px;
|
||||
background: linear-gradient(180deg, #ffffff 0%, #fbfefd 100%);
|
||||
color: #4b5563;
|
||||
box-shadow: 0 8px 20px rgba(15, 23, 42, 0.04);
|
||||
}
|
||||
|
||||
.empty strong {
|
||||
color: #111827;
|
||||
font-size: 14px;
|
||||
}
|
||||
|
||||
.empty span {
|
||||
font-size: 13px;
|
||||
}
|
||||
|
||||
.message {
|
||||
display: grid;
|
||||
grid-template-columns: 28px minmax(0, 1fr);
|
||||
gap: 8px;
|
||||
align-items: start;
|
||||
}
|
||||
|
||||
.message.user {
|
||||
grid-template-columns: minmax(0, 1fr) 28px;
|
||||
}
|
||||
|
||||
.message.user .avatar {
|
||||
grid-column: 2;
|
||||
}
|
||||
|
||||
.message.user .bubble {
|
||||
grid-column: 1;
|
||||
justify-self: end;
|
||||
}
|
||||
|
||||
.avatar {
|
||||
display: grid;
|
||||
place-items: center;
|
||||
width: 28px;
|
||||
height: 28px;
|
||||
border-radius: 8px;
|
||||
background: #e6fffb;
|
||||
color: #0f766e;
|
||||
font-size: 11px;
|
||||
font-weight: 800;
|
||||
}
|
||||
|
||||
.message.user .avatar {
|
||||
background: #111827;
|
||||
color: #fff;
|
||||
}
|
||||
|
||||
.bubble {
|
||||
display: grid;
|
||||
gap: 4px;
|
||||
max-width: 100%;
|
||||
}
|
||||
|
||||
.message-label {
|
||||
color: #6b7280;
|
||||
font-size: 11px;
|
||||
}
|
||||
|
||||
.message.user .message-label {
|
||||
text-align: right;
|
||||
}
|
||||
|
||||
.message-body {
|
||||
max-width: 330px;
|
||||
padding: 10px 12px;
|
||||
border: 1px solid #e5e7eb;
|
||||
border-radius: 8px;
|
||||
background: #fff;
|
||||
color: #171717;
|
||||
white-space: pre-wrap;
|
||||
overflow-wrap: anywhere;
|
||||
box-shadow: 0 1px 2px rgba(15, 23, 42, 0.04);
|
||||
}
|
||||
|
||||
.message.user .message-body {
|
||||
border-color: #0f172a;
|
||||
background: #0f172a;
|
||||
color: #fff;
|
||||
}
|
||||
|
||||
.message-body.pending {
|
||||
color: #6b7280;
|
||||
}
|
||||
|
||||
.message-body.pending::after {
|
||||
content: "";
|
||||
display: inline-block;
|
||||
width: 18px;
|
||||
height: 4px;
|
||||
margin-left: 6px;
|
||||
border-radius: 999px;
|
||||
background: linear-gradient(90deg, #0f766e 0 33%, transparent 33% 100%);
|
||||
background-size: 9px 4px;
|
||||
animation: zncm-thinking 900ms linear infinite;
|
||||
vertical-align: middle;
|
||||
}
|
||||
|
||||
@keyframes zncm-thinking {
|
||||
from {
|
||||
background-position: 0 0;
|
||||
}
|
||||
to {
|
||||
background-position: 18px 0;
|
||||
}
|
||||
}
|
||||
|
||||
.composer-wrap {
|
||||
display: grid;
|
||||
gap: 10px;
|
||||
padding: 12px 14px 14px;
|
||||
background: #fff;
|
||||
border-top: 1px solid #e5e7eb;
|
||||
}
|
||||
|
||||
.quick {
|
||||
display: grid;
|
||||
grid-template-columns: repeat(2, minmax(0, 1fr));
|
||||
gap: 7px;
|
||||
}
|
||||
|
||||
.quick button {
|
||||
border: 1px solid #d6d8dc;
|
||||
border-radius: 8px;
|
||||
padding: 8px 10px;
|
||||
background: #fbfcfd;
|
||||
color: #374151;
|
||||
cursor: pointer;
|
||||
font: inherit;
|
||||
font-size: 12px;
|
||||
text-align: left;
|
||||
transition: background 140ms ease, border-color 140ms ease, color 140ms ease, transform 140ms ease;
|
||||
}
|
||||
|
||||
.quick button:hover {
|
||||
border-color: #0f766e;
|
||||
color: #0f766e;
|
||||
background: #f0fdfa;
|
||||
transform: translateY(-1px);
|
||||
}
|
||||
|
||||
.composer {
|
||||
display: grid;
|
||||
grid-template-columns: 1fr auto;
|
||||
gap: 8px;
|
||||
align-items: end;
|
||||
}
|
||||
|
||||
.composer textarea {
|
||||
min-height: 46px;
|
||||
max-height: 130px;
|
||||
resize: vertical;
|
||||
border: 1px solid #cfd6dd;
|
||||
border-radius: 8px;
|
||||
padding: 10px 11px;
|
||||
color: #171717;
|
||||
background: #fbfcfd;
|
||||
font: inherit;
|
||||
outline: none;
|
||||
}
|
||||
|
||||
.composer textarea:focus {
|
||||
border-color: #0f766e;
|
||||
box-shadow: 0 0 0 3px rgba(15, 118, 110, 0.12);
|
||||
}
|
||||
|
||||
.composer button {
|
||||
height: 42px;
|
||||
border: 0;
|
||||
border-radius: 8px;
|
||||
padding: 0 14px;
|
||||
background: #0f766e;
|
||||
color: #fff;
|
||||
cursor: pointer;
|
||||
font: inherit;
|
||||
font-weight: 750;
|
||||
box-shadow: 0 8px 18px rgba(15, 118, 110, 0.2);
|
||||
transition: transform 140ms ease, background 140ms ease, box-shadow 140ms ease;
|
||||
}
|
||||
|
||||
.composer button:hover {
|
||||
background: #115e59;
|
||||
transform: translateY(-1px);
|
||||
box-shadow: 0 10px 22px rgba(15, 118, 110, 0.24);
|
||||
}
|
||||
|
||||
.panel.is-busy .brand-mark,
|
||||
.panel.is-busy .launcher-mark {
|
||||
opacity: 0.82;
|
||||
}
|
||||
|
||||
.composer button:disabled,
|
||||
.quick button:disabled,
|
||||
.icon-button:disabled {
|
||||
cursor: not-allowed;
|
||||
opacity: 0.6;
|
||||
}
|
||||
|
||||
@media (max-width: 480px) {
|
||||
.launcher {
|
||||
right: 12px;
|
||||
bottom: 12px;
|
||||
min-width: 96px;
|
||||
}
|
||||
|
||||
.panel {
|
||||
right: 12px;
|
||||
bottom: 64px;
|
||||
width: calc(100vw - 24px);
|
||||
max-height: calc(100vh - 80px);
|
||||
}
|
||||
|
||||
.message-body {
|
||||
max-width: calc(100vw - 104px);
|
||||
}
|
||||
|
||||
.quick {
|
||||
grid-template-columns: 1fr;
|
||||
}
|
||||
}
|
||||
`;
|
||||
|
||||
const launcher = makeElement("button", "launcher");
|
||||
launcher.type = "button";
|
||||
launcher.title = "打开 zncm 页面问答";
|
||||
launcher.setAttribute("aria-expanded", "false");
|
||||
launcher.append(makeElement("span", "launcher-mark", "zn"), makeElement("span", "", "页面问答"));
|
||||
|
||||
const panel = makeElement("section", "panel");
|
||||
const header = makeElement("header", "header");
|
||||
const headerTop = makeElement("div", "header-top");
|
||||
const brand = makeElement("div", "brand");
|
||||
const brandCopy = makeElement("div", "brand-copy");
|
||||
const title = makeElement("strong", "", "zncm 页面问答");
|
||||
const subtitle = makeElement("span", "", "读取当前页面并回答");
|
||||
brandCopy.append(title, subtitle);
|
||||
brand.append(makeElement("div", "brand-mark", "zn"), brandCopy);
|
||||
|
||||
const actions = makeElement("div", "header-actions");
|
||||
const settings = makeElement("button", "icon-button", "设置");
|
||||
settings.type = "button";
|
||||
settings.title = "打开插件设置";
|
||||
const minimize = makeElement("button", "icon-button", "－");
|
||||
minimize.type = "button";
|
||||
minimize.title = "收起";
|
||||
actions.append(settings, minimize);
|
||||
headerTop.append(brand, actions);
|
||||
|
||||
const pageChip = makeElement("div", "page-chip");
|
||||
const pageTitle = makeElement("span", "page-chip-text", document.title || location.hostname);
|
||||
pageChip.append(makeElement("span", "page-chip-dot"), pageTitle);
|
||||
header.append(headerTop, pageChip);
|
||||
|
||||
const status = makeElement("div", "status", "已连接当前页面。");
|
||||
const messages = makeElement("div", "messages");
|
||||
|
||||
const composerWrap = makeElement("div", "composer-wrap");
|
||||
const quick = makeElement("div", "quick");
|
||||
for (const text of ["总结此页", "提取关键数据", "列出待办", "解释表格"]) {
|
||||
const chip = makeElement("button", "", text);
|
||||
chip.type = "button";
|
||||
chip.addEventListener("click", () => ask(text));
|
||||
quick.appendChild(chip);
|
||||
}
|
||||
|
||||
const composer = makeElement("form", "composer");
|
||||
const input = makeElement("textarea");
|
||||
input.placeholder = "问当前页面...";
|
||||
input.rows = 2;
|
||||
const send = makeElement("button", "", "发送");
|
||||
send.type = "submit";
|
||||
composer.append(input, send);
|
||||
composerWrap.append(quick, composer);
|
||||
|
||||
panel.append(header, status, messages, composerWrap);
|
||||
shadow.append(style, panel, launcher);
|
||||
|
||||
ui = { launcher, panel, subtitle, pageTitle, status, messages, input, send };
|
||||
renderMessages();
|
||||
|
||||
launcher.addEventListener("click", () => setOpen(!state.open));
|
||||
minimize.addEventListener("click", () => setOpen(false));
|
||||
settings.addEventListener("click", () => {
|
||||
sendRuntimeMessage({ type: "zncm-open-options" }).catch((error) => {
|
||||
setStatus(error instanceof Error ? error.message : "无法打开插件设置。");
|
||||
});
|
||||
});
|
||||
composer.addEventListener("submit", (event) => {
|
||||
event.preventDefault();
|
||||
ask(input.value);
|
||||
});
|
||||
input.addEventListener("keydown", (event) => {
|
||||
if (event.key === "Enter" && !event.shiftKey) {
|
||||
event.preventDefault();
|
||||
ask(input.value);
|
||||
}
|
||||
});
|
||||
}
|
||||
|
||||
function ensureWidget() {
|
||||
if (!state.enabled) {
|
||||
destroyWidget();
|
||||
return;
|
||||
}
|
||||
if (document.getElementById(HOST_ID)) {
|
||||
return;
|
||||
}
|
||||
buildWidget();
|
||||
}
|
||||
|
||||
async function loadSettings() {
|
||||
const settings = await chrome.storage.sync.get(DEER_CLIPPER_DEFAULTS);
|
||||
state.enabled = Boolean(settings.pageAssistantEnabled);
|
||||
ensureWidget();
|
||||
}
|
||||
|
||||
chrome.storage.onChanged.addListener((changes, areaName) => {
|
||||
if (areaName !== "sync" || !changes.pageAssistantEnabled) {
|
||||
return;
|
||||
}
|
||||
state.enabled = Boolean(changes.pageAssistantEnabled.newValue);
|
||||
ensureWidget();
|
||||
if (state.enabled) {
|
||||
setStatus("页面问答框已开启。");
|
||||
}
|
||||
});
|
||||
|
||||
setInterval(() => {
|
||||
if (!state.enabled || !ui.pageTitle) {
|
||||
return;
|
||||
}
|
||||
if (state.pageUrl !== location.href) {
|
||||
resetThreadIfPageChanged();
|
||||
}
|
||||
ui.pageTitle.textContent = document.title || location.hostname;
|
||||
}, 1200);
|
||||
|
||||
void loadSettings();
|
||||
})();
|
||||
@@ -1,24 +0,0 @@
|
||||
{
|
||||
"manifest_version": 3,
|
||||
"name": "zncm Web Clipper",
|
||||
"version": "0.2.0",
|
||||
"description": "Send pages to zncm and execute browser-backed fetch tasks.",
|
||||
"permissions": ["activeTab", "alarms", "cookies", "scripting", "storage", "tabs"],
|
||||
"host_permissions": ["<all_urls>"],
|
||||
"background": {
|
||||
"service_worker": "background.js"
|
||||
},
|
||||
"action": {
|
||||
"default_title": "zncm Web Clipper",
|
||||
"default_popup": "popup.html"
|
||||
},
|
||||
"options_page": "options.html",
|
||||
"content_scripts": [
|
||||
{
|
||||
"matches": ["http://*/*", "https://*/*"],
|
||||
"js": ["config.js", "shared.js", "content.js"],
|
||||
"run_at": "document_idle",
|
||||
"all_frames": false
|
||||
}
|
||||
]
|
||||
}
|
||||
@@ -1,74 +0,0 @@
|
||||
<!doctype html>
|
||||
<html lang="zh-CN">
|
||||
<head>
|
||||
<meta charset="utf-8" />
|
||||
<title>zncm Web Clipper Settings</title>
|
||||
<link rel="stylesheet" href="styles.css" />
|
||||
</head>
|
||||
<body>
|
||||
<main class="options">
|
||||
<h1>zncm Web Clipper Settings</h1>
|
||||
<label>
|
||||
API 地址
|
||||
<input id="apiBaseUrl" />
|
||||
</label>
|
||||
<label>
|
||||
前端地址
|
||||
<input id="webBaseUrl" />
|
||||
</label>
|
||||
<label class="checkbox">
|
||||
<input id="autoDiscoverServiceUrl" type="checkbox" />
|
||||
自动识别已打开 zncm 页面的服务地址
|
||||
</label>
|
||||
<div class="actions">
|
||||
<button id="discoverService" type="button">重新识别服务地址</button>
|
||||
</div>
|
||||
<p id="serviceStatus" class="status"></p>
|
||||
<label>
|
||||
Access Token
|
||||
<textarea id="accessToken" rows="5" placeholder="会自动从已登录 zncm 页面或 access_token Cookie 获取。也可以手动粘贴。"></textarea>
|
||||
</label>
|
||||
<label class="checkbox">
|
||||
<input id="tokenSyncEnabled" type="checkbox" />
|
||||
自动同步已登录 zncm 的 token
|
||||
</label>
|
||||
<div class="actions">
|
||||
<button id="syncToken" type="button">Sync token now</button>
|
||||
</div>
|
||||
<p id="tokenStatus" class="status"></p>
|
||||
<label class="checkbox">
|
||||
<input id="pageAssistantEnabled" type="checkbox" />
|
||||
启用页面问答框
|
||||
</label>
|
||||
<label>
|
||||
默认首问
|
||||
<textarea id="defaultPrompt" rows="4"></textarea>
|
||||
</label>
|
||||
<label class="checkbox">
|
||||
<input id="taskPollingEnabled" type="checkbox" />
|
||||
启用 zncm 聊天触发的浏览器抓取任务
|
||||
</label>
|
||||
<label>
|
||||
Worker ID
|
||||
<input id="workerId" placeholder="自动生成" />
|
||||
</label>
|
||||
<label>
|
||||
任务轮询间隔秒数
|
||||
<input id="taskPollIntervalSeconds" type="number" min="3" step="1" />
|
||||
</label>
|
||||
<label>
|
||||
页面加载超时秒数
|
||||
<input id="taskPageLoadTimeoutSeconds" type="number" min="5" step="1" />
|
||||
</label>
|
||||
<label class="checkbox">
|
||||
<input id="closeTaskTabs" type="checkbox" />
|
||||
抓取完成后关闭任务标签页
|
||||
</label>
|
||||
<button id="save" type="button">保存</button>
|
||||
<p id="status" class="status"></p>
|
||||
</main>
|
||||
<script src="config.js"></script>
|
||||
<script src="shared.js"></script>
|
||||
<script src="options.js"></script>
|
||||
</body>
|
||||
</html>
|
||||
@@ -1,148 +0,0 @@
|
||||
const textFields = [
|
||||
"apiBaseUrl",
|
||||
"webBaseUrl",
|
||||
"accessToken",
|
||||
"defaultPrompt",
|
||||
"workerId",
|
||||
"taskPollIntervalSeconds",
|
||||
"taskPageLoadTimeoutSeconds",
|
||||
];
|
||||
const checkboxFields = [
|
||||
"autoDiscoverServiceUrl",
|
||||
"tokenSyncEnabled",
|
||||
"pageAssistantEnabled",
|
||||
"taskPollingEnabled",
|
||||
"closeTaskTabs",
|
||||
];
|
||||
|
||||
function formatServiceStatus(settings) {
|
||||
if (!settings.lastServiceSyncAt) {
|
||||
return `当前服务地址：${settings.webBaseUrl || DEER_CLIPPER_DEFAULTS.webBaseUrl}；API：${settings.apiBaseUrl || DEER_CLIPPER_DEFAULTS.apiBaseUrl}`;
|
||||
}
|
||||
return `当前服务地址：${settings.webBaseUrl || ""}；API：${settings.apiBaseUrl || ""}\n上次识别：${settings.lastServiceSyncAt}（${settings.lastServiceSyncSource || "unknown source"}）`;
|
||||
}
|
||||
|
||||
function formatTokenStatus(settings) {
|
||||
if (!settings.lastTokenSyncAt) {
|
||||
return "Token 尚未自动同步。";
|
||||
}
|
||||
return `上次 token 同步：${settings.lastTokenSyncAt}（${settings.lastTokenSyncSource || "unknown source"}）`;
|
||||
}
|
||||
|
||||
async function loadSettings() {
|
||||
const settings = await chrome.storage.sync.get(DEER_CLIPPER_DEFAULTS);
|
||||
if (!settings.workerId) {
|
||||
settings.workerId = deerMakeWorkerId();
|
||||
await chrome.storage.sync.set({ workerId: settings.workerId });
|
||||
}
|
||||
document.getElementById("apiBaseUrl").placeholder =
|
||||
DEER_CLIPPER_DEFAULTS.apiBaseUrl || "在 config.js 中配置默认 API 地址";
|
||||
document.getElementById("webBaseUrl").placeholder =
|
||||
DEER_CLIPPER_DEFAULTS.webBaseUrl || "在 config.js 中配置默认前端地址";
|
||||
for (const field of textFields) {
|
||||
document.getElementById(field).value = settings[field] ?? "";
|
||||
}
|
||||
for (const field of checkboxFields) {
|
||||
document.getElementById(field).checked = Boolean(settings[field]);
|
||||
}
|
||||
document.getElementById("serviceStatus").textContent = formatServiceStatus(settings);
|
||||
document.getElementById("tokenStatus").textContent = formatTokenStatus(settings);
|
||||
}
|
||||
|
||||
async function saveSettings() {
|
||||
const next = {};
|
||||
for (const field of textFields) {
|
||||
next[field] = document.getElementById(field).value.trim();
|
||||
}
|
||||
for (const field of checkboxFields) {
|
||||
next[field] = document.getElementById(field).checked;
|
||||
}
|
||||
next.apiBaseUrl = deerNormalizeBaseUrl(next.apiBaseUrl || DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
next.webBaseUrl = deerNormalizeBaseUrl(next.webBaseUrl || DEER_CLIPPER_DEFAULTS.webBaseUrl);
|
||||
next.defaultPrompt = next.defaultPrompt || DEER_CLIPPER_DEFAULT_PROMPT;
|
||||
next.workerId = next.workerId || deerMakeWorkerId();
|
||||
next.taskPollIntervalSeconds = Math.max(3, Number(next.taskPollIntervalSeconds || DEER_CLIPPER_DEFAULTS.taskPollIntervalSeconds));
|
||||
next.taskPageLoadTimeoutSeconds = Math.max(5, Number(next.taskPageLoadTimeoutSeconds || DEER_CLIPPER_DEFAULTS.taskPageLoadTimeoutSeconds));
|
||||
await chrome.storage.sync.set(next);
|
||||
const settings = await chrome.storage.sync.get(DEER_CLIPPER_DEFAULTS);
|
||||
document.getElementById("status").textContent = "已保存。";
|
||||
document.getElementById("serviceStatus").textContent = formatServiceStatus(settings);
|
||||
document.getElementById("tokenStatus").textContent = formatTokenStatus(settings);
|
||||
}
|
||||
|
||||
function sendDiscoverServiceMessage() {
|
||||
return new Promise((resolve, reject) => {
|
||||
chrome.runtime.sendMessage({ type: "zncm-discover-service" }, (response) => {
|
||||
const error = chrome.runtime.lastError;
|
||||
if (error) {
|
||||
reject(new Error(error.message));
|
||||
return;
|
||||
}
|
||||
resolve(response || { ok: false, message: "服务地址识别没有返回结果。" });
|
||||
});
|
||||
});
|
||||
}
|
||||
|
||||
function sendSyncTokenMessage() {
|
||||
return new Promise((resolve, reject) => {
|
||||
chrome.runtime.sendMessage({ type: "zncm-sync-token" }, (response) => {
|
||||
const error = chrome.runtime.lastError;
|
||||
if (error) {
|
||||
reject(new Error(error.message));
|
||||
return;
|
||||
}
|
||||
resolve(response || { ok: false, message: "Token 同步没有返回结果。" });
|
||||
});
|
||||
});
|
||||
}
|
||||
|
||||
async function discoverServiceNow() {
|
||||
const button = document.getElementById("discoverService");
|
||||
button.disabled = true;
|
||||
document.getElementById("serviceStatus").textContent = "正在识别服务地址...";
|
||||
try {
|
||||
await chrome.storage.sync.set({
|
||||
autoDiscoverServiceUrl: document.getElementById("autoDiscoverServiceUrl").checked,
|
||||
apiBaseUrl: deerNormalizeBaseUrl(document.getElementById("apiBaseUrl").value || DEER_CLIPPER_DEFAULTS.apiBaseUrl),
|
||||
webBaseUrl: deerNormalizeBaseUrl(document.getElementById("webBaseUrl").value || DEER_CLIPPER_DEFAULTS.webBaseUrl),
|
||||
});
|
||||
const response = await sendDiscoverServiceMessage();
|
||||
const settings = await chrome.storage.sync.get(DEER_CLIPPER_DEFAULTS);
|
||||
document.getElementById("apiBaseUrl").value = settings.apiBaseUrl || "";
|
||||
document.getElementById("webBaseUrl").value = settings.webBaseUrl || "";
|
||||
document.getElementById("serviceStatus").textContent = response.message || formatServiceStatus(settings);
|
||||
} catch (error) {
|
||||
document.getElementById("serviceStatus").textContent = error instanceof Error ? error.message : "服务地址识别失败。";
|
||||
} finally {
|
||||
button.disabled = false;
|
||||
}
|
||||
}
|
||||
|
||||
async function syncTokenNow() {
|
||||
const button = document.getElementById("syncToken");
|
||||
button.disabled = true;
|
||||
document.getElementById("tokenStatus").textContent = "正在同步 zncm token...";
|
||||
try {
|
||||
await chrome.storage.sync.set({
|
||||
tokenSyncEnabled: document.getElementById("tokenSyncEnabled").checked,
|
||||
apiBaseUrl: deerNormalizeBaseUrl(document.getElementById("apiBaseUrl").value || DEER_CLIPPER_DEFAULTS.apiBaseUrl),
|
||||
webBaseUrl: deerNormalizeBaseUrl(document.getElementById("webBaseUrl").value || DEER_CLIPPER_DEFAULTS.webBaseUrl),
|
||||
});
|
||||
const response = await sendSyncTokenMessage();
|
||||
const settings = await chrome.storage.sync.get(DEER_CLIPPER_DEFAULTS);
|
||||
document.getElementById("accessToken").value = settings.accessToken || "";
|
||||
document.getElementById("apiBaseUrl").value = settings.apiBaseUrl || "";
|
||||
document.getElementById("webBaseUrl").value = settings.webBaseUrl || "";
|
||||
document.getElementById("serviceStatus").textContent = formatServiceStatus(settings);
|
||||
document.getElementById("tokenStatus").textContent = response.message || formatTokenStatus(settings);
|
||||
} catch (error) {
|
||||
document.getElementById("tokenStatus").textContent = error instanceof Error ? error.message : "Token 同步失败。";
|
||||
} finally {
|
||||
button.disabled = false;
|
||||
}
|
||||
}
|
||||
|
||||
document.getElementById("save").addEventListener("click", saveSettings);
|
||||
document.getElementById("discoverService").addEventListener("click", discoverServiceNow);
|
||||
document.getElementById("syncToken").addEventListener("click", syncTokenNow);
|
||||
void loadSettings();
|
||||
@@ -1,31 +0,0 @@
|
||||
<!doctype html>
|
||||
<html lang="zh-CN">
|
||||
<head>
|
||||
<meta charset="utf-8" />
|
||||
<title>zncm Web Clipper</title>
|
||||
<link rel="stylesheet" href="styles.css" />
|
||||
</head>
|
||||
<body>
|
||||
<main class="popup">
|
||||
<h1>zncm Web Clipper</h1>
|
||||
<label>
|
||||
首问
|
||||
<textarea id="prompt" rows="4"></textarea>
|
||||
</label>
|
||||
<label class="checkbox">
|
||||
<input id="pageAssistantEnabled" type="checkbox" />
|
||||
启用页面问答框
|
||||
</label>
|
||||
<div class="actions">
|
||||
<button id="send" type="button">发送当前页</button>
|
||||
<button id="syncToken" type="button">同步 token</button>
|
||||
<button id="pollNow" type="button">检查任务</button>
|
||||
<button id="options" type="button" class="secondary">设置</button>
|
||||
</div>
|
||||
<p id="status" class="status"></p>
|
||||
</main>
|
||||
<script src="config.js"></script>
|
||||
<script src="shared.js"></script>
|
||||
<script src="popup.js"></script>
|
||||
</body>
|
||||
</html>
|
||||
@@ -1,151 +0,0 @@
|
||||
function setStatus(message) {
|
||||
document.getElementById("status").textContent = message;
|
||||
}
|
||||
|
||||
async function getActiveTab() {
|
||||
const [tab] = await chrome.tabs.query({ active: true, currentWindow: true });
|
||||
if (!tab?.id) {
|
||||
throw new Error("No readable active tab found.");
|
||||
}
|
||||
return tab;
|
||||
}
|
||||
|
||||
async function captureFromActiveTab(tabId) {
|
||||
const [result] = await chrome.scripting.executeScript({
|
||||
target: { tabId },
|
||||
func: deerExtractCurrentPage,
|
||||
});
|
||||
if (!result?.result?.content) {
|
||||
throw new Error("The current page has no readable text to send.");
|
||||
}
|
||||
return result.result;
|
||||
}
|
||||
|
||||
async function submitToZncm(page, settings, prompt) {
|
||||
const apiBaseUrl = deerNormalizeBaseUrl(settings.apiBaseUrl || DEER_CLIPPER_DEFAULTS.apiBaseUrl);
|
||||
const response = await fetch(`${apiBaseUrl}/api/browser-context/capture`, {
|
||||
method: "POST",
|
||||
headers: {
|
||||
"Content-Type": "application/json",
|
||||
...deerAuthHeaders(settings),
|
||||
},
|
||||
body: JSON.stringify({
|
||||
title: page.title,
|
||||
url: page.url,
|
||||
content: page.content,
|
||||
content_type: page.contentType,
|
||||
prompt,
|
||||
}),
|
||||
});
|
||||
|
||||
const data = await response.json().catch(() => ({}));
|
||||
if (!response.ok) {
|
||||
const detail = typeof data.detail === "string" ? data.detail : JSON.stringify(data.detail || data);
|
||||
throw new Error(detail || `HTTP ${response.status}`);
|
||||
}
|
||||
return data;
|
||||
}
|
||||
|
||||
async function sendCurrentPage() {
|
||||
const sendButton = document.getElementById("send");
|
||||
sendButton.disabled = true;
|
||||
setStatus("正在读取当前页面...");
|
||||
try {
|
||||
await syncToken({ silent: true });
|
||||
const settings = await chrome.storage.sync.get(DEER_CLIPPER_DEFAULTS);
|
||||
const prompt = document.getElementById("prompt").value.trim() || settings.defaultPrompt || DEER_CLIPPER_DEFAULT_PROMPT;
|
||||
const tab = await getActiveTab();
|
||||
const page = await captureFromActiveTab(tab.id);
|
||||
setStatus("正在发送到 zncm...");
|
||||
const result = await submitToZncm(page, settings, prompt);
|
||||
const webBaseUrl = deerNormalizeBaseUrl(settings.webBaseUrl || DEER_CLIPPER_DEFAULTS.webBaseUrl);
|
||||
await chrome.tabs.create({ url: `${webBaseUrl}/${result.chat_url}`.replace(/\/#/g, "/#") });
|
||||
setStatus("已发送，zncm 聊天页面已打开。");
|
||||
} catch (error) {
|
||||
setStatus(error instanceof Error ? error.message : "发送失败。");
|
||||
} finally {
|
||||
sendButton.disabled = false;
|
||||
}
|
||||
}
|
||||
|
||||
function sendSyncTokenMessage() {
|
||||
return new Promise((resolve, reject) => {
|
||||
chrome.runtime.sendMessage({ type: "zncm-sync-token" }, (response) => {
|
||||
const error = chrome.runtime.lastError;
|
||||
if (error) {
|
||||
reject(new Error(error.message));
|
||||
return;
|
||||
}
|
||||
resolve(response || { ok: false, message: "Token 同步没有返回结果。" });
|
||||
});
|
||||
});
|
||||
}
|
||||
|
||||
function sendPollNowMessage() {
|
||||
return new Promise((resolve, reject) => {
|
||||
chrome.runtime.sendMessage({ type: "zncm-poll-now" }, (response) => {
|
||||
const error = chrome.runtime.lastError;
|
||||
if (error) {
|
||||
reject(new Error(error.message));
|
||||
return;
|
||||
}
|
||||
resolve(response || { ok: true, message: "Poll requested." });
|
||||
});
|
||||
});
|
||||
}
|
||||
|
||||
async function syncToken(options = {}) {
|
||||
const button = document.getElementById("syncToken");
|
||||
button.disabled = true;
|
||||
if (!options.silent) {
|
||||
setStatus("正在同步 zncm token...");
|
||||
}
|
||||
try {
|
||||
const response = await sendSyncTokenMessage();
|
||||
if (!options.silent) {
|
||||
setStatus(response.message || "Token sync finished.");
|
||||
}
|
||||
return response;
|
||||
} catch (error) {
|
||||
if (!options.silent) {
|
||||
setStatus(error instanceof Error ? error.message : "Token 同步失败。");
|
||||
}
|
||||
return { ok: false, message: error instanceof Error ? error.message : "Token 同步失败。" };
|
||||
} finally {
|
||||
button.disabled = false;
|
||||
}
|
||||
}
|
||||
|
||||
async function pollNow() {
|
||||
const button = document.getElementById("pollNow");
|
||||
button.disabled = true;
|
||||
setStatus("正在检查 zncm 浏览器任务...");
|
||||
try {
|
||||
await syncToken({ silent: true });
|
||||
const response = await sendPollNowMessage();
|
||||
setStatus(response.message || "任务检查完成。");
|
||||
} catch (error) {
|
||||
setStatus(error instanceof Error ? error.message : "任务检查失败。");
|
||||
} finally {
|
||||
button.disabled = false;
|
||||
}
|
||||
}
|
||||
|
||||
async function togglePageAssistant() {
|
||||
const enabled = document.getElementById("pageAssistantEnabled").checked;
|
||||
await chrome.storage.sync.set({ pageAssistantEnabled: enabled });
|
||||
setStatus(enabled ? "页面问答框已开启。" : "页面问答框已关闭。");
|
||||
}
|
||||
|
||||
async function init() {
|
||||
const settings = await chrome.storage.sync.get(DEER_CLIPPER_DEFAULTS);
|
||||
document.getElementById("prompt").value = settings.defaultPrompt || DEER_CLIPPER_DEFAULT_PROMPT;
|
||||
document.getElementById("pageAssistantEnabled").checked = Boolean(settings.pageAssistantEnabled);
|
||||
document.getElementById("send").addEventListener("click", sendCurrentPage);
|
||||
document.getElementById("syncToken").addEventListener("click", () => syncToken());
|
||||
document.getElementById("pollNow").addEventListener("click", pollNow);
|
||||
document.getElementById("pageAssistantEnabled").addEventListener("change", togglePageAssistant);
|
||||
document.getElementById("options").addEventListener("click", () => chrome.runtime.openOptionsPage());
|
||||
}
|
||||
|
||||
void init();
|
||||
@@ -1,389 +0,0 @@
|
||||
const DEER_CLIPPER_DEFAULT_PROMPT =
|
||||
"\u8bf7\u57fa\u4e8e\u521a\u521a\u53d1\u9001\u7684\u7f51\u9875\u5185\u5bb9\uff0c\u5148\u7528\u4e2d\u6587\u603b\u7ed3\u8981\u70b9\uff0c\u5e76\u7b49\u5f85\u6211\u7ee7\u7eed\u63d0\u95ee\u3002";
|
||||
|
||||
const ZNCM_EXTENSION_CONFIG = globalThis.ZNCM_WEB_CLIPPER_CONFIG || {};
|
||||
const ZNCM_DEFAULT_WEB_BASE_URL = String(ZNCM_EXTENSION_CONFIG.defaultWebBaseUrl || "").trim().replace(/\/+$/, "");
|
||||
const ZNCM_DEFAULT_API_BASE_URL = String(ZNCM_EXTENSION_CONFIG.defaultApiBaseUrl || "").trim().replace(/\/+$/, "");
|
||||
|
||||
const DEER_CLIPPER_DEFAULTS = {
|
||||
apiBaseUrl: ZNCM_DEFAULT_API_BASE_URL,
|
||||
webBaseUrl: ZNCM_DEFAULT_WEB_BASE_URL,
|
||||
autoDiscoverServiceUrl: true,
|
||||
lastServiceSyncAt: "",
|
||||
lastServiceSyncSource: "",
|
||||
accessToken: "",
|
||||
tokenSyncEnabled: true,
|
||||
lastTokenSyncAt: "",
|
||||
lastTokenSyncSource: "",
|
||||
defaultPrompt: DEER_CLIPPER_DEFAULT_PROMPT,
|
||||
pageAssistantEnabled: true,
|
||||
taskPollingEnabled: true,
|
||||
workerId: "",
|
||||
taskPollIntervalSeconds: 5,
|
||||
taskPageLoadTimeoutSeconds: 30,
|
||||
closeTaskTabs: false,
|
||||
};
|
||||
|
||||
function deerNormalizeBaseUrl(value) {
|
||||
return String(value || "").trim().replace(/\/+$/, "");
|
||||
}
|
||||
|
||||
function deerAuthHeaders(settings) {
|
||||
const token = String(settings.accessToken || "").trim();
|
||||
if (!token) {
|
||||
return {};
|
||||
}
|
||||
return {
|
||||
Authorization: /^bearer\s+/i.test(token) ? token : `Bearer ${token}`,
|
||||
};
|
||||
}
|
||||
|
||||
function deerMakeWorkerId() {
|
||||
return `zncm-${Date.now().toString(36)}-${Math.random().toString(36).slice(2, 10)}`;
|
||||
}
|
||||
|
||||
function deerExtractAuthTokenFromPage() {
|
||||
const storageKeys = ["deerflow.auth"];
|
||||
const tokenFieldNames = new Set(["access_token", "accessToken", "token", "id_token", "idToken"]);
|
||||
|
||||
function looksLikeToken(value) {
|
||||
const text = String(value || "").trim();
|
||||
return text.length > 20 && !/\s/.test(text);
|
||||
}
|
||||
|
||||
function findToken(value, depth = 0) {
|
||||
if (value == null || depth > 4) {
|
||||
return "";
|
||||
}
|
||||
if (typeof value === "string") {
|
||||
const text = value.trim();
|
||||
if (!text) {
|
||||
return "";
|
||||
}
|
||||
if ((text.startsWith("{") && text.endsWith("}")) || (text.startsWith("[") && text.endsWith("]"))) {
|
||||
try {
|
||||
return findToken(JSON.parse(text), depth + 1);
|
||||
} catch {
|
||||
return looksLikeToken(text) ? text : "";
|
||||
}
|
||||
}
|
||||
return looksLikeToken(text) ? text : "";
|
||||
}
|
||||
if (typeof value !== "object") {
|
||||
return "";
|
||||
}
|
||||
for (const [key, nested] of Object.entries(value)) {
|
||||
if (tokenFieldNames.has(key) && looksLikeToken(nested)) {
|
||||
return String(nested).trim();
|
||||
}
|
||||
}
|
||||
for (const nested of Object.values(value)) {
|
||||
const token = findToken(nested, depth + 1);
|
||||
if (token) {
|
||||
return token;
|
||||
}
|
||||
}
|
||||
return "";
|
||||
}
|
||||
|
||||
function scanStorage(storage, storageName) {
|
||||
for (const key of storageKeys) {
|
||||
const token = findToken(storage.getItem(key));
|
||||
if (token) {
|
||||
return { token, source: `${storageName}:${key}` };
|
||||
}
|
||||
}
|
||||
|
||||
for (let index = 0; index < storage.length; index += 1) {
|
||||
const key = storage.key(index);
|
||||
if (!key || !/(auth|token|session)/i.test(key)) {
|
||||
continue;
|
||||
}
|
||||
const token = findToken(storage.getItem(key));
|
||||
if (token) {
|
||||
return { token, source: `${storageName}:${key}` };
|
||||
}
|
||||
}
|
||||
return null;
|
||||
}
|
||||
|
||||
try {
|
||||
return scanStorage(window.localStorage, "localStorage") || scanStorage(window.sessionStorage, "sessionStorage");
|
||||
} catch (error) {
|
||||
return { token: "", source: "", error: error instanceof Error ? error.message : String(error) };
|
||||
}
|
||||
}
|
||||
|
||||
function deerExtractCurrentPage() {
|
||||
const maxLocalChars = 220000;
|
||||
|
||||
function cleanText(value) {
|
||||
return String(value || "")
|
||||
.replace(/\s+\n/g, "\n")
|
||||
.replace(/\n{3,}/g, "\n\n")
|
||||
.trim();
|
||||
}
|
||||
|
||||
function escapeCell(value) {
|
||||
return cleanText(value).replace(/\|/g, "\\|").replace(/\n/g, " ");
|
||||
}
|
||||
|
||||
function tableToMarkdown(table, index) {
|
||||
const rows = Array.from(table.rows || []).map((row) =>
|
||||
Array.from(row.cells || []).map((cell) => escapeCell(cell.innerText)),
|
||||
);
|
||||
if (rows.length === 0 || rows.every((row) => row.length === 0)) {
|
||||
return "";
|
||||
}
|
||||
const width = Math.max(...rows.map((row) => row.length));
|
||||
const normalized = rows.map((row) => {
|
||||
const next = row.slice();
|
||||
while (next.length < width) next.push("");
|
||||
return next;
|
||||
});
|
||||
const header = normalized[0];
|
||||
const divider = Array.from({ length: width }, () => "---");
|
||||
const body = normalized.slice(1);
|
||||
return [
|
||||
`### Table ${index + 1}`,
|
||||
"",
|
||||
`| ${header.join(" | ")} |`,
|
||||
`| ${divider.join(" | ")} |`,
|
||||
...body.map((row) => `| ${row.join(" | ")} |`),
|
||||
].join("\n");
|
||||
}
|
||||
|
||||
const clone = document.body ? document.body.cloneNode(true) : null;
|
||||
if (clone) {
|
||||
clone.querySelectorAll("script, style, noscript, svg, canvas, iframe").forEach((node) => node.remove());
|
||||
}
|
||||
|
||||
const title = document.title || "";
|
||||
const url = location.href;
|
||||
const description = document.querySelector('meta[name="description"]')?.getAttribute("content") || "";
|
||||
const main = clone?.querySelector("article, main, [role='main']") || clone;
|
||||
const text = cleanText(main?.innerText || document.body?.innerText || "");
|
||||
const tables = Array.from(document.querySelectorAll("table"))
|
||||
.map((table, index) => tableToMarkdown(table, index))
|
||||
.filter(Boolean);
|
||||
|
||||
const parts = [];
|
||||
if (description.trim()) {
|
||||
parts.push(`## Meta Description\n\n${cleanText(description)}`);
|
||||
}
|
||||
if (text) {
|
||||
parts.push(`## Page Text\n\n${text}`);
|
||||
}
|
||||
if (tables.length > 0) {
|
||||
parts.push(`## Tables\n\n${tables.join("\n\n")}`);
|
||||
}
|
||||
|
||||
let content = parts.join("\n\n");
|
||||
if (content.length > maxLocalChars) {
|
||||
content = `${content.slice(0, maxLocalChars)}\n\n[Truncated locally by zncm Web Clipper]`;
|
||||
}
|
||||
|
||||
return { title, url, content, contentType: "markdown" };
|
||||
}
|
||||
|
||||
// Injected into the page to build an indexed list of interactive elements plus
|
||||
// the visible page text. Each interactive element is tagged with
|
||||
// `data-deer-index` so a follow-up deerPerformAction call can find it by index.
|
||||
function deerObservePage(maxElements) {
|
||||
const limit = Math.max(10, Number(maxElements || 150));
|
||||
const DEER_ATTR = "data-deer-index";
|
||||
|
||||
document.querySelectorAll(`[${DEER_ATTR}]`).forEach((node) => node.removeAttribute(DEER_ATTR));
|
||||
|
||||
function isVisible(el) {
|
||||
if (!(el instanceof Element)) return false;
|
||||
const rect = el.getBoundingClientRect();
|
||||
if (rect.width <= 1 && rect.height <= 1) return false;
|
||||
const style = window.getComputedStyle(el);
|
||||
if (style.visibility === "hidden" || style.display === "none" || style.opacity === "0") return false;
|
||||
if (el.disabled) return false;
|
||||
return true;
|
||||
}
|
||||
|
||||
function labelFor(el) {
|
||||
const aria = el.getAttribute("aria-label");
|
||||
if (aria && aria.trim()) return aria.trim();
|
||||
const placeholder = el.getAttribute("placeholder");
|
||||
const title = el.getAttribute("title");
|
||||
const name = el.getAttribute("name");
|
||||
const text = (el.innerText || el.textContent || "").trim();
|
||||
if (text) return text.replace(/\s+/g, " ");
|
||||
if (placeholder && placeholder.trim()) return placeholder.trim();
|
||||
if (title && title.trim()) return title.trim();
|
||||
if (el.value && String(el.value).trim()) return String(el.value).trim();
|
||||
if (name && name.trim()) return name.trim();
|
||||
const alt = el.querySelector?.("img[alt]")?.getAttribute("alt");
|
||||
if (alt && alt.trim()) return alt.trim();
|
||||
return "";
|
||||
}
|
||||
|
||||
const selector = [
|
||||
"a[href]",
|
||||
"button",
|
||||
"input:not([type=hidden])",
|
||||
"select",
|
||||
"textarea",
|
||||
"[role=button]",
|
||||
"[role=link]",
|
||||
"[role=tab]",
|
||||
"[role=menuitem]",
|
||||
"[role=checkbox]",
|
||||
"[role=radio]",
|
||||
"[role=switch]",
|
||||
"[contenteditable=true]",
|
||||
"[onclick]",
|
||||
].join(",");
|
||||
|
||||
const seen = new Set();
|
||||
const elements = [];
|
||||
let index = 0;
|
||||
for (const el of document.querySelectorAll(selector)) {
|
||||
if (index >= limit) break;
|
||||
if (seen.has(el) || !isVisible(el)) continue;
|
||||
seen.add(el);
|
||||
el.setAttribute(DEER_ATTR, String(index));
|
||||
const tag = el.tagName.toLowerCase();
|
||||
let label = labelFor(el);
|
||||
if (label.length > 200) label = label.slice(0, 200);
|
||||
const entry = { index, tag, label };
|
||||
const type = el.getAttribute("type") || el.getAttribute("role");
|
||||
if (type) entry.type = type;
|
||||
if (tag === "input" || tag === "textarea" || tag === "select") {
|
||||
const value = String(el.value || "");
|
||||
if (value) entry.value = value.slice(0, 80);
|
||||
}
|
||||
if (tag === "a") {
|
||||
const href = el.getAttribute("href") || "";
|
||||
if (href) entry.href = href.slice(0, 200);
|
||||
}
|
||||
elements.push(entry);
|
||||
index += 1;
|
||||
}
|
||||
|
||||
// Inline text extraction so this function stays self-contained when injected
|
||||
// via chrome.scripting.executeScript (no dependency on other content-script fns).
|
||||
let text = "";
|
||||
try {
|
||||
const clone = document.body ? document.body.cloneNode(true) : null;
|
||||
if (clone) {
|
||||
clone.querySelectorAll(`script, style, noscript, svg, canvas, iframe, [${DEER_ATTR}-skip]`).forEach((node) => node.remove());
|
||||
}
|
||||
const main = clone?.querySelector("article, main, [role='main']") || clone;
|
||||
text = String(main?.innerText || document.body?.innerText || "")
|
||||
.replace(/\s+\n/g, "\n")
|
||||
.replace(/\n{3,}/g, "\n\n")
|
||||
.trim();
|
||||
} catch {
|
||||
text = "";
|
||||
}
|
||||
|
||||
return {
|
||||
url: location.href,
|
||||
title: document.title || "",
|
||||
elements,
|
||||
text,
|
||||
};
|
||||
}
|
||||
|
||||
// Injected into the page to perform a single action on the element previously
|
||||
// tagged by deerObservePage (by index). Returns {status, message}.
|
||||
async function deerPerformAction(action) {
|
||||
const DEER_ATTR = "data-deer-index";
|
||||
const type = String(action?.type || "").toLowerCase();
|
||||
|
||||
function findByIndex(idx) {
|
||||
if (idx === undefined || idx === null || idx < 0) return null;
|
||||
return document.querySelector(`[${DEER_ATTR}="${idx}"]`);
|
||||
}
|
||||
|
||||
function fireInput(el, value) {
|
||||
const proto = el instanceof HTMLTextAreaElement ? HTMLTextAreaElement.prototype : HTMLInputElement.prototype;
|
||||
const setter = Object.getOwnPropertyDescriptor(proto, "value")?.set;
|
||||
if (setter) {
|
||||
setter.call(el, value);
|
||||
} else {
|
||||
el.value = value;
|
||||
}
|
||||
el.dispatchEvent(new Event("input", { bubbles: true }));
|
||||
el.dispatchEvent(new Event("change", { bubbles: true }));
|
||||
}
|
||||
|
||||
try {
|
||||
if (type === "observe" || type === "extract") {
|
||||
return { status: "ok", message: type === "extract" ? "Page text re-read." : "Observed page." };
|
||||
}
|
||||
|
||||
if (type === "scroll") {
|
||||
const el = findByIndex(action.index);
|
||||
if (el) {
|
||||
el.scrollIntoView({ block: "center", behavior: "instant" });
|
||||
return { status: "ok", message: "Scrolled element into view." };
|
||||
}
|
||||
window.scrollBy(0, Math.round(window.innerHeight * 0.85));
|
||||
return { status: "ok", message: "Scrolled down one viewport." };
|
||||
}
|
||||
|
||||
if (type === "wait") {
|
||||
const selector = String(action.wait_for || "").trim();
|
||||
const timeoutMs = Math.max(500, Number(action.timeout_ms || 8000));
|
||||
if (!selector) {
|
||||
await new Promise((resolve) => setTimeout(resolve, Math.min(timeoutMs, 3000)));
|
||||
return { status: "ok", message: "Waited." };
|
||||
}
|
||||
const start = Date.now();
|
||||
while (Date.now() - start < timeoutMs) {
|
||||
if (document.querySelector(selector)) {
|
||||
return { status: "ok", message: `Element ${selector} appeared.` };
|
||||
}
|
||||
await new Promise((resolve) => setTimeout(resolve, 200));
|
||||
}
|
||||
return { status: "timeout", message: `Element ${selector} did not appear within ${timeoutMs}ms.` };
|
||||
}
|
||||
|
||||
if (type === "click") {
|
||||
const el = findByIndex(action.index);
|
||||
if (!el) return { status: "error", message: `No element with index ${action.index}. Re-observe the page.` };
|
||||
el.scrollIntoView({ block: "center", behavior: "instant" });
|
||||
if (typeof el.focus === "function") el.focus();
|
||||
el.click();
|
||||
return { status: "ok", message: `Clicked element ${action.index}.` };
|
||||
}
|
||||
|
||||
if (type === "type") {
|
||||
const el = findByIndex(action.index);
|
||||
if (!el) return { status: "error", message: `No element with index ${action.index}. Re-observe the page.` };
|
||||
el.scrollIntoView({ block: "center", behavior: "instant" });
|
||||
if (typeof el.focus === "function") el.focus();
|
||||
const value = String(action.text ?? "");
|
||||
if (el.isContentEditable) {
|
||||
el.textContent = value;
|
||||
el.dispatchEvent(new Event("input", { bubbles: true }));
|
||||
} else {
|
||||
fireInput(el, value);
|
||||
}
|
||||
if (action.submit) {
|
||||
const enterInit = { bubbles: true, cancelable: true, key: "Enter", code: "Enter", keyCode: 13, which: 13 };
|
||||
el.dispatchEvent(new KeyboardEvent("keydown", enterInit));
|
||||
el.dispatchEvent(new KeyboardEvent("keyup", enterInit));
|
||||
const form = el.form || el.closest("form");
|
||||
if (form && typeof form.requestSubmit === "function") {
|
||||
try {
|
||||
form.requestSubmit();
|
||||
} catch {
|
||||
form.submit?.();
|
||||
}
|
||||
}
|
||||
}
|
||||
return { status: "ok", message: `Typed into element ${action.index}${action.submit ? " and submitted" : ""}.` };
|
||||
}
|
||||
|
||||
return { status: "error", message: `Unsupported action type: ${type}` };
|
||||
} catch (error) {
|
||||
return { status: "error", message: error instanceof Error ? error.message : String(error) };
|
||||
}
|
||||
}
|
||||
@@ -1,96 +0,0 @@
|
||||
* {
|
||||
box-sizing: border-box;
|
||||
}
|
||||
|
||||
body {
|
||||
margin: 0;
|
||||
color: #171717;
|
||||
background: #fafafa;
|
||||
font: 14px/1.5 system-ui, -apple-system, BlinkMacSystemFont, "Segoe UI", sans-serif;
|
||||
}
|
||||
|
||||
h1 {
|
||||
margin: 0 0 14px;
|
||||
font-size: 16px;
|
||||
}
|
||||
|
||||
.popup,
|
||||
.options {
|
||||
display: grid;
|
||||
gap: 12px;
|
||||
padding: 14px;
|
||||
}
|
||||
|
||||
.popup {
|
||||
width: 340px;
|
||||
}
|
||||
|
||||
.options {
|
||||
max-width: 720px;
|
||||
}
|
||||
|
||||
label {
|
||||
display: grid;
|
||||
gap: 6px;
|
||||
font-weight: 600;
|
||||
}
|
||||
|
||||
label.checkbox {
|
||||
grid-template-columns: auto 1fr;
|
||||
align-items: center;
|
||||
font-weight: 500;
|
||||
}
|
||||
|
||||
label.checkbox input {
|
||||
width: auto;
|
||||
}
|
||||
|
||||
.actions {
|
||||
display: flex;
|
||||
flex-wrap: wrap;
|
||||
gap: 8px;
|
||||
}
|
||||
|
||||
input,
|
||||
textarea {
|
||||
width: 100%;
|
||||
border: 1px solid #d4d4d4;
|
||||
border-radius: 6px;
|
||||
padding: 8px;
|
||||
background: #fff;
|
||||
color: inherit;
|
||||
font: inherit;
|
||||
font-weight: 400;
|
||||
}
|
||||
|
||||
textarea {
|
||||
resize: vertical;
|
||||
}
|
||||
|
||||
button {
|
||||
border: 0;
|
||||
border-radius: 6px;
|
||||
padding: 9px 12px;
|
||||
background: #111827;
|
||||
color: #fff;
|
||||
cursor: pointer;
|
||||
font: inherit;
|
||||
font-weight: 600;
|
||||
}
|
||||
|
||||
button.secondary {
|
||||
background: #e5e7eb;
|
||||
color: #111827;
|
||||
}
|
||||
|
||||
button:disabled {
|
||||
cursor: not-allowed;
|
||||
opacity: 0.6;
|
||||
}
|
||||
|
||||
.status {
|
||||
min-height: 20px;
|
||||
margin: 0;
|
||||
color: #525252;
|
||||
white-space: pre-wrap;
|
||||
}
|
||||
@@ -1,18 +1,6 @@
|
||||
# 本地开发（pnpm dev）— 与原先 .env 保持一致
|
||||
# 个人覆盖请新建 .env.development.local（已 gitignore，优先级更高）
|
||||
|
||||
# 任务看板数据源：true=真实接口（默认）；改成 false 可让任务分析/三情分析两页走本地假数据。
|
||||
VITE_USE_REAL_DATA=true
|
||||
|
||||
# 任务详情（登录深链 cop-task-detail）数据源：默认真实接口；改成 true 走本地假数据（consumer 外网不可达时调试用）。
|
||||
VITE_MOCK_COP_TASK_DETAIL=true
|
||||
|
||||
# 研判情报报告（大屏绘制智能体深链 selectSqReport）数据源：true 走本地样例假数据（真实接口未就绪时调试用）；联调通过后置 false。
|
||||
VITE_MOCK_SQ_REPORT=true
|
||||
|
||||
# 深链任务详情（cop-task-detail：任务名称/描述）数据源：true 走本地假数据（真实接口未就绪时调试用）；联调通过后置 false。
|
||||
VITE_MOCK_COP_TASK_DETAIL=true
|
||||
|
||||
# Gateway / DeerFlow API
|
||||
VITE_BACKEND_BASE_URL=http://127.0.0.1:8001
|
||||
VITE_LANGGRAPH_BASE_URL=http://127.0.0.1:8001/api
|
||||
@@ -36,17 +24,10 @@ VITE_AIPI_B2ACTOR_XWT_PATH=/fromB/toXWT?isFrame=1
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_API_URL=https://ch1.b.uat.4.cn/knowes/api/iframeKnowledgeUrl
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_SKIN=red
|
||||
VITE_AIPI_MOPECHAT_URL=https://ch1.b.uat.4.cn/likun_b1_mope
|
||||
VITE_AIPI_BACKENP_URL=http://47.88.25.99:8000
|
||||
# 工作室/空间后端接口（侧边栏“工作室”列表、空间详情等都走这里，不是 VITE_NOTE_API_BASE_URL）
|
||||
# 工作室/空间前端地址（用于 auth/callback 等跳转）
|
||||
# VITE_AIPI_FRONTEND_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
VITE_AIPI_FRONTEND_URL=http://localhost:3000
|
||||
VITE_AIPI_BACKENP_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
# 我的笔记前端
|
||||
VITE_AIPI_FRONTEND_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
VITE_AIPI_CONSUMER_URL=https://ch1.b.uat.4.cn/consumer
|
||||
# A 系统跳转地址（人物/组织/事件等列表项打开链接）
|
||||
VITE_AIPI_PRESENT_URL=https://present.a.uat.4.cn/present/domain
|
||||
# B 系统知识库 iframe 根地址
|
||||
VITE_AIPI_KNOWES_URL=https://ch1.b.uat.4.cn/knowes
|
||||
|
||||
# MaxKey 鉴权（getTokenByOnlineTicket）
|
||||
VITE_AUTH_BASE_URL=https://auth.y.sit.4.cn
|
||||
VITE_AIPI_PGLINKTO_URL=https://pg.b.uat.4.cn
|
||||
@@ -54,3 +35,5 @@ VITE_AIPI_XKLINKTO_URL=https://kz.b.uat.4.cn
|
||||
# 可选：Canvas 独立 LangGraph 地址（不设则回退到 VITE_LANGGRAPH_BASE_URL）
|
||||
# VITE_CANVAS_LANGGRAPH_BASE_URL=
|
||||
|
||||
# 我的笔记是否显示，侧边栏
|
||||
VITE_STUDIO_ENABLED=false
|
||||
@@ -15,15 +15,8 @@ VITE_AIPI_B2ACTOR_XWT_PATH=/fromB/toXWT?isFrame=1
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_API_URL=https://ch1.b.uat.4.cn/knowes/api/iframeKnowledgeUrl
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_SKIN=red
|
||||
VITE_AIPI_MOPECHAT_URL=https://ch1.b.uat.4.cn/likun_b1_mope
|
||||
# 工作室/空间后端接口（侧边栏“工作室”列表、空间详情等都走这里，不是 VITE_NOTE_API_BASE_URL）
|
||||
VITE_AIPI_BACKENP_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
# 工作室/空间前端地址（用于 auth/callback 等跳转）
|
||||
VITE_AIPI_FRONTEND_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
VITE_AIPI_CONSUMER_URL=https://ch1.b.uat.4.cn/consumer
|
||||
# A 系统跳转地址（人物/组织/事件等列表项打开链接）
|
||||
VITE_AIPI_PRESENT_URL=http://present.a.uat.4.cn/present/domain
|
||||
# B 系统知识库 iframe 根地址
|
||||
VITE_AIPI_KNOWES_URL=http://ch1.b.uat.4.cn/knowes
|
||||
VITE_AUTH_BASE_URL=https://auth.y.sit.4.cn
|
||||
VITE_AIPI_PGLINKTO_URL=https://pg.b.uat.4.cn
|
||||
VITE_AIPI_XKLINKTO_URL=https://kz.b.uat.4.cn
|
||||
|
||||
@@ -14,30 +14,17 @@
|
||||
# - 本机连 127.0.0.1:8001：复制 .env.development.local.example → .env.development.local
|
||||
# - 内网打包部署：改 .env.network 后执行 pnpm build:network（或改 .env.production 后 pnpm build）
|
||||
#
|
||||
# 若浏览器直连后端报 CORS，请在 Gateway 配置 GATEWAY_CORS_ORIGINS，包含前端源站；无界嵌入时还要加宿主源站，例如：
|
||||
# GATEWAY_CORS_ORIGINS=http://127.0.0.1:5174,http://localhost:5174,http://127.0.0.1:5183,http://localhost:5183
|
||||
# 若浏览器直连后端报 CORS，请在 Gateway 配置 CORS_ORIGINS，包含前端源站，例如：
|
||||
# CORS_ORIGINS=http://127.0.0.1:5173,http://localhost:5173,http://127.0.0.1:5174
|
||||
|
||||
# --- 开发环境（见 .env.development）---
|
||||
# VITE_BACKEND_BASE_URL=https://ch1.b.uat.4.cn/deerflow
|
||||
# VITE_LANGGRAPH_BASE_URL=https://ch1.b.uat.4.cn/deerflow/api
|
||||
# VITE_NOTE_API_BASE_URL=http://47.88.25.99:8000
|
||||
# VITE_AIPI_BACKENP_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
# VITE_AIPI_FRONTEND_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
|
||||
# --- 生产打包（见 .env.production）---
|
||||
# 部署前修改 .env.production 中的地址，再执行 pnpm build
|
||||
|
||||
# --- 任务看板（任务分析 / 三情分析）数据源开关 ---
|
||||
# true=走真实接口；false 或未设=这两页使用本地假数据（演示用，无需后端）。
|
||||
# .env.development / .env.production / .env.network / .env.sit 均默认 true（真实数据），
|
||||
# 仅本机 .env 默认 false（假数据）。改后需重启 dev / 重新打包生效。
|
||||
# VITE_USE_REAL_DATA=true
|
||||
|
||||
# --- 任务详情（登录深链 cop-task-detail）数据源开关 ---
|
||||
# 默认真实接口（未设/false）；改成 true 时该接口走本地假数据（consumer 外网不可达时调试用）。
|
||||
# 与 VITE_USE_REAL_DATA 独立。改后需重启 dev / 重新打包生效。
|
||||
# VITE_MOCK_COP_TASK_DETAIL=false
|
||||
|
||||
# MaxKey 鉴权根地址（无界 onlineTicket 换 token）
|
||||
# VITE_AUTH_BASE_URL=https://auth.y.sit.4.cn
|
||||
# VITE_AIPI_CONSUMER_URL=https://ch1.b.uat.4.cn/consumer
|
||||
@@ -50,6 +37,4 @@
|
||||
# 可选
|
||||
# VITE_CANVAS_LANGGRAPH_BASE_URL=
|
||||
# VITE_STATIC_WEBSITE_ONLY=
|
||||
# 部署级紧凑导航：true 时固定收起工作台左侧栏，并隐藏换肤与全量/简洁模式切换。
|
||||
# VITE_FORCE_COLLAPSED_SIDEBAR=false
|
||||
# VITE_GITHUB_OAUTH_TOKEN=
|
||||
+23
-48
@@ -1,57 +1,32 @@
|
||||
# 本地开发（pnpm dev）— 与原先 .env 保持一致
|
||||
# 个人覆盖请新建 .env.development.local（已 gitignore，优先级更高）
|
||||
# 内网 / 局域网生产打包（npm run build:network 或 pnpm build:network）
|
||||
# 打包前请将 192.168.1.100 改为实际服务器 IP；变量会写入 dist，运行时不可再改
|
||||
|
||||
# 任务看板数据源：true=真实接口（默认）；改成 false 可让任务分析/三情分析两页走本地假数据。
|
||||
VITE_USE_REAL_DATA=true
|
||||
|
||||
# 任务详情（登录深链 cop-task-detail）数据源：默认真实接口；改成 true 走本地假数据（调试用）。
|
||||
VITE_MOCK_COP_TASK_DETAIL=false
|
||||
|
||||
# 研判情报报告（大屏绘制智能体深链 selectSqReport）数据源：默认真实接口；真实接口未就绪可改 true 走本地样例假数据。
|
||||
VITE_MOCK_SQ_REPORT=false
|
||||
|
||||
# 深链任务详情（cop-task-detail）数据源：默认真实接口；真实接口未就绪可改 true 走本地假数据。
|
||||
VITE_MOCK_COP_TASK_DETAIL=false
|
||||
|
||||
# Gateway / DeerFlow API
|
||||
VITE_BACKEND_BASE_URL=http://10.19.200.5:2026
|
||||
VITE_LANGGRAPH_BASE_URL=http://10.19.200.5:2026/api
|
||||
# Gateway / DeerFlow API（与 offline-backend DEPLOY 文档一致）
|
||||
VITE_BACKEND_BASE_URL=http://47.88.25.99:7001
|
||||
VITE_LANGGRAPH_BASE_URL=http://47.88.25.99:7001/api
|
||||
|
||||
# 轻应用「新建笔记」系列接口
|
||||
VITE_NOTE_API_BASE_URL=http://47.88.25.99:8000
|
||||
# CMZS后端接口
|
||||
VITE_API_BASE_URL=https://ch1.b.uat.4.cn/zncm
|
||||
VITE_NOTE_API_BASE_URL=http://192.168.1.100:8000
|
||||
# CMZS 后端接口
|
||||
VITE_API_BASE_URL=http://192.168.1.100/zncm
|
||||
# 佳来模型对话接口
|
||||
VITE_AIPI_CHAT_URL=https://ch1.b.uat.4.cn/gpt
|
||||
VITE_AIPI_CHAT_URL=http://192.168.1.100/gpt
|
||||
# 多轮检索接口
|
||||
VITE_AIPI_GENERATE_CHAT_URL=https://ch1.b.uat.4.cn/jialai_b1_chat
|
||||
VITE_AIPI_TASK_CHAT_URL=https://ch1.b.uat.4.cn/qq_b1_chat
|
||||
VITE_AIPI_KNOWLEDGE_URL=https://chlib.b.uat.4.cn/knowlibrary
|
||||
VITE_AIPI_B1COH_URL=https://ch1.b.uat.4.cn/web
|
||||
VITE_AIPI_B2ACTOR_URL=https://ty.b.uat.4.cn
|
||||
VITE_AIPI_GENERATE_CHAT_URL=http://192.168.1.100/jialai_b1_chat
|
||||
VITE_AIPI_TASK_CHAT_URL=http://192.168.1.100/qq_b1_chat
|
||||
VITE_AIPI_KNOWLEDGE_URL=http://192.168.1.100/knowlibrary
|
||||
VITE_AIPI_B1COH_URL=http://192.168.1.100/web
|
||||
VITE_AIPI_B2ACTOR_URL=http://192.168.1.100
|
||||
VITE_AIPI_B2ACTOR_XWT_PATH=/fromB/toXWT?isFrame=1
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_API_URL=https://ch1.b.uat.4.cn/knowes/api/iframeKnowledgeUrl
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_API_URL=http://192.168.1.100/knowes/api/iframeKnowledgeUrl
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_SKIN=red
|
||||
VITE_AIPI_B2ACTOR_XWT_PATH=/fromB/toXWT?isFrame=1
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_API_URL=https://ch1.b.uat.4.cn/knowes/api/iframeKnowledgeUrl
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_SKIN=red
|
||||
VITE_AIPI_MOPECHAT_URL=https://ch1.b.uat.4.cn/likun_b1_mope
|
||||
VITE_AIPI_BACKENP_URL=http://47.88.25.99:8000
|
||||
# 工作室/空间后端接口（侧边栏“工作室”列表、空间详情等都走这里，不是 VITE_NOTE_API_BASE_URL）
|
||||
# 工作室/空间前端地址（用于 auth/callback 等跳转）
|
||||
# VITE_AIPI_FRONTEND_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
VITE_NOTEBOOK_API_BASE_URL=http://47.88.25.99:8000
|
||||
VITE_AIPI_FRONTEND_URL=http://47.88.25.99:3000
|
||||
VITE_AIPI_CONSUMER_URL=http://10.19.200.5:8002/consumer
|
||||
# A 系统跳转地址（人物/组织/事件等列表项打开链接）
|
||||
VITE_AIPI_PRESENT_URL=http://present.a.uat.4.cn/present/domain
|
||||
# B 系统知识库 iframe 根地址
|
||||
VITE_AIPI_KNOWES_URL=http://ch1.b.uat.4.cn/knowes
|
||||
|
||||
# MaxKey 鉴权（getTokenByOnlineTicket）
|
||||
VITE_AIPI_MOPECHAT_URL=http://192.168.1.100/likun_b1_mope
|
||||
VITE_AIPI_BACKENP_URL=http://192.168.1.100
|
||||
VITE_AIPI_CONSUMER_URL=http://192.168.1.100/consumer
|
||||
# MaxKey 鉴权（getTokenByOnlineTicket，内网部署请改为实际地址）
|
||||
VITE_AUTH_BASE_URL=https://auth.y.sit.4.cn
|
||||
VITE_AIPI_PGLINKTO_URL=https://pg.b.uat.4.cn
|
||||
VITE_AIPI_XKLINKTO_URL=https://kz.b.uat.4.cn
|
||||
# 可选：Canvas 独立 LangGraph 地址（不设则回退到 VITE_LANGGRAPH_BASE_URL）
|
||||
# VITE_CANVAS_LANGGRAPH_BASE_URL=
|
||||
VITE_AIPI_PGLINKTO_URL=http://192.168.1.100
|
||||
VITE_AIPI_XKLINKTO_URL=http://192.168.1.100
|
||||
|
||||
# 可选：Canvas 独立 LangGraph 地址
|
||||
# VITE_CANVAS_LANGGRAPH_BASE_URL=
|
||||
@@ -1,23 +1,12 @@
|
||||
# 生产打包（pnpm build）— 部署前请改成实际内网/生产地址
|
||||
# 打包后变量会写入 dist，运行时不可再改
|
||||
|
||||
# 任务看板数据源：true=真实接口（默认）；改成 false 可让任务分析/三情分析两页走本地假数据。
|
||||
VITE_USE_REAL_DATA=true
|
||||
|
||||
# 任务详情（登录深链 cop-task-detail）数据源：默认真实接口；改成 true 走本地假数据（调试用）。
|
||||
VITE_MOCK_COP_TASK_DETAIL=false
|
||||
|
||||
# 研判情报报告（大屏绘制智能体深链 selectSqReport）数据源：默认真实接口；真实接口未就绪可改 true 走本地样例假数据。
|
||||
VITE_MOCK_SQ_REPORT=false
|
||||
|
||||
# Gateway / DeerFlow API
|
||||
VITE_BACKEND_BASE_URL=https://ch1.b.uat.4.cn/deerflow
|
||||
VITE_LANGGRAPH_BASE_URL=https://ch1.b.uat.4.cn/deerflow/api
|
||||
|
||||
# 轻应用「新建笔记」系列接口
|
||||
VITE_NOTE_API_BASE_URL=https://your-intranet-host/note-api
|
||||
# CMZS后端接口
|
||||
VITE_API_BASE_URL=https://ch1.b.uat.4.cn/zncm
|
||||
# 佳来模型对话接口
|
||||
VITE_AIPI_CHAT_URL=https://ch1.b.uat.4.cn/gpt
|
||||
# 多轮检索接口
|
||||
@@ -30,21 +19,13 @@ VITE_AIPI_B2ACTOR_XWT_PATH=/fromB/toXWT?isFrame=1
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_API_URL=https://ch1.b.uat.4.cn/knowes/api/iframeKnowledgeUrl
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_SKIN=red
|
||||
VITE_AIPI_MOPECHAT_URL=https://ch1.b.uat.4.cn/likun_b1_mope
|
||||
# 工作室/空间后端接口（侧边栏“工作室”列表、空间详情等都走这里，不是 VITE_NOTE_API_BASE_URL）
|
||||
VITE_AIPI_BACKENP_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
# 我的空间前端
|
||||
VITE_AIPI_FRONTEND_URL= xxx
|
||||
VITE_AIPI_CONSUMER_URL=https://ch1.b.uat.4.cn/consumer
|
||||
# A 系统跳转地址（人物/组织/事件等列表项打开链接）
|
||||
VITE_AIPI_PRESENT_URL=https://present.a.uat.4.cn/present/domain
|
||||
# B 系统知识库 iframe 根地址
|
||||
VITE_AIPI_KNOWES_URL=https://ch1.b.uat.4.cn/knowes
|
||||
# MaxKey 鉴权（getTokenByOnlineTicket）
|
||||
VITE_AUTH_BASE_URL=https://auth.y.uat.4.cn
|
||||
VITE_AIPI_PGLINKTO_URL=https://pg.b.uat.4.cn
|
||||
VITE_AIPI_XKLINKTO_URL=https://kz.b.uat.4.cn
|
||||
# 工作室/空间前端地址（用于 auth/callback 等跳转）
|
||||
VITE_AIPI_FRONTEND_URL=https://knowledgetoolweb.b.uat.4.cn
|
||||
VITE_AIPI_FRONTEND_URL=https://knowledgetoolserver.b.uat.4.cn
|
||||
|
||||
# 可选：Canvas 独立 LangGraph 地址
|
||||
# VITE_CANVAS_LANGGRAPH_BASE_URL=
|
||||
@@ -3,5 +3,3 @@
|
||||
# VITE_BACKEND_BASE_URL=https://your-intranet-host/deerflow
|
||||
# VITE_LANGGRAPH_BASE_URL=https://your-intranet-host/deerflow/api
|
||||
# VITE_NOTE_API_BASE_URL=https://your-intranet-host/note-api
|
||||
# VITE_AIPI_BACKENP_URL=https://your-notebook-backend-host
|
||||
# VITE_AIPI_FRONTEND_URL=https://your-notebook-frontend-host
|
||||
@@ -1,26 +1,12 @@
|
||||
# SIT 环境打包（pnpm build:sit / npm run build:sit）— 接口地址为 UAT 的 sit 版本
|
||||
# 打包后变量会写入 dist，运行时不可再改
|
||||
|
||||
# 任务看板数据源：true=真实接口（默认）；改成 false 可让任务分析/三情分析两页走本地假数据。
|
||||
VITE_USE_REAL_DATA=true
|
||||
|
||||
# 任务详情（登录深链 cop-task-detail）数据源：默认真实接口；改成 true 走本地假数据（调试用）。
|
||||
VITE_MOCK_COP_TASK_DETAIL=false
|
||||
|
||||
# 研判情报报告（大屏绘制智能体深链 selectSqReport）数据源：默认真实接口；真实接口未就绪可改 true 走本地样例假数据。
|
||||
VITE_MOCK_SQ_REPORT=false
|
||||
|
||||
# 深链任务详情（cop-task-detail）数据源：默认真实接口；真实接口未就绪可改 true 走本地假数据。
|
||||
VITE_MOCK_COP_TASK_DETAIL=false
|
||||
|
||||
# Gateway / DeerFlow API
|
||||
VITE_BACKEND_BASE_URL=https://ch1.b.sit.4.cn/deerflow
|
||||
VITE_LANGGRAPH_BASE_URL=https://ch1.b.sit.4.cn/deerflow/api
|
||||
|
||||
# 轻应用「新建笔记」系列接口
|
||||
VITE_NOTE_API_BASE_URL=https://your-intranet-host/note-api
|
||||
# CMZS后端接口
|
||||
VITE_API_BASE_URL=https://ch1.b.sit.4.cn/zncm
|
||||
# 佳来模型对话接口
|
||||
VITE_AIPI_CHAT_URL=https://ch1.b.sit.4.cn/gpt
|
||||
# 多轮检索接口
|
||||
@@ -33,24 +19,12 @@ VITE_AIPI_B2ACTOR_XWT_PATH=/fromB/toXWT?isFrame=1
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_API_URL=https://ch1.b.sit.4.cn/knowes/api/iframeKnowledgeUrl
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_SKIN=red
|
||||
VITE_AIPI_MOPECHAT_URL=https://ch1.b.sit.4.cn/likun_b1_mope
|
||||
# 工作室/空间后端接口（侧边栏“工作室”列表、空间详情等都走这里，不是 VITE_NOTE_API_BASE_URL）
|
||||
VITE_AIPI_BACKENP_URL=https://knowledgetoolserver.b.sit.4.cn
|
||||
VITE_NOTEBOOK_API_BASE_URL=https://knowledgetoolserver.b.sit.4.cn
|
||||
# 工作室侧边栏显示开关；仅控制入口显示，展开内容仍依赖 notebook 接口成功返回数据
|
||||
VITE_STUDIO_ENABLED=true
|
||||
# 工作室/笔记本前端地址（用于 auth/callback 等跳转）
|
||||
VITE_AIPI_FRONTEND_URL=https://knowledgetoolserver.b.sit.4.cn
|
||||
VITE_AIPI_CONSUMER_URL=https://ch1.b.sit.4.cn/consumer
|
||||
# A 系统跳转地址（人物/组织/事件等列表项打开链接）
|
||||
VITE_AIPI_PRESENT_URL=http://present.a.sit.4.cn/present/domain
|
||||
# B 系统知识库 iframe 根地址
|
||||
VITE_AIPI_KNOWES_URL=https://ch1.b.sit.4.cn/knowes
|
||||
# MaxKey 鉴权（getTokenByOnlineTicket）
|
||||
VITE_AUTH_BASE_URL=https://auth.y.sit.4.cn
|
||||
VITE_AIPI_PGLINKTO_URL=https://pg.b.sit.4.cn
|
||||
VITE_AIPI_XKLINKTO_URL=https://kz.b.sit.4.cn
|
||||
# 工作室/空间前端地址（用于 auth/callback 等跳转）
|
||||
VITE_AIPI_FRONTEND_URL=https://knowledgetoolserver.b.sit.4.cn
|
||||
|
||||
# 可选：Canvas 独立 LangGraph 地址
|
||||
# VITE_CANVAS_LANGGRAPH_BASE_URL=
|
||||
@@ -1,253 +0,0 @@
|
||||
# AI 写作「四阶段内置智能体化」改造方案
|
||||
|
||||
> 目标：把 AI 写作现有的 **素材收集专家 / 写作大纲 / 作家写作 / 编辑审核** 四个功能，
|
||||
> 从「硬编码的 LangGraph 节点 + 硬编码提示词」改造为由**内置功能型智能体**驱动——
|
||||
> 每个阶段对应一个可配置的内置 agent（独立人设 SOUL、可配模型、可挂技能/工具）。
|
||||
|
||||
本文给出现状梳理、目标设计、以及**分步可独立交付的实现计划**。
|
||||
|
||||
---
|
||||
|
||||
## 0. 一句话结论
|
||||
|
||||
不推倒重来。**保留现有图的编排骨架、4 个用户干预点、流式与状态机，只把每个阶段节点的「大脑」从「硬编码 prompt 直调 LLM」替换为「加载对应内置 agent 的 SOUL+配置、在子线程上跑该 agent、按结构化契约解析其输出」。** 这样前端几乎不动，复杂逻辑（严格模式、素材不足求助、逐章流式、技能检索步骤）全部保留。
|
||||
|
||||
---
|
||||
|
||||
## 1. 现状梳理（精确到文件）
|
||||
|
||||
AI 写作是一条 LangGraph 图：`packages/harness/deerflow/agents/ai_writing/graph.py`
|
||||
|
||||
```
|
||||
intent_parser → researcher → (pause 素材确认) → writer_outline → (pause 大纲确认)
|
||||
→ writer_draft → (pause 草稿确认 / 素材不足求助) → editor → (pause 编辑打回) → done
|
||||
```
|
||||
|
||||
四个待改造阶段（均为**硬编码节点**，直接调 LLM、不走 agent 运行时）：
|
||||
|
||||
| 阶段（用户语言） | 节点文件 | 提示词 | 产出（`state.py`） |
|
||||
|---|---|---|---|
|
||||
| 素材收集专家 | `nodes/researcher.py` | `prompts/researcher_prompts.py` | `material_package`（keywords + materials） |
|
||||
| 写作大纲 | `nodes/writer_outline.py` | `prompts/writer_prompts.py` | `current_outline`（Outline JSON） |
|
||||
| 作家写作 | `nodes/writer_draft.py` | `prompts/writer_prompts.py` | `current_draft`（Markdown，逐章流式） |
|
||||
| 编辑审核 | `nodes/editor.py` | `prompts/editor_prompts.py` | `review_result`（评分 + 问题列表 JSON） |
|
||||
|
||||
关键事实：
|
||||
- **LLM 调用方式**：节点通过 `nodes/_utils.py` 的 `get_model()` + `llm_json()` **直接调 LLM**，提示词写死在 `prompts/*.py`。**没有**走 agent 人设 / 技能 / 工具体系。
|
||||
- **模型**：全程单一 `state.model_name`（用户在写作表单里选的主模型），各阶段无法分别配。
|
||||
- **干预点 / 流式 / 严格模式 / 素材不足求助**：都在 `graph.py` 的 pause 节点和 `writer_draft` 里，逻辑较重，需原样保留。
|
||||
- **前端**：`frontend-web/src/open-canvas/contexts/AIWritingContext.tsx` 是状态机（`researching/writing_outline/writing_draft/reviewing/...`），时间线 `AIWritingTimeline.tsx` 按 `progress_events`（已带 `agent_name` 字段）渲染。
|
||||
|
||||
---
|
||||
|
||||
## 2. 可复用的现成机制（不用造轮子）
|
||||
|
||||
代码库里**已经有**一套成熟的「内置功能型智能体」范式——圆桌的 `roundtable-intent / roundtable-recommender / roundtable-coordinator`：
|
||||
|
||||
1. **资源即代码**：`app/gateway/routers/_roundtable_seed_assets/<agent-id>/`
|
||||
- `config.yaml`（`id` / `name` / `description`，可扩展 `model` / `skills` / `tool_groups`）
|
||||
- `SOUL.md`（人设 + 职责 + 工作流 + **结构化输出契约**）
|
||||
2. **种子器**：`app/gateway/routers/_roundtable_seed.py::ensure_roundtable_functional_agents`
|
||||
- 启动时 + 相关路由入口自愈：目录不存在才从资源字节级落盘到 `.deer-flow/agents/<id>/` + DB（**已存在则永不覆盖**，尊重本地编辑）。
|
||||
3. **运行**：在**独立 thread** 上跑该内置 agent（见 `app/gateway/routers/intent.py`，`INTENT_AGENT_ID = "roundtable-intent"`），通过 lead agent 运行时把 SOUL.md 作系统提示，流式经 LangGraph `custom`/`messages` 通道透传。
|
||||
4. **结构化输出约定**：SOUL 规定模型末尾输出固定 **marker + ```json```**（如 `[INTENT_READY]\n```json{...}```），后端用正则解析（`intent.py::_INTENT_READY_PATTERN`）。
|
||||
|
||||
> 我们要做的，本质就是**照搬这套范式，给 AI 写作做 4 个功能型 agent**，并让对应的图节点改成「跑这个 agent + 解析其结构化输出」。
|
||||
|
||||
另有 agents 通用体系可复用：`agents` 表、`/api/agents` CRUD、智能体管理页（配模型/技能/工具）——4 个写作 agent 落盘后天然出现在这里，可被 admin 编辑。
|
||||
|
||||
---
|
||||
|
||||
## 3. 目标设计
|
||||
|
||||
### 3.1 四个内置写作智能体（建议 id / 职责）
|
||||
|
||||
| Agent id | 名称 | 取代节点 | 结构化输出 marker | 产出 schema |
|
||||
|---|---|---|---|---|
|
||||
| `ai-writing-researcher` | 素材收集专家 | `researcher` | `[MATERIALS_READY]` | `MaterialPackage` |
|
||||
| `ai-writing-outliner` | 大纲规划师 | `writer_outline` | `[OUTLINE_READY]` | `Outline` |
|
||||
| `ai-writing-writer` | 作家 | `writer_draft` | `[SECTION_READY]`（逐章） | 章节 Markdown |
|
||||
| `ai-writing-editor` | 编辑 | `editor` | `[REVIEW_READY]` | `ReviewResult` |
|
||||
|
||||
每个 agent 自带：
|
||||
- `config.yaml`：`id / name / description` +（扩展）`model`（该阶段专用模型，可空=回退用户主模型）、`skills`（如 researcher 挂检索技能）、`tool_groups`。
|
||||
- `SOUL.md`：由现有 `prompts/*.py` 内容迁移改写，**追加结构化输出契约**（严格 marker + JSON 格式 + 反例，仿 `roundtable-intent` SOUL 的「硬规则」写法，压制弱模型乱输出）。
|
||||
|
||||
### 3.2 节点改造方式（核心）
|
||||
|
||||
每个阶段节点保留**输入/输出契约不变**（仍读写 `state` 的同一批字段），内部从：
|
||||
|
||||
```
|
||||
# 旧：硬编码 prompt 直调
|
||||
model = get_model(config, state["model_name"])
|
||||
result = llm_json(model, SYSTEM_PROMPT, USER_PROMPT.format(...))
|
||||
```
|
||||
|
||||
改为：
|
||||
|
||||
```
|
||||
# 新：跑对应内置 agent，按 marker 解析
|
||||
result = await run_writing_agent(
|
||||
agent_id="ai-writing-outliner",
|
||||
state=state, config=config,
|
||||
user_payload=<本阶段输入>, # 意图/素材/大纲/草稿等拼成的一段输入
|
||||
marker="[OUTLINE_READY]", # 结构化输出契约
|
||||
)
|
||||
```
|
||||
|
||||
新增统一助手 `nodes/_agent_runner.py::run_writing_agent(...)`：在子 thread 上跑指定内置 agent（复用 `intent.py` 的 run/stream 机制）、把流式 chunk 透传到 `custom` 通道（前端时间线不回归）、用复用的 `_utils.parse_json` 多级容错解析 marker 后的 JSON。
|
||||
|
||||
### 3.3 为什么选这个方案（A）而非全多智能体重写（B）
|
||||
|
||||
- **方案 A（推荐）**：保留图骨架，逐节点替换大脑。前端基本不动；干预点/流式/严格模式/素材求助原样保留；可**按阶段灰度 + feature flag 回退**，风险可控。
|
||||
- **方案 B（不推荐）**：仿圆桌重做成多 agent 编排，需重写编排、状态、流式、4 个干预点对接，工作量数倍、回归面巨大，收益（更"纯"的多 agent）对本场景并不必要。
|
||||
|
||||
---
|
||||
|
||||
## 4. 配置与可管理性
|
||||
|
||||
- **落盘**：仿 roundtable 新增 `_ai_writing_seed_assets/` + 一个 `ensure_ai_writing_functional_agents()` 种子器，在 Gateway 启动与 ai_writing 会话入口自愈。
|
||||
- **管理界面**：4 个 agent 落盘后即出现在智能体管理页，admin 可编辑 **SOUL（人设）/ 模型 / 挂载技能**。可选：在「文章类型配置页」旁加一个「写作智能体配置」入口，或给这 4 个打一个 `tag` 便于筛选。
|
||||
- **模型优先级**：节点跑 agent 时 `该 agent config.model > 用户写作表单选的主模型 model_name`（保留现有"用户选模型"语义，新增 per-阶段覆盖能力）。
|
||||
- **技能**：researcher 现有的技能检索能力（`search/skill_search.py`）改为 `ai-writing-researcher` 的**挂载技能**，admin 可增减；检索步骤事件（`SkillStep`）形态保留给前端展示。
|
||||
|
||||
---
|
||||
|
||||
## 5. 分步实现计划（每步可独立交付 / 验收）
|
||||
|
||||
### Phase 0 · 契约冻结
|
||||
- 定 4 个 agent 的 `id / name`、每个的 **marker + JSON schema**（与 `state.py` 的 `MaterialPackage/Outline/DraftArticle/ReviewResult` 严格对齐）。
|
||||
- **验收**：本文 §3.1 / §6 的契约评审通过、字段与现有 TypedDict 一一对应。
|
||||
|
||||
### Phase 1 · 内置 agent 种子化（后端，纯新增，不接管流程）
|
||||
- 新建 `app/gateway/routers/_ai_writing_seed_assets/<4 个 agent>/{config.yaml, SOUL.md}`；SOUL 由现有 `prompts/*.py` 迁移改写 + 追加结构化输出契约。
|
||||
- 仿 `_roundtable_seed.py` 写 `ensure_ai_writing_functional_agents()`，挂到启动 lifespan + ai_writing 入口。
|
||||
- **验收**：启动后 4 个 agent 出现在 `.deer-flow/agents/`、DB、`/api/agents`，能在智能体管理页查看/编辑；**此阶段不改图，写作流程行为不变**。
|
||||
|
||||
### Phase 2 · 节点接入「跑内置 agent」（后端，逐阶段灰度）
|
||||
- 新增 `nodes/_agent_runner.py::run_writing_agent(...)`（子线程跑 agent + 流式透传 + marker 解析）。
|
||||
- 加 feature flag：`config.yaml → ai_writing.use_builtin_agents.{editor,outliner,researcher,writer}`（默认 false=走旧节点）。
|
||||
- **接入顺序（由易到难）**：
|
||||
1. **editor**（最简单：小 JSON、无检索、无逐章流式）→ flag 打开端到端验证。
|
||||
2. **outliner**（输出 Outline JSON）。
|
||||
3. **researcher**（保留技能检索步骤事件、素材确认暂停）。
|
||||
4. **writer**（最难：保留**逐章流式** + **严格模式素材不足求助** `blocked_sections`；建议**仍由节点控制"逐章循环"**，每章调一次 `ai-writing-writer`，`section_help` 暂停逻辑不动）。
|
||||
- 每接一个，旧 prompt 与旧逻辑**保留**，节点按 flag 二选一。
|
||||
- **验收**：每个阶段 flag 打开后端到端走通，产出结构与旧版一致，干预点 / 流式 / 严格模式 / 求助暂停均不回归。
|
||||
|
||||
### Phase 3 · 每阶段 agent 配置化
|
||||
- per-agent 模型覆盖生效（`agent.config.model > model_name`）。
|
||||
- researcher 的检索能力改为挂载技能，admin 可增减。
|
||||
- 智能体管理页可编辑这 4 个的 SOUL / 模型 / 技能；config 热加载，新会话即生效、不重启前端。
|
||||
- **验收**：admin 改某阶段 agent 的模型 / SOUL / 技能后，新会话立即按新配置运行。
|
||||
|
||||
### Phase 4 · 前端适配（小改）
|
||||
- 时间线 / 状态文案：阶段名读对应 agent 的 `name`（`progress_events.agent_name` 已存在，基本沿用，去掉写死文案）。
|
||||
- 可选：写作表单 / 设置加「查看·配置写作智能体」入口，跳智能体管理页。
|
||||
- **验收**：前端展示各阶段对应的智能体名；配置入口可达。
|
||||
|
||||
### Phase 5 · 清理收口
|
||||
- 四阶段稳定后，默认打开全部 flag、移除旧 prompt 直调路径与 flag（prompts 历史留在 git）。
|
||||
- 更新 `offline-backend-20260512/backend/CLAUDE.md` 架构描述。
|
||||
- **验收**：回归测试通过、冗余删除、文档同步。
|
||||
|
||||
---
|
||||
|
||||
## 6. 结构化输出契约（与现有 state 对齐）
|
||||
|
||||
各 agent SOUL 末尾必须输出 `marker + ```json``` 块，schema 直接对齐 `state.py`：
|
||||
|
||||
- `ai-writing-researcher` → `[MATERIALS_READY]`：`{ "keywords": string[], "materials": Material[] }`（`Material` 见 `state.py`，含 `id/title/content/source/relevance_score/source_type/...`）。
|
||||
- `ai-writing-outliner` → `[OUTLINE_READY]`：`Outline = { "title": string, "sections": OutlineSection[] }`（`OutlineSection = { section_title, key_points: KeyPoint[], material_ids }`）。
|
||||
- `ai-writing-writer` → 逐章 `[SECTION_READY]`：`{ "section_title": string, "markdown": string, "needs_more_material"?: { "reason": string } }`（`needs_more_material` 触发严格模式求助 `blocked_sections`）。
|
||||
- `ai-writing-editor` → `[REVIEW_READY]`：`ReviewResult = { "verdict": "pass"|"reject", "fact_score", "logic_score", "language_score", "overall_score", "pass_threshold", "revision_notes": ReviewIssue[] }`。
|
||||
|
||||
解析复用 `nodes/_utils.py::parse_json`（多级容错），marker 正则仿 `intent.py::_INTENT_READY_PATTERN`。
|
||||
|
||||
---
|
||||
|
||||
## 7. 风险与注意
|
||||
|
||||
1. **弱模型结构化输出不稳**：SOUL 写严格格式 + 反例 + 「输出 marker 后立刻闭嘴」硬规则（仿 `roundtable-intent` SOUL §0）；解析层复用 `parse_json` 多级降级。
|
||||
2. **writer 的逐章流式 + 严格模式求助最复杂**：建议**循环控制权留在节点**（每章调一次 agent），不要让 agent 一次吐全文，否则 `blocked_sections` / `section_help` / 逐章流式都要重做。
|
||||
3. **researcher 技能检索步骤事件**（`SkillStep`）前端有专门卡片：迁移时保留这些 `progress_events` 形态。
|
||||
4. **种子器 never-overwrite**：迭代 SOUL 后**老部署不会自动更新**（与圆桌同一问题）。运维更新办法：删 `.deer-flow/agents/<id>/` 重新种子，或后续给种子器加「内置功能 agent 按版本/hash 覆盖」开关。
|
||||
5. **子线程隔离**：跑写作 agent 的子 thread 与主 ai_writing thread 状态隔离，注意 checkpointer 与清理（沿用 ai_writing 既有的会话清理）。
|
||||
6. **离线内网**：researcher 的检索一律走技能（不联网），与现有「离线部署禁 web_search、走 skill」策略一致。
|
||||
|
||||
---
|
||||
|
||||
## 8. 涉及文件清单（落地时对照）
|
||||
|
||||
**后端（主要改动）**
|
||||
- 新增 `app/gateway/routers/_ai_writing_seed_assets/{ai-writing-researcher,ai-writing-outliner,ai-writing-writer,ai-writing-editor}/{config.yaml,SOUL.md}`
|
||||
- 新增 `app/gateway/routers/_ai_writing_seed.py`（种子器，仿 `_roundtable_seed.py`）
|
||||
- 新增 `packages/harness/deerflow/agents/ai_writing/nodes/_agent_runner.py`（`run_writing_agent`）
|
||||
- 改 `nodes/{researcher,writer_outline,writer_draft,editor}.py`（按 flag 二选一）
|
||||
- 改 `config.yaml`（新增 `ai_writing.use_builtin_agents.*` 开关；可选 per-agent 默认模型）
|
||||
- 保留 `prompts/*.py`（回退用，Phase 5 删）
|
||||
|
||||
**前端（小改）**
|
||||
- 改 `src/open-canvas/components/ai-writing/AIWritingTimeline.tsx`（阶段名读 agent name）
|
||||
- 可选：写作表单/设置加「配置写作智能体」入口（跳 `/page/...` 智能体管理页）
|
||||
|
||||
**文档**
|
||||
- 本文件
|
||||
- Phase 5 收口时更新 `offline-backend-20260512/backend/CLAUDE.md`
|
||||
|
||||
---
|
||||
|
||||
## 9. 实施记录（2026-06-11，Phase 1-4 已落地）
|
||||
|
||||
### 已交付
|
||||
|
||||
- **Phase 1**：`_ai_writing_seed_assets/{ai-writing-researcher,ai-writing-outliner,ai-writing-writer,ai-writing-editor}/{config.yaml,SOUL.md}` + `_ai_writing_seed.py::ensure_ai_writing_functional_agents()`，挂启动 lifespan（`_sync_legacy_agents` 之前）+ `ai_writing.list_sessions / list_article_types` 入口自愈。
|
||||
- **Phase 2**：`nodes/_agent_runner.py`（`stream_writing_agent` / `run_writing_agent` / `parse_marker_json` / `resolve_agent_model_name` / `use_builtin_agent`）；四节点按 `config.yaml → ai_writing.use_builtin_agents.{researcher,outliner,writer,editor}` 二选一（默认全 false，config 按 mtime 热加载，改完即对新会话生效）。agent 目录缺失时节点**静默回退旧 prompt 路径**，写作流程永不因 agent 被删而中断。
|
||||
- **Phase 3（部分）**：per-agent 模型覆盖已生效（优先级 `agent config.model > 节点显式回退模型（keyword_model/rank_model）> 表单主模型 model_name`）；SOUL/模型在智能体管理页可编辑、按 mtime 热加载。researcher「检索能力改挂载技能」未做（仍走节点内 SkillPicker 编排）。
|
||||
- **Phase 4**：时间线 agent 名由后端 `progress_events.agent_name` 动态下发（flag 开启时读 agent config 的 `name`）；`TimelineStep.tsx` 补「大纲规划师」配色（未命中名字回退灰色，admin 改名不报错）。
|
||||
- **测试**：`tests/test_ai_writing_seed.py`（资源完整性 / 契约 marker / never-overwrite）+ `tests/test_ai_writing_agent_runner.py`（marker 多级解析 / 模型优先级 / flag / 流式端到端 / SOUL 缺失降级），共 31 例。
|
||||
|
||||
### 与 §6 契约的修订（以实现为准）
|
||||
|
||||
为保住「逐章流式打字机」「review_chunk/review_item_done 分维度卡片」「素材确认/求助暂停」等前端形态，编排控制权全部留在节点，agent 只承担**单次结构化产出**：
|
||||
|
||||
| Agent | 实际契约 |
|
||||
|---|---|
|
||||
| `ai-writing-researcher` | 两个子任务：`[KEYWORDS_READY]` + `{"keywords": []}`；`[MATERIALS_RANKED]` + `{"ranked_ids": [], "summary": ""}`。素材本体来自节点的检索编排（agent 不联网），`MaterialPackage` 仍由节点组装 |
|
||||
| `ai-writing-outliner` | `[OUTLINE_READY]` + 完整 `Outline` JSON（新规划 / 修改模式共用，payload 开头标明任务模式） |
|
||||
| `ai-writing-writer` | **无 JSON marker**：每章直接输出 Markdown 正文（保流式）；严格模式素材不足只输出一行 `[[SECTION_BLOCKED]] 原因`（沿用原哨兵） |
|
||||
| `ai-writing-editor` | 每个维度一次调用：`[REVIEW_READY]` + `{"score": int, "issues": []}`；综合分/verdict/pass_threshold 仍由节点按 config 计算（不信任 LLM 自判通过） |
|
||||
|
||||
另一处架构修订：**不在子 thread 上经 lead agent 运行时跑**（§3.2 原设想复用 `intent.py` 机制，但其 loopback 在 app 层，harness 图节点不可 import；且单轮结构化调用无需 thread/checkpointer/中间件成本）。`run_writing_agent` 直接「SOUL 作 system prompt + 单条输入」调模型，规避了 §7.5 的子线程隔离/清理风险。researcher 挂技能（Phase 3 剩余项）如需完整工具体系再评估接入运行时。
|
||||
|
||||
### 已知存量测试失败（与本次无关）
|
||||
|
||||
`test_ai_writing_graph.py`（mock 已删除的 `intent_parser.llm_json`）、`test_ai_writing_word_count.py`（mock 已删除的 `writer_draft.get_model`）、`test_ai_writing_draft_headings.py`（断言三级结构升级前的 `####→###` 行为）、`test_ai_writing_outline_revise.py`（断言旧 prompt 文案），改造前即失败，待另行清理。
|
||||
|
||||
## 10. Phase 3 收尾：检索类型两项化 + 配置技能驱动检索（2026-06-13 已实施）
|
||||
|
||||
本节落地了 §9 遗留的「researcher 检索改挂载技能」，并按新需求重新设计了检索类型与 PAUSE-1 交互。
|
||||
|
||||
### 10.1 检索类型（material_source）收敛为两项
|
||||
|
||||
- 写作表单「素材来源」改名「检索类型」，只剩 `general`（通用检索，默认）/ `notebook`（我的空间）；旧值 `knowledge_base` / `skill` 由 researcher 节点按 `general` 兼容（老会话 resume 不受影响）。
|
||||
- 前端：`AIWritingForm.tsx`（选项/标签）、`WritingFormContext.tsx`（默认值/类型/applyProposal 旧值映射）、`types/ai-writing.ts`、`message-list.tsx`（审批卡标签兼容旧值）。
|
||||
|
||||
### 10.2 通用检索 = 素材收集专家配置的技能编排
|
||||
|
||||
- `ai-writing-researcher` agent 的 `config.yaml → skills` 是权威配置（智能体管理页可增删/替换）；种子默认 `[knowledge-base-search]`。
|
||||
- 内置技能 **knowledge-base-search（知识库检索）**：把原 web/intranet 关键词并发检索包装成技能。researcher 对它特殊处理——不拼 description，直接对 Step1 生成的检索词做原生并发检索（行为与旧 knowledge_base 模式一致）。其它配置技能按 `意图 + description` 拼定向 query 检索。按配置顺序执行、累积达 `max_materials` 早停（`stop_early`）。
|
||||
- 技能配置为空/全部失效 → 内存兜底回退知识库检索（`researcher._load_configured_skills`），检索永不瘫痪。
|
||||
- 种子：`_ai_writing_seed_assets/skills/knowledge-base-search/SKILL.md` → `skills/public/`（never-overwrite）；旧部署的 researcher `config.yaml` 缺 `skills` 键时就地补默认（显式 `skills: []` 不动）。`_ai_writing_seed.py::ensure_ai_writing_functional_agents()` 一并处理。
|
||||
|
||||
### 10.3 PAUSE-1 输入框 → 匹配技能检索并**追加**
|
||||
|
||||
- 用户在素材确认卡输入自然语言（如「检索 A 的数据」）→ `re_search + user_query`：配置技能 >1 个时 `SkillPicker(candidates=配置技能)` 让模型按输入匹配排序（流式打字机保留），单技能直接用；用用户原话（非知识库技能附 description）检索。
|
||||
- **追加语义**：保留原素材列表与顺序，新素材按 URL 去重后追加在末尾（ID 续号），跳过 LLM 重排与 max_materials 截断（仅 100 条硬上限），summary 显示「追加检索到 X 条新素材，当前共 Y 条」。原 20 条 + 新 20 条 = 40 条。
|
||||
- 「技能检索」独立按钮已删除（通用检索本来就走技能编排）；后端继续兼容旧 `enable_skill_search` action（按 user_query 追加处理）。无输入的「重新搜索」仍为覆盖式重检。
|
||||
- 前端：`InterventionCard.tsx`（按钮/文案，hasQuery 时按钮文案「检索并追加素材」）、`MaterialsToolbar.tsx`（提示语）、`useAIWritingStream.ts`（素材卡回到 researching 一律置 skillSearchPending，让 SkillStepsCard 占位即时挂载）。
|
||||
|
||||
### 10.4 测试
|
||||
|
||||
- `tests/test_ai_writing_material_source.py` 重写（9 例）：缺省/旧值→通用检索、自定义技能 description query、user_query 追加+URL 去重+多技能 picker 匹配（candidates 断言）、无输入重检覆盖、空配置回退。
|
||||
- `tests/test_ai_writing_seed.py` 扩展（+5 例）：技能资产存在、researcher 种子含默认 skills、技能种子 never-overwrite、缺 skills 键就地补丁、显式 `[]` 不动。
|
||||
@@ -0,0 +1,228 @@
|
||||
# AI 写作 · 右侧文本编辑器（滑词 / 段落菜单）
|
||||
|
||||
> 说明 AI 写作页、开放画布中文稿右侧编辑器的实现方式：**编辑器本体基于 Tiptap v3**；选中文本浮层（「滑词」）与段落悬停按钮为**自研 React 组件**，未使用 `@tiptap/extension-bubble-menu` 官方 BubbleMenu。
|
||||
|
||||
---
|
||||
|
||||
## 一、功能概览
|
||||
|
||||
| 交互 | 用户感知 | 组件 | 触发方式 |
|
||||
|------|----------|------|----------|
|
||||
| **滑词** | 拖选一段文字后，选区上方出现工具条（AI 改写、加粗、斜体、下划线、H1、复制） | `WritingBubbleMenu` | Tiptap `selectionUpdate` + 浏览器 `window.getSelection()` |
|
||||
| **段落菜单** | 鼠标移入某段/标题/引用块时，段落左侧出现 ✨ 按钮，点击可 AI 改写整段 | `WritingParagraphMenu` | 编辑器 DOM `mousemove` + ProseMirror `posAtCoords` |
|
||||
| **AI 改写面板** | 点击「AI 改写」后弹出润色/扩写等操作与预览 | `WritingAiMenu` / `WritingRewritePreview` | 由 `TextArtifactRenderer` + `useWritingRewrite` 串联 |
|
||||
|
||||
画布线程页（`/page/canvas/:thread_id`）与 AI 写作页（`/page/workspace/ai-writing`）右侧草稿**共用同一套** `TextArtifactRenderer` 与菜单组件。
|
||||
|
||||
---
|
||||
|
||||
## 二、在 AI 写作中的挂载路径
|
||||
|
||||
```mermaid
|
||||
flowchart TB
|
||||
AIWritingPage --> CanvasResizableLayout
|
||||
CanvasResizableLayout --> AIWritingDraftPanel
|
||||
AIWritingDraftPanel --> DraftStreamingPreview
|
||||
AIWritingDraftPanel --> TextArtifactRenderer
|
||||
DraftStreamingPreview["DraftStreamingPreview<br/>（流式只读 Markdown）"]
|
||||
TextArtifactRenderer --> WritingBubbleMenu
|
||||
TextArtifactRenderer --> WritingParagraphMenu
|
||||
TextArtifactRenderer --> TiptapEditor
|
||||
TextArtifactRenderer --> WritingAiMenu
|
||||
```
|
||||
|
||||
| 阶段 | 右侧展示 | 滑词 / 段落菜单 |
|
||||
|------|----------|-----------------|
|
||||
| 草稿流式生成中（`writing_draft` / `revising` 且有 section 流） | `DraftStreamingPreview`（ReactMarkdown 只读） | **不可用** |
|
||||
| 成稿可编辑（`isDraftEditable === true`） | `TextArtifactRenderer` + Tiptap | **可用** |
|
||||
| 流式结束但尚未 editable | `TextArtifactRenderer` 且 `isStreaming={true}` | **不可用**（`editable: false`） |
|
||||
|
||||
相关状态由 [CanvasUIContext.tsx](../src/open-canvas/contexts/CanvasUIContext.tsx) 的 `isDraftStreaming`、`isDraftEditable` 控制；注释中「禁用编辑/润色/滑词」即指生成完成前不进入可编辑渲染器。
|
||||
|
||||
入口文件：
|
||||
|
||||
- [AIWritingDraftPanel.tsx](../src/open-canvas/components/ai-writing/AIWritingDraftPanel.tsx) — 决定流式预览 vs `TextArtifactRenderer`
|
||||
- [TextArtifactRenderer.tsx](../src/open-canvas/components/artifact/TextArtifactRenderer.tsx) — 挂载编辑器与两套浮层菜单
|
||||
|
||||
---
|
||||
|
||||
## 三、技术栈：编辑器 vs 浮层
|
||||
|
||||
### 3.1 Tiptap 编辑器（有使用）
|
||||
|
||||
| 包 | 用途 |
|
||||
|----|------|
|
||||
| `@tiptap/react` | `useEditor`、`EditorContent` |
|
||||
| `@tiptap/starter-kit` | 段落、标题、加粗、斜体、下划线等 |
|
||||
| `@tiptap/extension-placeholder` | 空文档占位符 |
|
||||
| `tiptap-markdown` | Markdown ↔ 文档互转 |
|
||||
| 自研 `AiHighlightExtension` | AI 改写前高亮选区（ProseMirror Decoration） |
|
||||
|
||||
扩展注册见 [TiptapExtensions.ts](../src/open-canvas/editor/TiptapExtensions.ts)；实例创建见 [useWritingEditor.ts](../src/open-canvas/hooks/useWritingEditor.ts)。
|
||||
|
||||
### 3.2 浮层菜单（自研，非 Tiptap 官方插件）
|
||||
|
||||
| 包 | 是否用于滑词/段落菜单 |
|
||||
|----|----------------------|
|
||||
| `@tiptap/extension-bubble-menu` | **未使用**（`package.json` 中有依赖，代码未 import） |
|
||||
| `@tiptap/react` 的 `<BubbleMenu>` | **未使用** |
|
||||
|
||||
两套菜单均采用：
|
||||
|
||||
1. 监听 Tiptap / DOM 事件更新位置与选区；
|
||||
2. `createPortal(..., document.body)` 挂到 `document.body`，避免被右侧滚动容器 `overflow` 裁切；
|
||||
3. `position: fixed` + `getBoundingClientRect()` 定位。
|
||||
|
||||
命名上的 `WritingBubbleMenu` 仅表示「气泡式选区工具条」，实现上与 Tiptap 官方 Bubble Menu 扩展无关。
|
||||
|
||||
---
|
||||
|
||||
## 四、核心文件索引
|
||||
|
||||
| 文件 | 职责 |
|
||||
|------|------|
|
||||
| [TiptapEditor.tsx](../src/open-canvas/editor/TiptapEditor.tsx) | `EditorContent` 渲染容器 |
|
||||
| [TiptapExtensions.ts](../src/open-canvas/editor/TiptapExtensions.ts) | StarterKit、Placeholder、Markdown、AiHighlight |
|
||||
| [AiHighlightExtension.ts](../src/open-canvas/editor/AiHighlightExtension.ts) | `setAiHighlight` / `clearAiHighlight` 命令 |
|
||||
| [useWritingEditor.ts](../src/open-canvas/hooks/useWritingEditor.ts) | 创建 editor、同步外部 markdown、editable |
|
||||
| [WritingBubbleMenu.tsx](../src/open-canvas/components/artifact/WritingBubbleMenu.tsx) | **滑词** 选区浮层 |
|
||||
| [WritingParagraphMenu.tsx](../src/open-canvas/components/artifact/WritingParagraphMenu.tsx) | **段落** 左侧 ✨ 按钮 |
|
||||
| [WritingAiMenu.tsx](../src/open-canvas/components/artifact/WritingAiMenu.tsx) | AI 改写操作菜单 |
|
||||
| [WritingRewritePreview.tsx](../src/open-canvas/components/artifact/WritingRewritePreview.tsx) | 改写结果预览 |
|
||||
| [useWritingRewrite.ts](../src/open-canvas/hooks/useWritingRewrite.ts) | 调用润色 API、写回编辑器 |
|
||||
| [apply-rewrite.ts](../src/open-canvas/utils/apply-rewrite.ts) | 将改写结果应用到 ProseMirror 选区 |
|
||||
| [CanvasUIContext.tsx](../src/open-canvas/contexts/CanvasUIContext.tsx) | `writingSelection` 全局状态 |
|
||||
| [selection.ts](../src/open-canvas/types/selection.ts) | `WritingSelection` 类型定义 |
|
||||
|
||||
---
|
||||
|
||||
## 五、滑词：`WritingBubbleMenu`
|
||||
|
||||
### 5.1 行为
|
||||
|
||||
- 用户在编辑器内拖选非空文本时显示浮层；选区为空或位于 `codeBlock` 内时隐藏。
|
||||
- 浮层按钮：`AI 改写`、`B` / `I` / `U`、`H1`、复制。
|
||||
- 按钮 `onMouseDown` 使用 `e.preventDefault()`，避免点击时编辑器失焦导致选区丢失。
|
||||
|
||||
### 5.2 实现要点
|
||||
|
||||
```text
|
||||
editor.on("selectionUpdate")
|
||||
→ 读取 editor.state.selection (from, to, empty)
|
||||
→ window.getSelection() + Range.getBoundingClientRect() 算屏幕坐标
|
||||
→ setWritingSelection({ from, to, text }) 写入 CanvasUIContext
|
||||
→ createPortal 渲染工具条，useLayoutEffect 做贴边 / 上方空间不足时翻到选区下方
|
||||
```
|
||||
|
||||
关键代码位置：[WritingBubbleMenu.tsx](../src/open-canvas/components/artifact/WritingBubbleMenu.tsx) 中 `useEffect` 监听 `selectionUpdate` 与 `blur`。
|
||||
|
||||
点击「AI 改写」时：
|
||||
|
||||
1. `editor.commands.setAiHighlight(from, to)` 高亮选区；
|
||||
2. 调用父组件传入的 `onAiRewrite()`，由 `TextArtifactRenderer` 打开 `WritingAiMenu`。
|
||||
|
||||
---
|
||||
|
||||
## 六、段落菜单：`WritingParagraphMenu`
|
||||
|
||||
### 6.1 行为
|
||||
|
||||
- 鼠标在编辑器内移动时，检测光标下的**块级节点**；
|
||||
- 仅对 `paragraph`、`heading`、`blockquote` 且文本非空时，在块左侧显示 ✨ 按钮；
|
||||
- 鼠标离开编辑器或移出可改写块后，延迟约 **380ms** 隐藏（`HIDE_DELAY`），并在按钮与正文之间留有 hover 桥接区，避免移向按钮时闪烁消失。
|
||||
|
||||
### 6.2 实现要点
|
||||
|
||||
```text
|
||||
editor.view.dom mousemove
|
||||
→ editor.view.posAtCoords({ left, top })
|
||||
→ doc.resolve(pos).before(depth) 定位块起点
|
||||
→ editor.view.nodeDOM(nodeStart) 取 DOM 矩形
|
||||
→ createPortal 固定定位按钮在块左侧 (BTN_SIZE=24, BTN_GAP=6)
|
||||
```
|
||||
|
||||
点击 ✨ 时：
|
||||
|
||||
1. `setTextSelection` 选中整段 ProseMirror 范围（`nodeFrom` ~ `nodeTo`）；
|
||||
2. `setAiHighlight` + `setWritingSelection`；
|
||||
3. 触发与滑词相同的 `onAiRewrite()` 流程。
|
||||
|
||||
关键代码位置：[WritingParagraphMenu.tsx](../src/open-canvas/components/artifact/WritingParagraphMenu.tsx)。
|
||||
|
||||
---
|
||||
|
||||
## 七、AI 改写数据流
|
||||
|
||||
```mermaid
|
||||
sequenceDiagram
|
||||
participant User
|
||||
participant Menu as BubbleMenu / ParagraphMenu
|
||||
participant Ctx as CanvasUIContext
|
||||
participant Renderer as TextArtifactRenderer
|
||||
participant API as rewrite API
|
||||
|
||||
User->>Menu: 选中文本或点段落 ✨
|
||||
Menu->>Ctx: setWritingSelection
|
||||
Menu->>Renderer: onAiRewrite()
|
||||
Renderer->>Renderer: phase = menu, WritingAiMenu
|
||||
User->>Renderer: 选择润色/扩写等
|
||||
Renderer->>API: useWritingRewrite.rewrite
|
||||
API-->>Renderer: 流式结果
|
||||
Renderer->>Renderer: WritingRewritePreview
|
||||
User->>Renderer: 接受改写
|
||||
Renderer->>Renderer: applyReplace / applyInsert
|
||||
```
|
||||
|
||||
- 选区信息类型：`WritingSelection`（`from`、`to`、`text`），见 [selection.ts](../src/open-canvas/types/selection.ts)。
|
||||
- 与画布聊天侧选区 `CanvasTextSelection` 分离；后者用于快捷操作栏等，不用于 Tiptap 改写。
|
||||
|
||||
---
|
||||
|
||||
## 八、与顶部工具栏、目录栏的关系
|
||||
|
||||
同一 `TextArtifactRenderer` 编辑模式下还包含：
|
||||
|
||||
| 组件 | 作用 |
|
||||
|------|------|
|
||||
| [WritingToolbar.tsx](../src/open-canvas/components/artifact/WritingToolbar.tsx) | 顶部固定格式栏、导出等 |
|
||||
| [DraftOutlineSidebar.tsx](../src/open-canvas/components/artifact/DraftOutlineSidebar.tsx) | 左侧标题目录，点击滚动到对应 heading |
|
||||
|
||||
滑词 / 段落菜单与工具栏互不替代：工具栏面向全局格式；浮层面向**当前选区或当前段落**的快捷操作与 AI 改写入口。
|
||||
|
||||
---
|
||||
|
||||
## 九、开发与改造建议
|
||||
|
||||
### 9.1 若要改用官方 Bubble Menu
|
||||
|
||||
1. 在 [TiptapExtensions.ts](../src/open-canvas/editor/TiptapExtensions.ts) 或组件层引入 `@tiptap/react` 的 `BubbleMenu`（或注册 `@tiptap/extension-bubble-menu`）；
|
||||
2. 将 [WritingBubbleMenu.tsx](../src/open-canvas/components/artifact/WritingBubbleMenu.tsx) 内手写 `selectionUpdate` + Portal 逻辑替换为官方组件的 `shouldShow` / `tippyOptions`（v3 API 以官方文档为准）；
|
||||
3. **段落左侧按钮**仍建议保留自研 `WritingParagraphMenu`（官方 Bubble Menu 只覆盖选区，不覆盖 hover 段落场景）。
|
||||
|
||||
### 9.2 新增浮层按钮
|
||||
|
||||
- **滑词**：在 `WritingBubbleMenu` 的 `BubbleButton` 列表中增加，使用 `editor.chain().focus()....run()`。
|
||||
- **段落**：在 `WritingParagraphMenu` 的 `handleClick` 或拆出多按钮布局；注意 `REWRITABLE_TYPES` 与 hover 延迟逻辑。
|
||||
|
||||
### 9.3 流式阶段禁止交互
|
||||
|
||||
不要在 `DraftStreamingPreview` 中挂载 Tiptap 菜单：流式阶段 token 高频更新，挂载 `useWritingEditor` 易引发更新风暴（见 [DraftStreamingPreview.tsx](../src/open-canvas/components/ai-writing/DraftStreamingPreview.tsx) 文件头注释）。应等 `isDraftEditable` 后再走 `TextArtifactRenderer`。
|
||||
|
||||
---
|
||||
|
||||
## 十、相关文档
|
||||
|
||||
- [前端开发.md](./前端开发.md) §25.4 — Canvas / AI 写作路由与组件总表
|
||||
- [ai-writing-用户自带大纲-测试用例.md](./ai-writing-用户自带大纲-测试用例.md) — AI 写作业务流程测试（含右侧草稿层级校验）
|
||||
|
||||
---
|
||||
|
||||
## 十一、依赖版本（摘录）
|
||||
|
||||
以 [package.json](../package.json) 为准，编辑器相关主要为：
|
||||
|
||||
- `@tiptap/core`、`@tiptap/react`、`@tiptap/starter-kit`：`^3.23.4`
|
||||
- `@tiptap/extension-bubble-menu`：已安装但未接入滑词实现
|
||||
- `tiptap-markdown`：Markdown 扩展
|
||||
|
||||
升级 Tiptap 大版本前，请对照官方迁移指南，并重点回归：`WritingBubbleMenu` 选区坐标、`WritingParagraphMenu` 的 `posAtCoords`、以及 `AiHighlightExtension` 与 StarterKit 扩展名冲突（勿重复注册 `underline`）。
|
||||
@@ -1,507 +0,0 @@
|
||||
# AI 写作「对话驱动写作台」实现方案（可落地版）
|
||||
|
||||
> 需求：**复制**现有 AI 写作代码，新增一个独立页面（**原页面/原路由零改动**）。在新页面里
|
||||
> **大改交互流程**——去掉「素材收集 / 大纲规划 / 作家写作 / 编辑审核」四个阶段干预卡里的**输入框**，
|
||||
> 只**保留原来的确认按钮**；在页面**底部新增一个统一的多行智能输入框**，具备**强意图理解**能力：
|
||||
> 重新搜索、新增检索词、添加检索素材、重新规划大纲、按要求改大纲、提修改意见、审核意见、
|
||||
> 素材不足继续搜集、对写作内容提问……**覆盖现有全部交互甚至更多**。
|
||||
>
|
||||
> 交互节奏不变：**一步一确认**——进行到某个专家阶段时，交互就针对该专家（阶段作用域）；
|
||||
> 用户可以**点原来的确认按钮**，也可以**在底部输入框里发「继续下一步」的意图**推进。
|
||||
> 意图**拿不准时**主动向用户**发起协助/澄清**，绝不盲目触发高代价动作（定稿 / 重搜等）。
|
||||
|
||||
本文是**实现指引**，落地时按第 9 节路线图分阶段推进。文中所有文件路径、符号名均对齐当前代码。
|
||||
|
||||
---
|
||||
|
||||
## 1. 现状梳理（改造地基）
|
||||
|
||||
现有 AI 写作是一套**独立 LangGraph 子系统**（不是 lead_agent），前后端都已成型、可复用。
|
||||
|
||||
### 1.1 后端 `ai_writing` graph
|
||||
|
||||
- 图工厂：`offline-backend-20260512/backend/packages/harness/deerflow/agents/ai_writing/graph.py`
|
||||
（`make_ai_writing_graph`，registered as `assistant_id="ai_writing"`）
|
||||
- 状态：`.../ai_writing/state.py`（`AIWritingState`）
|
||||
- 四个「阶段大脑」节点：
|
||||
|
||||
| 阶段 | 节点 | 文件 |
|
||||
|------|------|------|
|
||||
| 素材收集 | `researcher` | `nodes/researcher.py` |
|
||||
| 大纲规划 | `writer_outline` | `nodes/writer_outline.py` |
|
||||
| 作家写作 | `writer_draft` | `nodes/writer_draft.py` |
|
||||
| 编辑审核 | `editor` | `nodes/editor.py` |
|
||||
|
||||
- **五个暂停点**（`interrupt()`，`Command(resume=payload)` 续跑）——这是对话改造的核心可复用资产：
|
||||
|
||||
| pause_point | 语义 | 现有 action |
|
||||
|-------------|------|-------------|
|
||||
| `material_confirm` | 素材确认 | `confirm` / `re_search` / `enable_skill_search` |
|
||||
| `outline_confirm` | 大纲确认 | `confirm` / `re_outline` |
|
||||
| `draft_confirm` | 草稿确认 | `to_editor` / `finalize` / `user_revise` |
|
||||
| `review_confirm` | 编辑打回确认 | `accept_review` / `force_finalize` |
|
||||
| `section_help` | 素材不足求助 | `section_help`(+`sectionDecisions`) |
|
||||
|
||||
- resume 载荷（`UserInterventionPayload`，`types/ai-writing.ts`）字段：`action`、`approvedMaterialIds`、
|
||||
`userQuery`（自然语言补充检索）、`editedOutline`、`outlineFeedback`、`userRevisionNotes`、
|
||||
`overrideVerdict`、`selectedIssueIndices`、`sectionDecisions`。
|
||||
**协议已同时支持「自然语言 + 结构化 action」**——这是对话层能复用的关键。
|
||||
|
||||
> 结论：**状态机 / 流式 / 暂停·续跑协议 / 持久化都可复用**，无需重写引擎。改造本质是
|
||||
> 「把分散在干预卡里的输入框，换成一个底部智能输入框 + 意图路由层」，再把意图翻译回**现有
|
||||
> `action + payload`**；仅少量新意图（对内容提问、素材不足自动回退重搜、追加审核意见）需新增能力。
|
||||
|
||||
### 1.2 前端 `open-canvas` 子系统
|
||||
|
||||
- 主页面：`frontend-web/src/open-canvas/pages/AIWritingPage.tsx`
|
||||
- 路由注册：`open-canvas/routes/OpenCanvasRoutes.tsx`（`/page/canvas/ai-writing/*`）、
|
||||
`pages/WorkspaceRoutes.tsx`（`/page/workspace/ai-writing/*`）、`pages/NotebookAIWritingPage.tsx`
|
||||
- 状态机 + 流式：`contexts/AIWritingContext.tsx`（`useReducer` + LangGraph `useStream`，`assistantId:'ai_writing'`）
|
||||
- 事件翻译 / reducer：`hooks/useAIWritingStream.ts`
|
||||
- 类型：`types/ai-writing.ts`
|
||||
- **左栏面板**：`components/ai-writing/AIWritingPanel.tsx`（时间线 + 干预卡 + 状态栏；idle 时挂启动表单）
|
||||
- **要改造的输入 UI**（本次重点）：
|
||||
- 干预卡 `components/ai-writing/InterventionCard.tsx`：内含
|
||||
`MaterialConfirmCard`（素材：有 `MaterialsToolbar` 输入框）、
|
||||
`OutlineConfirmCard`（大纲：有 feedback textarea）、
|
||||
`DraftConfirmCard`（草稿：有「修改意见」textarea）、
|
||||
`ReviewConfirmCard`（审核：有「补充修改指令」textarea）、
|
||||
`SectionHelpCard`（素材不足：按钮组）。
|
||||
**每张卡里都既有输入框又有确认/动作按钮**——本次要**去掉输入框、保留按钮**。
|
||||
- 启动表单 `components/ai-writing/AIWritingForm.tsx`（idle 阶段的配置表单）——本次由底部输入框接管启动。
|
||||
- **已存在的对话原型**（可借鉴）：`components/ai-writing/WritingSetupChat.tsx`
|
||||
（`useThreadStream` 起 lead_agent 线程 + `setup_writing` 审批卡回填表单；仅管启动前配置，不直接复用）。
|
||||
- 底部输入框可直接复用 UI 组件：`@/components/ai-elements/prompt-input`
|
||||
（`PromptInput` / `PromptInputTextarea` / `PromptInputSubmit` / `PromptInputFooter`，见 `WritingSetupChat`）。
|
||||
|
||||
---
|
||||
|
||||
## 2. 目标与范围
|
||||
|
||||
### 2.1 必须满足
|
||||
|
||||
1. **复制式新增**：新目录 / 新路由 / 新侧边栏入口，**旧页面、旧路由、旧组件零改动**（「初始的不用动」）。
|
||||
2. 页面**底部单一多行智能输入框**接管全流程交互；四个阶段干预卡里的**输入框全部移除**。
|
||||
3. **保留确认按钮**：每个阶段到达暂停点时，消息流里仍渲染**产物卡 + 原来的确认/动作按钮**
|
||||
（素材「确认素材」、大纲「确认大纲」、草稿「让编辑审核 / 确认定稿」、审核「按编辑意见修改 / 强制定稿」、
|
||||
素材不足的三选一按钮）。用户**点按钮**或**在底部输入框发自然语言意图**二选一均可推进。
|
||||
4. **强意图理解**：把用户自由文本准确路由到正确动作，**覆盖现有全部交互 + 更多**（见 3.2）。
|
||||
5. **阶段作用域**：意图解析必须知道「当前停在哪个专家阶段」，把同一句话按当前阶段翻译成对应 action。
|
||||
6. **素材不足自动回退重搜**：修改意见需要的信息现有素材没有时，先重搜补素材、补完自动继续原意图。
|
||||
7. **拿不准就协助**：低置信意图**不猜着执行**，走澄清追问（clarify）或用户协助。
|
||||
8. 保留右侧草稿实时预览（可编辑）与时间线 / 进度展示。
|
||||
|
||||
### 2.2 明确不做（本期）
|
||||
|
||||
- 不改旧 `AIWritingPage` 及其 `AIWritingForm` / `InterventionCard`（这两个是要在**副本**里改的）。
|
||||
- 不替换 `ai_writing` graph 四个阶段大脑逻辑，不动五个暂停点 / checkpoint / 持久化。
|
||||
- 不引新状态管理库（继续 `useReducer` + `useStream`）。
|
||||
|
||||
### 2.3 复制策略（关键决策）
|
||||
|
||||
「复制一份」= **复制会被大改的表现层组件**（页面 + 干预卡 + 面板），**复用不改的引擎层**：
|
||||
|
||||
| 复制（新建副本，大改） | 复用（原样，不改；仅对 `types` 做**追加式**扩展） |
|
||||
|------------------------|--------------------------------------------------|
|
||||
| 新页面 `AIWritingChatPage.tsx` | `contexts/AIWritingContext.tsx`（状态机 + 流式） |
|
||||
| 新面板 `WritingChatPanel.tsx`（消息流 + 底部输入框） | `hooks/useAIWritingStream.ts`（事件 → state） |
|
||||
| 新消息卡 `chat-cards/*`（产物卡+确认按钮，**无输入框**） | `components/ai-writing/AIWritingDraftPanel.tsx`（右侧草稿预览） |
|
||||
| 新意图客户端 `api/ai-writing-intent.ts` | 各流式展示子组件（`MaterialsCard`/`OutlineCard`/`ReviewCard`/`SkillStepsCard`…） |
|
||||
| （可选）新 `contexts/WritingChatContext.tsx` 包一层「pending 意图 / 澄清态」 | `types/ai-writing.ts`（**只加字段不改旧字段**，保证旧页面不受影响） |
|
||||
|
||||
> **不复用、也不改** 的三个「输入框组件」：`AIWritingForm` / `InterventionCard` / `WritingFormContext`——
|
||||
> 旧页面继续用它们，新页面完全不引用它们。这样「初始的不用动」得到硬保证。
|
||||
>
|
||||
> 为什么复用 `AIWritingContext` 而不复制它？它**不含任何输入框 UI**，只是「状态机 + 与后端 graph 的
|
||||
> 流式桥」，复用它**不会改动旧页面的任何呈现**（旧页面照常在自己的 Provider 实例里跑）。复用能显著
|
||||
> 降低维护成本、避免两套引擎漂移。若要绝对物理隔离，也可整目录复制 `AIWritingContext` +
|
||||
> `useAIWritingStream` 到新目录——**推荐先复用**，仅当后续需要给对话态加大量图内新协议时再分叉。
|
||||
|
||||
---
|
||||
|
||||
## 3. 核心设计
|
||||
|
||||
### 3.1 三层结构
|
||||
|
||||
```
|
||||
┌──────────────────────────────────────────────────────────────┐
|
||||
│ 对话 UI 层（新） WritingChatPanel │
|
||||
│ - 消息流：用户消息 + AI 阶段产物卡（含【保留的确认按钮】）+ 进度 │
|
||||
│ - 底部：统一多行智能输入框（唯一自由输入入口） │
|
||||
│ - 右侧：草稿实时预览（复用 AIWritingDraftPanel） │
|
||||
└───────────────┬──────────────────────────────────────────────┘
|
||||
│ 用户自然语言（+ 当前 status/pausePoint 上下文）
|
||||
▼
|
||||
┌──────────────────────────────────────────────────────────────┐
|
||||
│ 意图解析层（新，核心） POST /api/ai-writing/sessions/{id}/intent │
|
||||
│ 输出：{ intent, pausePoint, action, payload, │
|
||||
│ needResearch, researchQuery, answer, confidence, clarify }│
|
||||
└───────────────┬──────────────────────────────────────────────┘
|
||||
│ 结构化 action + payload
|
||||
▼
|
||||
┌──────────────────────────────────────────────────────────────┐
|
||||
│ 执行层（复用现有 AIWritingContext） │
|
||||
│ idle → startWriting(AIWritingRequest) │
|
||||
│ 暂停 → submitIntervention(UserInterventionPayload) │
|
||||
│ 问答 → 旁路回答（不改写作状态） │
|
||||
│ 素材不足 → 先重搜、缓存 pending、补完自动续跑 │
|
||||
└──────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
**设计原则**：意图解析层是「翻译官」——把自由文本翻译成**执行层已经懂的指令**。能复用
|
||||
`action + payload` 的绝不新造协议；确认按钮点击**直接走执行层**（不经意图层，零延迟、零误判）。
|
||||
|
||||
### 3.2 意图分类体系（覆盖全流程 + 更多）
|
||||
|
||||
按「意图 → 触发时机（当前 status/pausePoint）→ 映射执行」组织，即意图解析引擎的目标 schema。
|
||||
**「甚至更多」** = G（问答）、H（控制）、以及 B/C/D/E 里的「素材不足自动回退重搜」。
|
||||
|
||||
#### A. 启动类（status = `idle`）
|
||||
| 意图 | 用户说法示例 | 映射 |
|
||||
|------|-------------|------|
|
||||
| `start_writing` | 「写一篇关于新能源出口的深度分析，2000 字」 | 解析成 `AIWritingRequest` → `startWriting()` |
|
||||
| `start_with_outline` | 「按这个大纲写：…」 | `AIWritingRequest.userOutline` |
|
||||
| `start_imitate` | 「仿照这篇样文写…」 | `writingModeType='imitate'` + `sampleText` |
|
||||
|
||||
#### B. 素材类（pausePoint = `material_confirm`）
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `confirm_materials` | 「素材可以，继续 / 下一步」 | `{action:'confirm', approvedMaterialIds}` |
|
||||
| `research_again` | 「重新搜索素材」 | `{action:'re_search'}`（无 userQuery=覆盖式重检） |
|
||||
| `add_material`/`add_keyword` | 「再补点关于 X 的素材」「加个关键词 Y」 | `{action:'re_search', userQuery:'X/Y'}`（追加） |
|
||||
| `select_materials` | 「只用第 1、3 条」 | `{action:'confirm', approvedMaterialIds:[…]}` |
|
||||
| `drop_materials` | 「去掉第 2 条」 | 同上（反选后的子集） |
|
||||
|
||||
#### C. 大纲类（pausePoint = `outline_confirm`）
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `confirm_outline` | 「大纲 OK，开始写」 | `{action:'confirm'}` |
|
||||
| `reoutline` | 「重写大纲」 | `{action:'re_outline'}`（feedback 可空） |
|
||||
| `revise_outline` | 「第二章拆成两节」「调整顺序」 | `{action:'re_outline', outlineFeedback:自然语言}` |
|
||||
|
||||
#### D. 写作 / 草稿类（pausePoint = `draft_confirm`）
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `to_editor` | 「交给编辑审核」 | `{action:'to_editor'}` |
|
||||
| `finalize` | 「就这样定稿」 | `{action:'finalize'}` **（高代价：低置信必澄清）** |
|
||||
| `revise_draft` | 「第三段太啰嗦，精简」「补个 2024 案例」 | `{action:'user_revise', userRevisionNotes:意见}` |
|
||||
| `rewrite_section` | 「重写第二章」 | `user_revise` + 章节定位写进 notes |
|
||||
| `change_tone/style` | 「更口语一点 / 更正式」 | `user_revise` + notes |
|
||||
| `apply_review`（若已有审核结果） | 「按编辑意见改」「只接受第 1、2 条」 | `{action:'user_revise', selectedIssueIndices:[…]}` |
|
||||
|
||||
#### E. 编辑审核类（pausePoint = `review_confirm`）
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `accept_review` | 「按编辑意见改」 | `{action:'accept_review', selectedIssueIndices}` |
|
||||
| `partial_accept_review` | 「只接受第 1、2 条」 | `accept_review` + `selectedIssueIndices` 子集 |
|
||||
| `add_review_opinion` | 「再补一条：核查数据准确性」 | `accept_review` + `userRevisionNotes`（并入补充指令，现协议已支持） |
|
||||
| `force_finalize` | 「不改了，强制定稿」 | `{action:'force_finalize', overrideVerdict:true}` **（高代价：必澄清）** |
|
||||
|
||||
#### F. 素材不足求助类（pausePoint = `section_help`）
|
||||
| 意图 | 映射 |
|
||||
|------|------|
|
||||
| `section_supplement` | `{action:'section_help', sectionDecisions:[…'supplement']}` |
|
||||
| `section_loose` | `sectionDecisions:[…'loose']` |
|
||||
| `section_delete` | `sectionDecisions:[…'delete']` |
|
||||
|
||||
#### G. 问答类（**新**；任意 status；旁路，不改写作状态）
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `ask_about_content` | 「第二章讲了啥？」「为什么用这个论点？」 | 意图节点直接带回 `answer`，对话流渲染（见 5.4） |
|
||||
| `ask_meta` | 「现在到哪步了？」「用了哪些素材？」 | 读当前 state 直接回答，不走图 |
|
||||
|
||||
#### H. 控制类（**新 / 半新**）
|
||||
| 意图 | 映射 |
|
||||
|------|------|
|
||||
| `next_step`（通用「继续/下一步」） | 按**当前 pausePoint** 映射到该阶段的默认确认（material/outline→`confirm`，draft→`to_editor`…；等价点确认按钮） |
|
||||
| `restart_stage` | 重跑当前阶段（复用 `resumeWriting()` 空 command 续跑） |
|
||||
| `run_background` | 转后台自动跑：`suspendAIWritingToBackground()` |
|
||||
| `cancel` | `pauseWriting()`（`stream.stop()`） |
|
||||
| `unknown/clarify` | 无法判定 → 反问澄清 / 用户协助（**防误触发硬指标**） |
|
||||
|
||||
### 3.3 「素材不足自动重搜」判定与回退（用户强调点）
|
||||
|
||||
触发：用户在**大纲/草稿/审核**阶段提出修改意见，但意见需要的信息现有素材里没有。
|
||||
|
||||
判定来源（组合）：
|
||||
1. **意图节点直接判**：把「素材摘要（keywords + 每条 title/source）」喂给意图 prompt，让它判断
|
||||
「这条意见能否用现有素材满足」；不能 → `needResearch=true` + `researchQuery`。
|
||||
2. **执行阶段兜底**：`writer_draft` 严格模式已有 `blocked_sections` → `section_help` 暂停点，
|
||||
对话层把它翻译成「素材不足，是否重搜？」的追问。
|
||||
|
||||
回退动作（**首期纯前端实现，不动后端协议**）：`needResearch=true` 时，前端把用户的修改
|
||||
`payload` 缓存为 `pendingIntent`，先 `submitIntervention({action:'re_search', userQuery:researchQuery})`
|
||||
补素材；监听到新一轮 `material_confirm`（素材追加完成）后，**自动提交** `pendingIntent`。
|
||||
需处理「重搜期间用户又改主意」→ 允许覆盖 pending。
|
||||
|
||||
---
|
||||
|
||||
## 4. 意图解析引擎：方案与契约
|
||||
|
||||
**采用方案 A：后端 LLM 意图节点（独立 REST）**，叠加高频短句**规则快路径**优化。
|
||||
|
||||
理由：`needResearch` 判定需要**完整写作上下文**（素材/大纲/草稿），后端天然持有 checkpoint state；
|
||||
与 `setup_writing` 一脉相承，离线模型选择策略统一；输出 schema 稳定、前端只消费。
|
||||
|
||||
### 4.1 后端入口
|
||||
|
||||
`app/gateway/routers/ai_writing.py` 新增：
|
||||
|
||||
```
|
||||
POST /api/ai-writing/sessions/{session_id}/intent
|
||||
Body: { text: string, status: string, pausePoint?: string }
|
||||
Resp: IntentResult（见 4.3）
|
||||
```
|
||||
|
||||
实现：读该 session 的 checkpoint state（素材摘要 / 大纲纲要 / 草稿字数·章节 / 审核意见）→
|
||||
用 `ai_writing` 配置模型（复用 `intent_parser.py` 同套 `create_chat_model` + `parse_json` +
|
||||
模型容错 `_model_fallback`）解析 → 返回结构化契约。**不改图拓扑**。
|
||||
|
||||
新增 prompt：`agents/ai_writing/prompts/intent_router_prompts.py`
|
||||
- 输入：用户文本、`current_status`/`pause_point`、素材摘要、大纲纲要、草稿字数+章节标题、审核意见列表。
|
||||
- 输出：严格 JSON（4.3 契约）。硬要求：**只在证据充分时给高 confidence，否则给 `clarify`**；
|
||||
`finalize`/`force_finalize`/`re_search` 等高代价意图**必须**高置信才不澄清。
|
||||
|
||||
### 4.2 前端快路径（先跑规则，未命中再调后端）
|
||||
|
||||
`api/ai-writing-intent.ts` 里对**当前 pausePoint** 做少量确定性短语命中，直接产出 payload，省一次往返：
|
||||
- 「继续/下一步/确认/可以/ok/好的」→ 当前阶段默认 `confirm`（draft 阶段=`to_editor`）。
|
||||
- 「重新搜索/重搜」（material）→ `re_search`。「重写大纲」（outline）→ `re_outline`。
|
||||
- 「定稿/强制定稿」→ **不进快路径**（高代价，一律走后端 + 可能澄清）。
|
||||
- 命中即执行；未命中 → 调后端 `intent` 接口。
|
||||
|
||||
### 4.3 意图解析输出契约
|
||||
|
||||
```jsonc
|
||||
{
|
||||
"intent": "revise_draft", // 见 3.2 分类
|
||||
"pausePoint": "draft_confirm", // 需要 resume 时；否则 null
|
||||
"action": "user_revise", // 映射到现有 action
|
||||
"payload": { // 直接可喂 submitIntervention 的字段（camelCase）
|
||||
"userRevisionNotes": "第三段精简，并补一个 2024 年的案例"
|
||||
},
|
||||
"needResearch": true, // 素材是否不足
|
||||
"researchQuery": "2024 行业案例", // needResearch 时的检索意图
|
||||
"answer": null, // 问答类意图的直接回答（旁路）
|
||||
"confidence": 0.86,
|
||||
"clarify": null // 低置信时的反问话术（非空则不执行、只追问）
|
||||
}
|
||||
```
|
||||
|
||||
前端消费顺序：
|
||||
1. `answer` 非空 → 对话流渲染回答（问答旁路，不动写作状态）。
|
||||
2. `clarify` 非空 **或** `confidence < 阈值(建议 0.6)` → 渲染反问气泡，等用户澄清（**防误触发**）。
|
||||
3. `needResearch=true` → 缓存 `pendingIntent=payload`，先重搜，收到新素材后自动提交。
|
||||
4. 否则 → `idle` 走 `startWriting(...)`；暂停态走 `submitIntervention({action, ...payload})`。
|
||||
|
||||
---
|
||||
|
||||
## 5. 前端改造方案
|
||||
|
||||
### 5.1 新页面与路由（复制式）
|
||||
|
||||
- 新页面：`open-canvas/pages/AIWritingChatPage.tsx`
|
||||
（骨架照抄 `AIWritingPage.tsx` 的 Provider 包裹与左右分栏；左栏换成 `WritingChatPanel`，右栏仍
|
||||
`AIWritingDraftPanel`；idle 不再挂 `AIWritingForm`，改由底部输入框启动）。
|
||||
- 路由注册（**新增，不动旧行**）：
|
||||
- `open-canvas/routes/OpenCanvasRoutes.tsx` 加 `path="ai-writing-chat/*"`。
|
||||
- 侧边栏 `core/page-layout/sidebar-menu.ts`：在 `qa-research` 旁新增
|
||||
`{ id:"qa-research-chat", label:"AI写作(对话版)", icon:FileText, path:"/page/canvas/ai-writing-chat" }`。
|
||||
（或先不加、内部灰度直链验证。）
|
||||
|
||||
### 5.2 新面板 `WritingChatPanel.tsx`（左栏主体）
|
||||
|
||||
职责 = 消息流容器 + 底部输入框，替代 `AIWritingPanel` 里的「时间线 + InterventionCard + 状态栏」。
|
||||
复用 `useAIWriting()` 读 state / 调 `startWriting` / `submitIntervention` / `pauseWriting` /
|
||||
`suspendAIWritingToBackground`。
|
||||
|
||||
消息流渲染（把 `ProgressEvent` / 阶段产物适配成消息卡，复用现有展示子组件）：
|
||||
- 运行中：`research_progress`/`outline_chunk`/`draft_chunk`/`review_chunk` → 进度 / 流式卡
|
||||
（复用 `SkillStepsCard`/`OutlineStreamingView`/`DraftGeneratingView`/`ReviewStreamingView` 的展示部分）。
|
||||
- 到达暂停点：渲染 **产物卡 + 保留的确认按钮**（见 5.3），不再有输入框。
|
||||
- 用户消息、AI 澄清追问、问答回答：普通气泡。
|
||||
|
||||
底部输入框：直接复用 `@/components/ai-elements/prompt-input`（`PromptInput`+`PromptInputTextarea`+
|
||||
`PromptInputSubmit`），Enter 发送、Shift+Enter 换行；发送时读当前 `state.status`/`currentPause?.pausePoint`
|
||||
交给意图路由（见 5.5）。运行中（非暂停点）可禁用发送或转「排队/提示当前在生成」。
|
||||
|
||||
### 5.3 新消息卡 `chat-cards/*`（**保留按钮、去掉输入框**）
|
||||
|
||||
把 `InterventionCard.tsx` 的五张卡各复制一份到 `open-canvas/components/ai-writing-chat/chat-cards/`，
|
||||
**删掉其中的 textarea / MaterialsToolbar 输入部分，保留产物展示 + 确认/动作按钮**：
|
||||
|
||||
| 新卡 | 展示（复用） | 保留的按钮 | 删除的输入 |
|
||||
|------|-------------|-----------|-----------|
|
||||
| `MaterialConfirmChatCard` | `MaterialsCard`（勾选保留）+ `SkillStepsCard` | 「确认素材，规划大纲」「重新搜索」 | `MaterialsToolbar` 的 userQuery 输入框 |
|
||||
| `OutlineConfirmChatCard` | `OutlineCard`（可点标题微调保留） | 「确认大纲，开始写作」 | feedback textarea + 发送 |
|
||||
| `DraftConfirmChatCard` | `ReviewCard`（若有审核，勾选保留） | 「按编辑审核意见修改」「让编辑继续审核」「确认定稿」 | 「修改意见」textarea |
|
||||
| `ReviewConfirmChatCard` | `ReviewCard`（勾选保留） | 「按编辑意见修改」「强制定稿」 | 「补充修改指令」textarea |
|
||||
| `SectionHelpChatCard` | 章节 + 三选一按钮组 | 三选一 + 「提交并继续」 | （本就无自由输入，原样保留） |
|
||||
|
||||
> 勾选态（素材/审核意见）**保留**——它是「点按钮」路径的一部分，不算自由输入框。用户想用自然语言
|
||||
> 「只用第 1、3 条」也能走底部输入框（意图层产出 `approvedMaterialIds`/`selectedIssueIndices`）。
|
||||
> 按钮点击 **直接调 `submitIntervention`**，与旧卡完全一致，零改动风险。
|
||||
|
||||
### 5.4 问答旁路（新能力）
|
||||
|
||||
「对写作内容提问」不改写作 state。首期**轻量实现**：意图节点检出 `ask_*` 时读 state 直接带回 `answer`，
|
||||
前端在消息流渲染为普通 AI 气泡即可（一次调用出答案）。完整版（后续）可另起只读 lead_agent 线程注入
|
||||
草稿/大纲作背景，类似 `WritingSetupChat` 的 `useThreadStream`。
|
||||
|
||||
### 5.5 交互流程（对话版）串起来
|
||||
|
||||
1. **空闲**：底部输入「写一篇关于 X 的深度分析，2000 字」→ 意图 `start_writing` → 组 `AIWritingRequest`
|
||||
→ `startWriting()`。（缺关键字段时先 `clarify` 追问字数/读者/类型。）
|
||||
2. **运行中**：进度/流式卡渲染。
|
||||
3. **到暂停点**：渲染产物卡 + 按钮 + 一句提示「你可以点上面的按钮，或直接说：确认继续 / 补充素材 / 改大纲…」。
|
||||
- **点按钮** → 直接 `submitIntervention`（旧逻辑）。
|
||||
- **发文字** → `resolveIntent(text, status, pausePoint)`（先快路径、后后端）→ 按 4.3 消费。
|
||||
4. **问答**：任意时刻提问 → 旁路 `answer`，不打断写作。
|
||||
5. **素材不足**：`needResearch` 或 `section_help` → 追问/执行「重搜 → 自动续跑 pending 意图」。
|
||||
6. **拿不准**：`clarify`/低置信 → 反问气泡，不执行。
|
||||
|
||||
### 5.6 pending 意图 & 澄清态（新增本地状态）
|
||||
|
||||
在 `WritingChatPanel`（或新 `WritingChatContext`）维护：
|
||||
- `pendingIntent: UserInterventionPayload | null` —— 素材不足重搜完成后自动提交。
|
||||
- `awaitingClarify: { originalText: string } | null` —— 澄清态，用户下一句与上一句拼接再解析。
|
||||
- pending 生命周期：设置 → 提交 `re_search` → 监听 reducer 里新一轮 `awaiting_material_confirm`
|
||||
（`state.status` 从 `researching` 回到 `awaiting_material_confirm`）→ 自动 `submitIntervention(pending)` → 清空。
|
||||
|
||||
---
|
||||
|
||||
## 6. 后端改造方案
|
||||
|
||||
### 6.1 意图解析入口（新增，唯一后端改动主体）
|
||||
|
||||
- `app/gateway/routers/ai_writing.py`：新增 `POST /sessions/{id}/intent`（见 4.1）。
|
||||
读 checkpoint state 用 `app.state.checkpointer` + `make_ai_writing_graph().aget_state(...)`
|
||||
（参照 `ai_writing_job_executor.py` 的读法）取素材/大纲/草稿/审核摘要。
|
||||
- `agents/ai_writing/prompts/intent_router_prompts.py`：新增意图 prompt（见 4.1）。
|
||||
- 复用 `nodes/intent_parser.py` 的模型创建 / `parse_json` / `_model_fallback` 容错。
|
||||
|
||||
### 6.2 不需要改的部分
|
||||
|
||||
- 四个阶段大脑节点、五个暂停点、checkpoint、StreamBridge、持久化、sessions/transcript REST——**全不动**。
|
||||
- 「素材不足自动续跑」「问答旁路」首期都在**前端 / 意图节点**实现，**不扩 resume 协议**。
|
||||
|
||||
### 6.3 可选增强（阶段 3，按需）
|
||||
|
||||
- `add_review_opinion` 完整版：`review_confirm` resume 增字段 `extra_review_notes`（首期已用现有
|
||||
`userRevisionNotes` 降级承载，够用）。
|
||||
- `resume_intent` 后端化（把「重搜完自动续跑」搬到后端，减少前端状态机复杂度）。
|
||||
- undo / 回退到上一暂停点（依赖 checkpoint 回滚）。
|
||||
|
||||
---
|
||||
|
||||
## 7. 数据流时序（改稿 + 素材不足自动重搜）
|
||||
|
||||
```
|
||||
用户（draft_confirm 暂停）: "第三段补个2024年的行业案例"
|
||||
│ 底部输入框 → resolveIntent(text, status='awaiting_draft_confirm', pausePoint='draft_confirm')
|
||||
▼ POST /api/ai-writing/sessions/{id}/intent { text, status, pausePoint }
|
||||
后端: 读 state(素材摘要) + LLM 解析
|
||||
│ 返回 { intent:revise_draft, action:user_revise,
|
||||
│ payload:{userRevisionNotes:"第三段…补2024案例"},
|
||||
│ needResearch:true, researchQuery:"2024 行业案例", confidence:0.9 }
|
||||
▼
|
||||
前端: needResearch=true → pendingIntent = payload
|
||||
│ submitIntervention({action:'re_search', userQuery:"2024 行业案例"})
|
||||
▼ LangGraph resume → researcher 追加素材 → status 回到 awaiting_material_confirm
|
||||
前端: 监听到新一轮素材确认 → 自动提交 pendingIntent
|
||||
│ submitIntervention({action:'user_revise', userRevisionNotes:"第三段…补2024案例"})
|
||||
▼ writer_draft 用新素材改稿 → draft_chunk… → draft_ready → draft_confirm
|
||||
前端: 消息流展示新草稿卡 + "已按你的意见改写并补充了 2024 案例"
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 8. 关键实现细节与坑
|
||||
|
||||
1. **阶段作用域**：`resolveIntent` 必须带 `state.currentPause?.pausePoint` + `state.status`——
|
||||
同一句「重新来」在 material 是重搜、在 outline 是重写大纲。`AIWritingContext` 已有这两个字段，直接读。
|
||||
2. **确认按钮零改动**：新卡的按钮 `onClick` 原样调 `submitIntervention(payload)`，与旧卡逐字一致，
|
||||
风险最低；**只删输入框**。
|
||||
3. **防误触发**：低置信（<0.6）或 `clarify` 非空一律追问；`finalize`/`force_finalize`/`re_search`
|
||||
即使命中也要求高置信，否则澄清。绝不盲目定稿/覆盖重搜。
|
||||
4. **pending 意图队列**：素材不足是「先插重搜、再回原意图」，用 5.6 的小状态机；重搜完成前允许覆盖 pending。
|
||||
5. **问答不污染 transcript**：问答/澄清消息在前端标记 `kind:'qa'|'clarify'`，与写作进度事件区分（
|
||||
`AIWritingContext` 的 transcript 只存 progressEvents/completedInterventions，问答气泡是纯前端消息，
|
||||
默认不入 transcript，避免历史回看错乱）。
|
||||
6. **流式产物→消息卡**：复用 `OutlineStreamingView`/`DraftGeneratingView`/`ReviewStreamingView`/
|
||||
`SkillStepsCard` 的**展示部分**，套进消息容器，不重写渲染。
|
||||
7. **右侧草稿预览**：`AIWritingDraftPanel` + `DraftSync` 原样复用（`DraftSync` 依据 status/markdown
|
||||
同步右侧画布）。
|
||||
8. **离线内网**：意图模型走 `config.yaml → ai_writing` 配置模型；检索仍走技能/内网 ES，不联网。
|
||||
9. **模型选择透传**：`submitIntervention` 的 resume 已把 `state.request?.modelName` 一并回传（见
|
||||
`AIWritingContext`），新页面沿用即可。
|
||||
10. **旧页面隔离验证**：改完后回归旧 `/page/canvas/ai-writing` 与 `/page/workspace/ai-writing`、
|
||||
笔记本嵌入页，确认零变化（新代码不 import 旧的 `AIWritingForm`/`InterventionCard`/`WritingFormContext`）。
|
||||
|
||||
---
|
||||
|
||||
## 9. 分阶段实施路线图
|
||||
|
||||
**阶段 0 — 骨架（不含意图）**
|
||||
- 复制 `AIWritingChatPage` + 注册新路由 + 侧边栏入口；复用 `AIWritingContext`/`useStream`/`DraftPanel`。
|
||||
- 复制五张 `chat-cards`（删输入框、留按钮）；底部输入框先只做 idle 直发 `start_writing`
|
||||
（文本原样当 `userIntent`，其余字段取默认）+ 暂停点靠按钮推进。端到端跑通。
|
||||
|
||||
**阶段 1 — 意图解析（核心）**
|
||||
- 后端 `POST /sessions/{id}/intent` + `intent_router_prompts.py`，返回 4.3 契约。
|
||||
- 前端 `api/ai-writing-intent.ts`（快路径 + 后端）接入 `WritingChatPanel`，覆盖 A~F 类意图。
|
||||
- 低置信 `clarify` 澄清态（5.6）。
|
||||
|
||||
**阶段 2 — 问答旁路 + 素材不足自动重搜**
|
||||
- G 类问答（意图节点直接带 `answer`）。
|
||||
- `needResearch` + pending 意图机制（前端实现重搜后自动续跑，5.6）。
|
||||
|
||||
**阶段 3 — 增强（按需）**
|
||||
- `add_review_opinion` 扩协议、`resume_intent` 后端化、undo/回退、完整问答只读线程。
|
||||
|
||||
**阶段 4 — 打磨**
|
||||
- 消息卡视觉、历史回看（transcript）、后台挂起入口、灰度到侧边栏。
|
||||
|
||||
---
|
||||
|
||||
## 10. 关键文件清单（落地对照）
|
||||
|
||||
### 新增（前端）
|
||||
- `open-canvas/pages/AIWritingChatPage.tsx` — 新页面（复制 `AIWritingPage` 骨架改造）
|
||||
- `open-canvas/components/ai-writing-chat/WritingChatPanel.tsx` — 消息流 + 底部输入框
|
||||
- `open-canvas/components/ai-writing-chat/chat-cards/{MaterialConfirmChatCard,OutlineConfirmChatCard,DraftConfirmChatCard,ReviewConfirmChatCard,SectionHelpChatCard}.tsx` — 保留按钮、去掉输入框
|
||||
- `open-canvas/components/ai-writing-chat/progressToMessage.ts` — `ProgressEvent` → 消息卡适配
|
||||
- `open-canvas/api/ai-writing-intent.ts` — 意图解析客户端（快路径 + `POST …/intent`）
|
||||
- （可选）`open-canvas/contexts/WritingChatContext.tsx` — pending 意图 / 澄清态
|
||||
|
||||
### 修改（前端，仅追加，不动旧逻辑）
|
||||
- `open-canvas/routes/OpenCanvasRoutes.tsx` — 加一条 `ai-writing-chat/*` 路由
|
||||
- `core/page-layout/sidebar-menu.ts` — 加一个侧边栏入口（可选）
|
||||
- `open-canvas/types/ai-writing.ts` — **仅追加** `IntentResult` 等新类型，不改旧字段
|
||||
|
||||
### 新增（后端）
|
||||
- `app/gateway/routers/ai_writing.py` — 加 `POST /sessions/{id}/intent`
|
||||
- `agents/ai_writing/prompts/intent_router_prompts.py` — 意图 prompt
|
||||
- `tests/test_ai_writing_intent.py` — 意图接口单测（TDD 强制）
|
||||
|
||||
### 复用（不改）
|
||||
- `open-canvas/contexts/AIWritingContext.tsx`、`hooks/useAIWritingStream.ts`
|
||||
- `open-canvas/components/ai-writing/{AIWritingDraftPanel,DraftSync,MaterialsCard,OutlineCard,ReviewCard,SkillStepsCard,OutlineStreamingView,DraftGeneratingView,ReviewStreamingView}.tsx`
|
||||
- `agents/ai_writing/graph.py`、四个阶段节点、五个暂停点、`nodes/intent_parser.py`
|
||||
|
||||
### 绝不动（保证「初始的不用动」）
|
||||
- `open-canvas/pages/AIWritingPage.tsx` 及其引用的
|
||||
`components/ai-writing/{AIWritingForm,InterventionCard}.tsx`、`contexts/WritingFormContext.tsx`
|
||||
- 旧路由 `/page/canvas/ai-writing`、`/page/workspace/ai-writing`、笔记本嵌入页
|
||||
|
||||
---
|
||||
|
||||
## 11. 待确认（默认取值已给出，可直接推进）
|
||||
|
||||
1. 意图引擎：**后端 LLM 节点（方案 A）+ 前端短句快路径** —— 默认采用。
|
||||
2. `AIWritingContext` / `useAIWritingStream`：**复用**（不物理复制）——默认采用；需绝对隔离可改为复制。
|
||||
3. 新页面入口：**侧边栏新增「AI写作(对话版)」** —— 默认加；若要先灰度可只留直链。
|
||||
4. 「素材不足自动重搜」：**首期前端 pending 续跑**（不动后端 resume 协议）——默认采用。
|
||||
5. 问答旁路：**首期意图节点直接带 `answer`** —— 默认采用。
|
||||
|
||||
> 以上默认按「A + 复用引擎 + 侧边栏入口 + 前端 pending + 意图节点带答案」推进，如需调整请指出。
|
||||
@@ -1,454 +0,0 @@
|
||||
# AI 写作「对话驱动 + 意图解析」改造方案
|
||||
|
||||
> 目标：在**不动现有 AI 写作页面**的前提下，新建一个「对话驱动」的写作工作台。
|
||||
> 去掉素材收集 / 大纲规划 / 作家写作 / 编辑审核各阶段**分散的输入框/表单**，
|
||||
> 用**底部一个统一的智能输入框**接管全流程交互；该输入框具备**强意图解析**能力，
|
||||
> 能准确识别用户意图（重新搜索、加素材、加关键词、重写大纲、按要求改大纲、改写作、
|
||||
> 改编辑、提编辑意见、对写作内容提问……），并在**素材不足**时自动回退去重新检索素材。
|
||||
|
||||
本文只做**方案与实现指引**，不含代码改动。落地时按第 9 节的路线图分阶段推进。
|
||||
|
||||
---
|
||||
|
||||
## 1. 现状梳理（改造的地基）
|
||||
|
||||
现有 AI 写作是一套**独立的 LangGraph 子系统**（不是 lead_agent），前后端都已成型。
|
||||
|
||||
### 1.1 后端：`ai_writing` graph
|
||||
|
||||
- 图工厂：`offline-backend-20260512/backend/packages/harness/deerflow/agents/ai_writing/graph.py`
|
||||
- 状态：`.../agents/ai_writing/state.py`（`AIWritingState` + `UserInterventionPayload`）
|
||||
- 注册：`offline-backend-20260512/backend/langgraph.json` → `assistant_id = "ai_writing"`
|
||||
- 执行：与 lead agent **共用** RunManager + StreamBridge + checkpointer（`app/gateway/services.py` 里 `assistant_id=="ai_writing"` 分发到 `make_ai_writing_graph`）
|
||||
|
||||
节点拓扑（实际）：
|
||||
|
||||
```
|
||||
sample_analyzer → intent_parser → researcher
|
||||
→ [cond] pause_material | writer_draft
|
||||
→ writer_outline → pause_outline
|
||||
→ writer_draft → [cond] pause_section_help | pause_draft | END
|
||||
→ editor → [cond] pause_draft | pause_review
|
||||
→ pause_review → END | writer_draft
|
||||
```
|
||||
|
||||
四个「阶段大脑」节点：
|
||||
|
||||
| 阶段 | 节点 | 文件 |
|
||||
|------|------|------|
|
||||
| 素材收集 | `researcher` | `nodes/researcher.py` |
|
||||
| 大纲规划 | `writer_outline` | `nodes/writer_outline.py` |
|
||||
| 作家写作 | `writer_draft` | `nodes/writer_draft.py` |
|
||||
| 编辑审核 | `editor` | `nodes/editor.py` |
|
||||
|
||||
五个**暂停点**（`interrupt()`，靠 `Command(resume=payload)` 续跑）：
|
||||
|
||||
| pause_point | 语义 | 可用 action（现有） |
|
||||
|-------------|------|---------------------|
|
||||
| `material_confirm` | 素材确认 | `confirm` / `re_search` / `supplement_search` / `enable_skill_search` |
|
||||
| `outline_confirm` | 大纲确认 | `confirm` / `re_outline` |
|
||||
| `draft_confirm` | 草稿确认 | `to_editor` / `finalize` / `user_revise` / `custom_revise` |
|
||||
| `review_confirm` | 编辑打回确认 | `accept_review` / `force_finalize` |
|
||||
| `section_help` | 素材不足求助 | `supplement` / `loose` / `delete` |
|
||||
|
||||
> resume 载荷字段（`UserInterventionPayload`）：`action`、`approved_material_ids`、
|
||||
> `extra_keywords`、`user_query`(自然语言补充检索)、`edited_outline`、`outline_feedback`、
|
||||
> `user_revision_notes`、`override_verdict`、`selected_issue_indices`、`section_decisions`。
|
||||
> **关键点**：现有协议已经支持「自然语言 + 结构化 action」，这是对话改造的核心可复用资产。
|
||||
|
||||
REST（业务元数据，非图执行）：`app/gateway/routers/ai_writing.py`，前缀 `/api/ai-writing`
|
||||
（sessions CRUD、transcript 保存、article-types、后台挂起 `/background` + `/progress` 等）。
|
||||
|
||||
持久化：`ai_writing_sessions` 表（`transcript` 为 `PortableLongText`）
|
||||
+ `ai_writing_article_types`（`deerflow/persistence/...`）。
|
||||
|
||||
配置：`config.yaml → ai_writing`（`keyword_model`、`rank_model`、`max_materials`、
|
||||
`use_builtin_agents.{researcher,outliner,writer,editor}` 等）。
|
||||
|
||||
### 1.2 前端：`open-canvas` 子系统
|
||||
|
||||
- 主页面：`frontend-web/src/open-canvas/pages/AIWritingPage.tsx`
|
||||
- 路由：`/page/canvas/ai-writing`、`/page/workspace/ai-writing`
|
||||
- 状态机：`contexts/AIWritingContext.tsx`（`useReducer` + LangGraph `useStream`）
|
||||
- 事件翻译/reducer：`hooks/useAIWritingStream.ts`
|
||||
- 类型：`types/ai-writing.ts`（`WritingStatus`、`ProgressEvent`、`UserInterventionPayload`、`AIWritingRequest`…）
|
||||
- 流式：LangGraph `useStream`，`assistantId: 'ai_writing'`，`streamMode: ['values','updates','custom','messages-tuple']`
|
||||
- **要去掉的输入 UI**（改造重点）：
|
||||
- 启动表单 `components/ai-writing/AIWritingForm.tsx` + `WritingFormContext`
|
||||
- 阶段暂停输入 `components/ai-writing/InterventionCard.tsx`（素材/大纲/草稿/审核四套内嵌表单）
|
||||
- **已存在的对话原型**：`components/ai-writing/WritingSetupChat.tsx`
|
||||
（用 `useThreadStream` 起一条 lead_agent 线程 + `setup_writing` 工具审批卡回填表单）——
|
||||
这是「对话 → 结构化配置」的现成范式，可借鉴但**不直接复用**（它只管启动前配置）。
|
||||
|
||||
### 1.3 结论
|
||||
|
||||
- **状态机、流式、暂停/续跑协议、持久化都可复用**，无需重写引擎。
|
||||
- 改造本质 = **把「分散表单」换成「一个对话输入框 + 意图路由层」**，
|
||||
再把意图映射回**现有的 `action + payload`**（大部分场景），
|
||||
少量新意图（如「对内容提问」「素材不足自动重搜」）需要**新增协议/节点**。
|
||||
|
||||
---
|
||||
|
||||
## 2. 目标与范围
|
||||
|
||||
### 2.1 必须满足
|
||||
|
||||
1. 新页面/新路由，**旧页面零改动**（用户要求「初始的不用动」）。
|
||||
2. 底部**单一对话输入框**控制所有阶段交互，移除各阶段独立输入框。
|
||||
3. **强意图解析**：准确识别用户自然语言意图并路由到正确动作，覆盖全流程 + 更多。
|
||||
4. 用户给修改意见后，**若现有素材不足以满足要求 → 自动/建议重新检索素材**再继续。
|
||||
5. 保留右侧草稿实时预览（可编辑）与时间线/进度展示。
|
||||
|
||||
### 2.2 明确不做（本期）
|
||||
|
||||
- 不改旧 `AIWritingPage` 及其表单/干预卡组件。
|
||||
- 不替换底层 `ai_writing` graph 的四个阶段大脑逻辑（只加意图层与少量新协议）。
|
||||
- 不引入新的状态管理库（继续 `useReducer` + `useStream`）。
|
||||
|
||||
---
|
||||
|
||||
## 3. 核心设计
|
||||
|
||||
### 3.1 总体思路：三层结构
|
||||
|
||||
```
|
||||
┌─────────────────────────────────────────────────────────┐
|
||||
│ 对话 UI 层（新） │
|
||||
│ - 左：消息流（用户消息 + AI 阶段产物卡片 + 进度） │
|
||||
│ - 底：统一智能输入框（唯一交互入口） │
|
||||
│ - 右：草稿实时预览（复用 AIWritingDraftPanel） │
|
||||
└───────────────┬─────────────────────────────────────────┘
|
||||
│ 用户自然语言
|
||||
▼
|
||||
┌─────────────────────────────────────────────────────────┐
|
||||
│ 意图解析层（新，核心） │
|
||||
│ - 输入：用户文本 + 当前 WritingStatus + 上下文快照 │
|
||||
│ - 输出：{ intent, action, payload, needResearch, ... } │
|
||||
└───────────────┬─────────────────────────────────────────┘
|
||||
│ 结构化 action + payload
|
||||
▼
|
||||
┌─────────────────────────────────────────────────────────┐
|
||||
│ 执行层（复用现有） │
|
||||
│ - 空闲态：startWriting(AIWritingRequest) │
|
||||
│ - 暂停态：submitIntervention(UserInterventionPayload) │
|
||||
│ - 运行中：stream.stop() → 续跑 / 新命令 │
|
||||
│ - 提问态：旁路问答（不改写作状态） │
|
||||
└─────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
**设计原则**：意图解析层是「翻译官」——把自由文本翻译成**现有执行层已经懂的指令**。
|
||||
能复用 `action + payload` 的绝不新造协议；只有现有协议表达不了的才扩展。
|
||||
|
||||
### 3.2 意图分类体系（覆盖用户全流程 + 更多）
|
||||
|
||||
按「意图类别 → 触发时机（当前 status）→ 映射到的执行」组织。这是**意图解析引擎的目标 schema**。
|
||||
|
||||
#### A. 启动类（status = idle）
|
||||
|
||||
| 意图 | 说明 | 映射 |
|
||||
|------|------|------|
|
||||
| `start_writing` | 用户描述要写什么 | 解析成 `AIWritingRequest` → `startWriting()` |
|
||||
| `start_with_outline` | 用户自带大纲 | `AIWritingRequest.userOutline` |
|
||||
| `start_imitate` | 样文仿写 | `writingModeType=imitate` + 样文 |
|
||||
|
||||
#### B. 素材类（status ≈ awaiting_material_confirm / 运行中）
|
||||
|
||||
| 意图 | 用户说法示例 | 映射 |
|
||||
|------|-------------|------|
|
||||
| `confirm_materials` | 「素材可以，继续」 | `material_confirm` + `confirm`（可带勾选 ids） |
|
||||
| `research_again` | 「重新搜索素材」 | `material_confirm` + `re_search` |
|
||||
| `add_material` / `add_keyword` | 「补充关于 X 的素材」「加个关键词 Y」 | `material_confirm` + `supplement_search`，`user_query`=X/Y |
|
||||
| `use_skill_search` | 「用 XX 技能查」 | `material_confirm` + `enable_skill_search`，`user_query`=技能/来源 |
|
||||
| `select_materials` | 「只用第 1、3 条」 | `confirm` + `approved_material_ids` 子集 |
|
||||
| `drop_materials` | 「去掉第 2 条」 | 同上，反选后 `approved_material_ids` |
|
||||
|
||||
#### C. 大纲类（status ≈ awaiting_outline_confirm）
|
||||
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `confirm_outline` | 「大纲 OK」 | `outline_confirm` + `confirm` |
|
||||
| `reoutline` | 「重写大纲」 | `outline_confirm` + `re_outline`（`outline_feedback` 可空） |
|
||||
| `revise_outline` | 「第二章拆成两节」「调整顺序」 | `re_outline` + `outline_feedback`=自然语言 |
|
||||
| `edit_outline_manually` | 用户直接给出新大纲文本 | 解析成 `edited_outline` |
|
||||
|
||||
#### D. 写作/草稿类（status ≈ awaiting_draft_confirm / writing_draft）
|
||||
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `to_editor` | 「交给编辑审核」 | `draft_confirm` + `to_editor` |
|
||||
| `finalize` | 「就这样定稿」 | `draft_confirm` + `finalize` |
|
||||
| `revise_draft` | 「第三段太啰嗦，精简」「补个案例」 | `draft_confirm` + `user_revise`，`user_revision_notes`=意见 |
|
||||
| `rewrite_section` | 「重写第二章」 | `user_revise` + 定位信息（章节）写进 notes |
|
||||
| `change_tone/style` | 「更口语一点」 | `user_revise` + notes |
|
||||
|
||||
#### E. 编辑审核类（status ≈ awaiting_review_confirm）
|
||||
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `accept_review` | 「按编辑意见改」 | `review_confirm` + `accept_review`（可带 `selected_issue_indices`） |
|
||||
| `partial_accept_review` | 「只接受第 1、2 条意见」 | `accept_review` + `selected_issue_indices` |
|
||||
| `force_finalize` | 「不改了，强制定稿」 | `review_confirm` + `force_finalize` |
|
||||
| `add_review_opinion` | 「你再补一条：检查数据准确性」 | 追加自定义审核意见（见 5.3 扩展） |
|
||||
|
||||
#### F. 素材不足求助类（status = awaiting_section_help）
|
||||
|
||||
| 意图 | 映射 |
|
||||
|------|------|
|
||||
| `section_supplement` | `section_help` + `supplement`（去重搜） |
|
||||
| `section_loose` | `section_help` + `loose`（放宽续写） |
|
||||
| `section_delete` | `section_help` + `delete` |
|
||||
|
||||
#### G. 问答类（**新**，任意 status，旁路，不改写作状态）
|
||||
|
||||
| 意图 | 示例 | 映射 |
|
||||
|------|------|------|
|
||||
| `ask_about_content` | 「第二章讲了啥？」「为什么用这个论点？」 | 旁路问答（见 5.4） |
|
||||
| `ask_meta` | 「现在进行到哪一步？」「用了哪些素材？」 | 读当前 state 直接回答，不走图 |
|
||||
|
||||
#### H. 控制类（**新/半新**）
|
||||
|
||||
| 意图 | 映射 |
|
||||
|------|------|
|
||||
| `undo/back` | 回退到上一暂停点（依赖 checkpoint 回滚，见 8.4） |
|
||||
| `restart_stage` | 重跑当前阶段 |
|
||||
| `run_background` | 挂起后台自动跑 → `POST /sessions/{id}/background` |
|
||||
| `cancel` | `stream.stop()` |
|
||||
| `unknown/clarify` | 无法判定 → 反问澄清（对话追问，不误触发） |
|
||||
|
||||
> **覆盖「甚至更多」**：G（问答）、H（控制）、以及 D/E 里「素材不足自动回退重搜」都是超出原表单能力的新增。
|
||||
|
||||
### 3.3 「素材不足自动重搜」的判定与回退（用户强调的重点）
|
||||
|
||||
触发场景：用户在**大纲/草稿/审核**阶段提出修改意见，但意见需要的**信息现有素材里没有**。
|
||||
|
||||
判定放在**意图解析层**输出一个 `needResearch: boolean` + `researchQuery`，判定来源二选一（推荐组合）：
|
||||
|
||||
1. **意图解析 LLM 直接判断**：把「当前素材摘要（keywords + 每条 title/来源）」喂给意图解析
|
||||
prompt，让它判断「用户这条意见能否用现有素材满足」。不能 → `needResearch=true` 并给出检索词。
|
||||
2. **执行阶段兜底**：`writer_draft` 严格模式本就有 `blocked_sections` → `pause_section_help`，
|
||||
这是**已有的**素材不足闸门；对话层把它翻译成「素材不足，是否重新检索？」的对话追问。
|
||||
|
||||
回退动作：`needResearch=true` 时，先走 `material_confirm + supplement_search`（或新的
|
||||
「带目标的重搜」协议，见 8.3）补素材，**补完自动继续**用户原本的修改意图（需要在对话层
|
||||
记住「pending 修改意图」，重搜结束后再提交）。
|
||||
|
||||
---
|
||||
|
||||
## 4. 意图解析引擎：方案对比与推荐
|
||||
|
||||
| 方案 | 做法 | 优点 | 缺点 |
|
||||
|------|------|------|------|
|
||||
| **A. 后端 LLM 意图节点**（推荐） | 新增 `/api/ai-writing/intent/parse`（或图内新节点），后端用 LLM 把 `{text, status, contextSnapshot}` → 结构化意图 JSON | 能吃到完整 state（素材/大纲/草稿）；可判 `needResearch`；prompt/模型集中可控；离线内网可用配置模型 | 多一次网络往返（可接受，交互本就等待） |
|
||||
| B. 前端调模型接口分类 | 前端直接调现有 `/api/models` 聊天做分类 | 少一层后端改动 | 前端难拿全后端 state 快照；prompt 分散；鉴权/模型选择复杂 |
|
||||
| C. 规则 + LLM 混合 | 常见短句走关键词规则快路径，兜底 LLM | 快、省 token | 规则易漏、维护成本高；中文表达多样命中率低 |
|
||||
|
||||
**推荐 A（后端 LLM 意图节点）**，并在其上**叠加 C 的快路径**做优化（如「确认/继续/定稿」等
|
||||
高频短句先走轻量规则，未命中再调 LLM）。理由：
|
||||
|
||||
- 意图判定需要**当前写作上下文**（尤其 `needResearch` 判定），后端天然持有 state/checkpoint。
|
||||
- 与现有「配置助手 `setup_writing`」一脉相承，模型选择/离线部署策略统一。
|
||||
- 输出 schema 稳定，前端只消费结构化结果，UI 简单。
|
||||
|
||||
### 4.1 意图解析输出契约（建议）
|
||||
|
||||
```jsonc
|
||||
{
|
||||
"intent": "revise_draft", // 见第 3.2 分类
|
||||
"pausePoint": "draft_confirm", // 若需 resume；否则 null
|
||||
"action": "user_revise", // 映射到现有 action；新意图可为新值
|
||||
"payload": { // 直接可喂 submitIntervention 的字段
|
||||
"userRevisionNotes": "第三段精简，并补一个 2024 年的案例"
|
||||
},
|
||||
"needResearch": true, // 素材是否不足
|
||||
"researchQuery": "2024 行业案例", // needResearch 时的检索意图
|
||||
"answer": null, // 问答类意图的直接回答（旁路）
|
||||
"confidence": 0.86,
|
||||
"clarify": null // 低置信时的反问话术
|
||||
}
|
||||
```
|
||||
|
||||
前端拿到后：
|
||||
- `answer` 非空 → 直接在对话流渲染回答（问答旁路，不动写作状态）。
|
||||
- `clarify` 非空或 `confidence` 低 → 渲染反问，等用户澄清（**防误触发**）。
|
||||
- `needResearch=true` → 先重搜，缓存 `payload` 为 pending，重搜完再提交。
|
||||
- 否则 → `submitIntervention({action, ...payload})` 或 `startWriting(...)`。
|
||||
|
||||
---
|
||||
|
||||
## 5. 前端改造方案
|
||||
|
||||
### 5.1 新页面与路由
|
||||
|
||||
- 新目录：`frontend-web/src/open-canvas/pages/AIWritingChatPage.tsx`（或 `conversational/` 子目录）。
|
||||
- 新路由：`/page/canvas/ai-writing-chat`（在 `OpenCanvasRoutes.tsx` 注册；旧路由保留）。
|
||||
- 侧边栏：`sidebar-menu.ts` 加一个入口（如「AI 写作（对话版）」），或先不加、内部灰度。
|
||||
|
||||
### 5.2 复用 vs 新建
|
||||
|
||||
| 复用（尽量不改） | 新建 |
|
||||
|------------------|------|
|
||||
| `AIWritingContext` / `useAIWritingStream`（状态机 + 流式） | 统一对话输入框组件 `WritingChatComposer` |
|
||||
| `AIWritingDraftPanel`（右侧草稿预览） | 消息流容器 `WritingChatThread`（渲染阶段产物为消息卡片） |
|
||||
| `types/ai-writing.ts`（按需扩展） | 意图解析客户端 `api/ai-writing-intent.ts` |
|
||||
| `ProgressEvent` 事件模型 | 阶段产物 → 消息卡片的适配器 `progressToMessage.ts` |
|
||||
| 会话持久化 REST | 「pending 意图」暂存逻辑（重搜后续跑） |
|
||||
|
||||
**不复用**：`AIWritingForm` / `WritingFormContext` / `InterventionCard`（这三个就是要去掉的表单）。
|
||||
|
||||
### 5.3 交互流程（对话版）
|
||||
|
||||
1. 空闲：用户在底部输入「写一篇关于 X 的深度分析，2000 字」→ 意图解析 `start_writing`
|
||||
→ 组装 `AIWritingRequest` → `startWriting()`。
|
||||
2. 运行中：`ProgressEvent`（`research_progress`/`outline_chunk`/`draft_chunk`…）
|
||||
→ 适配成消息流里的**进度卡片/流式卡片**（复用现有流式视图组件的展示部分）。
|
||||
3. 到达暂停点：不再弹 `InterventionCard`，而是在消息流里展示**阶段产物卡片**
|
||||
（素材列表 / 大纲 / 草稿 / 审核结果）+ 一句提示「你可以说：确认继续 / 补充素材 / 改大纲…」。
|
||||
用户下一句自然语言 → 意图解析 → `submitIntervention`。
|
||||
4. 问答：任意时刻用户提问 → 旁路回答，不打断写作状态。
|
||||
5. 素材不足：意图解析 `needResearch` 或后端 `section_help` → 对话追问「素材不足，帮你重搜？」
|
||||
→ 确认后重搜 → 自动续跑 pending 意图。
|
||||
|
||||
### 5.4 问答旁路（新能力）
|
||||
|
||||
「对写作内容提问」不应改写作 state。两种实现：
|
||||
|
||||
- **轻量（推荐先做）**：意图解析节点直接带回 `answer`（后端能读 state，一次调用出答案）。
|
||||
- **完整**：新起一条**只读**问答线程（lead_agent，context 注入当前草稿/大纲作为背景），
|
||||
类似 `WritingSetupChat` 的 `useThreadStream` 模式，但只读不回填。
|
||||
|
||||
---
|
||||
|
||||
## 6. 后端改造方案
|
||||
|
||||
### 6.1 意图解析入口（二选一）
|
||||
|
||||
- **方案 A1（推荐，独立 REST）**：`app/gateway/routers/ai_writing.py` 新增
|
||||
`POST /api/ai-writing/sessions/{id}/intent`，入参 `{text}`，后端读该 session 的
|
||||
checkpoint state（素材/大纲/草稿摘要）+ 用 `ai_writing` 配置模型解析，返回第 4.1 的契约。
|
||||
- 好处：不改图拓扑；前端拿到结果后再决定调 `startWriting`/`submitIntervention`。
|
||||
- 方案 A2（图内节点）：在图里加 `conversation_router` 节点。改动大、状态耦合高，**不推荐**首期。
|
||||
|
||||
### 6.2 意图解析 prompt
|
||||
|
||||
新增 `agents/ai_writing/prompts/intent_router_prompts.py`（或复用/扩展 `intent_parser.py`）：
|
||||
- 输入：用户文本、`current_status`、素材摘要（keywords + 每条 title/source）、大纲纲要、
|
||||
草稿字数/章节标题、审核意见列表。
|
||||
- 输出：严格 JSON（第 4.1 契约）。要求模型**只在证据充分时给高 confidence**，
|
||||
否则给 `clarify`（防误触发是硬指标）。
|
||||
|
||||
### 6.3 「带目标的重搜」协议扩展（可选增强）
|
||||
|
||||
现有 `supplement_search` 用 `user_query` 追加检索。为支持「改稿意见 → 缺素材 → 重搜 → 自动续跑」，
|
||||
建议在 `material_confirm` 的 resume 里新增可选字段 `resume_intent`（重搜完成后要自动执行的原始意图）。
|
||||
- 前端也可**纯前端**实现（重搜完成事件回来后再自动提交缓存的 pending payload），**首期优先前端实现**，避免动后端协议。
|
||||
|
||||
### 6.4 追加自定义审核意见（`add_review_opinion`）
|
||||
|
||||
`editor`/`pause_review` 目前只接受「接受/强制定稿」。要支持「你再补一条审核意见」，
|
||||
需在 `review_confirm` resume 里接受 `extra_review_notes`，`editor` 或 `writer_draft`
|
||||
修订时并入。首期可先降级为 `user_revise` + notes（把用户的审核诉求当作改稿意见），
|
||||
**完整版**再扩协议。
|
||||
|
||||
### 6.5 不需要改的部分
|
||||
|
||||
- 四个阶段大脑节点（researcher/writer_outline/writer_draft/editor）逻辑不动。
|
||||
- 五个暂停点、checkpoint、StreamBridge、持久化不动。
|
||||
|
||||
---
|
||||
|
||||
## 7. 数据流时序（改稿 + 素材不足自动重搜）
|
||||
|
||||
```
|
||||
用户: "第三段补个2024年的行业案例"
|
||||
│
|
||||
▼ POST /api/ai-writing/sessions/{id}/intent { text }
|
||||
后端: 读 state(素材摘要) + LLM 解析
|
||||
│ 返回 { intent:revise_draft, action:user_revise,
|
||||
│ payload:{userRevisionNotes:...},
|
||||
│ needResearch:true, researchQuery:"2024 行业案例" }
|
||||
▼
|
||||
前端: needResearch=true → 缓存 pending=payload
|
||||
│ 对话追问/直接执行: submitIntervention(material_confirm, supplement_search, user_query="2024 行业案例")
|
||||
▼ LangGraph resume → researcher 补素材 → materials_ready
|
||||
前端: 收到新素材 → 自动提交 pending
|
||||
│ submitIntervention(draft_confirm, user_revise, userRevisionNotes=...)
|
||||
▼ writer_draft 用新素材改稿 → draft_chunk... → draft_ready → pause_draft
|
||||
前端: 消息流展示新草稿 + "已按你的意见改写并补充了案例"
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 8. 关键实现细节与坑
|
||||
|
||||
1. **暂停态识别**：对话层必须知道「现在停在哪个 pause_point」才能正确映射 action。
|
||||
`AIWritingContext` 已有 `currentPause`/`status`，直接读。
|
||||
2. **防误触发**：低置信意图一律走 `clarify` 反问，绝不猜着执行（尤其 `finalize`/`re_search` 这种代价大的）。
|
||||
3. **pending 意图队列**：素材不足重搜是「先插一段、再回原意图」，需要一个小状态机记住 pending，
|
||||
并处理「重搜后用户又改主意」的情况（重搜完成前允许覆盖 pending）。
|
||||
4. **回退/undo（H 类）**：依赖 checkpoint 回滚，`runtime/runs/worker.py` 有 `rollback` 概念，
|
||||
首期可**不做**或只做「重跑当前阶段」。
|
||||
5. **问答旁路不污染 transcript**：问答消息可标记 `kind:'qa'`，与写作进度事件区分持久化。
|
||||
6. **流式产物→消息卡片**：复用现有 `OutlineStreamingView`/`DraftStreamingPreview` 等的**展示部分**，
|
||||
套进消息气泡即可，不必重写渲染。
|
||||
7. **离线内网**：意图解析模型走 `config.yaml → ai_writing` 里的配置模型；检索仍走技能/内网 ES，
|
||||
不联网（与现有并行席位「禁 web_search」的离线策略一致）。
|
||||
|
||||
---
|
||||
|
||||
## 9. 分阶段实施路线图
|
||||
|
||||
**阶段 0 — 骨架（不含意图）**
|
||||
- 新路由 + 新页面 `AIWritingChatPage`，复用 `AIWritingContext`/`useStream`/`DraftPanel`。
|
||||
- 底部对话框先只做「空闲 `start_writing` + 暂停点固定按钮」跑通端到端（相当于把干预卡换成消息）。
|
||||
|
||||
**阶段 1 — 意图解析（核心）**
|
||||
- 后端 `POST /sessions/{id}/intent` + intent prompt，返回第 4.1 契约。
|
||||
- 前端接入：自然语言 → 意图 → `startWriting`/`submitIntervention`，覆盖 A~F 类意图。
|
||||
- 加高频短句规则快路径 + 低置信 `clarify` 反问。
|
||||
|
||||
**阶段 2 — 问答旁路 + 素材不足自动重搜**
|
||||
- G 类问答（先用 intent 节点直接带 `answer`）。
|
||||
- `needResearch` + pending 意图机制（前端实现重搜后自动续跑）。
|
||||
|
||||
**阶段 3 — 增强协议（按需）**
|
||||
- `add_review_opinion` 扩 `review_confirm` 协议、`resume_intent` 后端化、undo/回退。
|
||||
|
||||
**阶段 4 — 打磨**
|
||||
- 消息卡片视觉、历史回看（复用 transcript）、后台挂起入口、灰度到侧边栏。
|
||||
|
||||
---
|
||||
|
||||
## 10. 关键文件清单（落地时对照）
|
||||
|
||||
### 新增（前端）
|
||||
- `open-canvas/pages/AIWritingChatPage.tsx` — 新页面
|
||||
- `open-canvas/components/ai-writing-chat/WritingChatComposer.tsx` — 统一输入框
|
||||
- `open-canvas/components/ai-writing-chat/WritingChatThread.tsx` — 消息流
|
||||
- `open-canvas/components/ai-writing-chat/progressToMessage.ts` — 事件→消息适配
|
||||
- `open-canvas/api/ai-writing-intent.ts` — 意图解析客户端
|
||||
- 路由注册：`open-canvas/OpenCanvasRoutes.tsx`（+ 可选 `sidebar-menu.ts`）
|
||||
|
||||
### 新增（后端）
|
||||
- `app/gateway/routers/ai_writing.py` — 加 `POST /sessions/{id}/intent`
|
||||
- `agents/ai_writing/prompts/intent_router_prompts.py` — 意图 prompt
|
||||
- （阶段 3）`state.py` / `nodes/*` — 扩 `review_confirm`、`resume_intent`
|
||||
|
||||
### 复用（尽量不改）
|
||||
- `open-canvas/contexts/AIWritingContext.tsx`、`hooks/useAIWritingStream.ts`
|
||||
- `open-canvas/components/ai-writing/AIWritingDraftPanel.tsx` 及各流式展示组件
|
||||
- `open-canvas/types/ai-writing.ts`（按需扩字段）
|
||||
- `agents/ai_writing/graph.py`、四个阶段节点、五个暂停点
|
||||
|
||||
### 不动
|
||||
- 旧 `open-canvas/pages/AIWritingPage.tsx` 及 `AIWritingForm`/`WritingFormContext`/`InterventionCard`
|
||||
|
||||
---
|
||||
|
||||
## 11. 待确认
|
||||
|
||||
1. 意图解析引擎是否采用**推荐的方案 A（后端 LLM 节点）**？
|
||||
2. 新页面是**独立入口**（侧边栏新增）还是先**内部灰度**（仅直链）？
|
||||
3. 「素材不足自动重搜」首期是否接受**前端实现 pending 续跑**（不动后端 resume 协议）？
|
||||
4. 问答旁路首期是否接受**意图节点直接带答案**（不新起只读线程）？
|
||||
|
||||
> 以上默认按「A / 内部灰度 / 前端 pending / 意图节点带答案」推进，如需调整请指出。
|
||||
@@ -0,0 +1,274 @@
|
||||
# AI 写作 ·「用户自带大纲」手动测试用例
|
||||
|
||||
> 功能：用户在 AI 写作表单粘贴已有大纲后，系统**完全按用户大纲生成正文**，兼容一/二/三级目录。
|
||||
> 入口页面：`http://localhost:5173/#/page/workspace/ai-writing`（或 `2026` 端口经 nginx）
|
||||
|
||||
---
|
||||
|
||||
## 一、测试目标
|
||||
|
||||
验证三件事：
|
||||
1. **识别**——用户粘贴的各种格式大纲，能被正确解析成 章节(##) → 小节(###) → 子要点(####) 三级结构。
|
||||
2. **保真**——生成的大纲标题、层级**与用户粘贴的完全一致**，不增、不删、不改写。
|
||||
3. **兼容**——只有一级 / 一+二级 / 一+二+三级 三种深度都能正确生成对应层级的正文。
|
||||
|
||||
---
|
||||
|
||||
## 二、前置条件
|
||||
|
||||
| 项 | 要求 |
|
||||
|----|------|
|
||||
| 服务 | 后端（Gateway 8001）+ 前端（Vite 5174/5173）均已启动，可正常发起一次普通 AI 写作 |
|
||||
| 模型 | 写作模型已在表单可选、能正常输出（先用不带大纲的普通流程跑通一次作为基线） |
|
||||
| 账号 | 已登录（或无鉴权模式） |
|
||||
| 入口 | 进入 `/page/workspace/ai-writing`，能看到「写作配置选项」表单 |
|
||||
|
||||
---
|
||||
|
||||
## 三、通用操作步骤
|
||||
|
||||
除非用例另有说明，每条用例都按以下步骤执行：
|
||||
|
||||
1. 进入 AI 写作页面，点开**「已有大纲（可选）」** 旁的「我已有大纲，直接粘贴 ▾」。
|
||||
2. 在「写作意图」填一句话（如：`写一篇关于新能源汽车出口的分析文章`）。
|
||||
3. 把用例里的**「输入大纲」**整段粘贴进「已有大纲」输入框。
|
||||
4. 文章类型 / 目标字数 / 素材来源保持默认即可。
|
||||
5. 点击「开始 AI 写作」。
|
||||
6. 等待流程走到 **PAUSE-2「大纲确认」**，对照**「预期大纲结构」**核对左侧/卡片里的大纲。
|
||||
7. 点「确认」继续，等正文生成完，对照**「预期正文层级」**核对右侧草稿的标题层级。
|
||||
|
||||
> 关键校验点都在第 6 步（大纲确认卡）。正文层级（第 7 步）作为二次确认。
|
||||
|
||||
---
|
||||
|
||||
## 四、测试用例
|
||||
|
||||
### TC-01 Markdown 三级 + 文章标题
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
# 新能源汽车出口为何高速增长
|
||||
## 引言
|
||||
### 出口数据创新高
|
||||
#### 同比增幅
|
||||
#### 全球排名变化
|
||||
### 市场背景
|
||||
## 增长动因
|
||||
### 成本优势
|
||||
```
|
||||
|
||||
**预期大纲结构：**
|
||||
- 文章标题 = `新能源汽车出口为何高速增长`（首个孤立 `#` 被识别为标题，不作为章节）
|
||||
- 章节1 `引言`
|
||||
- 小节 `出口数据创新高` → 子要点：`同比增幅`、`全球排名变化`
|
||||
- 小节 `市场背景`（无子要点）
|
||||
- 章节2 `增长动因`
|
||||
- 小节 `成本优势`
|
||||
|
||||
**预期正文层级：** `#` 文章标题 → `##` 引言/增长动因 → `###` 小节 → `####` 同比增幅/全球排名变化，各级都有正文。
|
||||
|
||||
---
|
||||
|
||||
### TC-02 中文编号三级（一、/（一）/ 1.）
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
一、引言
|
||||
（一）背景
|
||||
（二）意义
|
||||
二、现状分析
|
||||
（一）市场数据
|
||||
1. 国内市场
|
||||
2. 海外市场
|
||||
```
|
||||
|
||||
**预期大纲结构：**
|
||||
- 章节1 `引言` → 小节 `背景`、`意义`
|
||||
- 章节2 `现状分析` → 小节 `市场数据` → 子要点 `国内市场`、`海外市场`
|
||||
- 标题留空 → 用「写作意图」兜底（卡片标题约等于意图前 30 字）
|
||||
|
||||
**校验重点：** `一、` > `（一）` > `1.` 的层级顺序正确（符合中文公文惯例）。
|
||||
|
||||
---
|
||||
|
||||
### TC-03 阿拉伯小数点编号（1. / 1.1 / 1.1.1）
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
1. 引言
|
||||
1.1 研究背景
|
||||
1.2 研究意义
|
||||
2. 现状
|
||||
2.1 行业概况
|
||||
```
|
||||
|
||||
**预期大纲结构：**
|
||||
- 章节1 `引言` → 小节 `研究背景`、`研究意义`
|
||||
- 章节2 `现状` → 小节 `行业概况`
|
||||
|
||||
---
|
||||
|
||||
### TC-04 缩进 bullet 表示层级
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
- 引言
|
||||
- 背景
|
||||
- 意义
|
||||
- 现状
|
||||
- 数据
|
||||
- 细节展开
|
||||
```
|
||||
|
||||
**预期大纲结构：**
|
||||
- 章节1 `引言` → 小节 `背景`、`意义`
|
||||
- 章节2 `现状` → 小节 `数据` → 子要点 `细节展开`
|
||||
|
||||
**校验重点：** 缩进深度正确映射成层级。
|
||||
|
||||
---
|
||||
|
||||
### TC-05 只有一级目录（平铺 Markdown）
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
## 引言
|
||||
## 现状
|
||||
## 对策
|
||||
## 结论
|
||||
```
|
||||
|
||||
**预期大纲结构：** 4 个章节 `引言/现状/对策/结论`，**均无小节**。
|
||||
**预期正文层级：** 每个章节 `##` 标题下是一整段连贯正文（不出现 `###`）。
|
||||
|
||||
**校验重点：** 一级目录场景不报错、不被强行拆出多余小节。
|
||||
|
||||
---
|
||||
|
||||
### TC-06 只有一级目录（中文编号）
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
一、引言
|
||||
二、现状
|
||||
三、结论
|
||||
```
|
||||
|
||||
**预期大纲结构：** 3 个章节 `引言/现状/结论`，无小节。与 TC-05 等价，验证另一种一级写法。
|
||||
|
||||
---
|
||||
|
||||
### TC-07 纯文本平铺（无任何标记）
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
引言
|
||||
现状
|
||||
对策
|
||||
总结
|
||||
```
|
||||
|
||||
**预期大纲结构：** 4 个章节 `引言/现状/对策/总结`（每行一个短标题 → 一级章节）。
|
||||
|
||||
**校验重点：** 无编号无缩进的短标题列表也能识别为一级大纲。
|
||||
|
||||
---
|
||||
|
||||
### TC-08 格式混乱 / 散文化（走 LLM 兜底）
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
我想写一篇分析今年AI大模型发展趋势的文章，重点关注国内企业的布局。
|
||||
希望先讲背景，再讲现状，最后给出展望。
|
||||
```
|
||||
|
||||
**预期行为：**
|
||||
- 确定性解析判定为「散文」→ 自动走 **LLM 归一兜底**，整理成一份结构化大纲。
|
||||
- 大纲确认卡能正常显示一份大纲（不报错、不卡死）。
|
||||
- 大纲内容应贴近原意（背景 / 现状 / 展望之类），**但不要求逐字一致**（这一条是兜底鲁棒性，不是保真用例）。
|
||||
|
||||
**校验重点：** 流程不崩、能产出可用大纲。
|
||||
|
||||
---
|
||||
|
||||
### TC-09 保真校验（标题原样保留）
|
||||
|
||||
**输入大纲：**
|
||||
```
|
||||
一、A股市场2024年回顾与展望
|
||||
（一）宏观经济（GDP/CPI）影响
|
||||
（二）流动性与利率
|
||||
二、行业轮动主线
|
||||
```
|
||||
|
||||
**预期大纲结构：** 标题文本**逐字保留**，仅编号被剥离：
|
||||
- 章节1 `A股市场2024年回顾与展望`
|
||||
- 小节 `宏观经济（GDP/CPI）影响`、`流动性与利率`
|
||||
- 章节2 `行业轮动主线`
|
||||
|
||||
**校验重点：** 标题里的中文括号、英文缩写、年份等**不被改写**。
|
||||
|
||||
---
|
||||
|
||||
### TC-10 大纲确认卡里再编辑 / 重新规划
|
||||
|
||||
**前置：** 先用 TC-02 的大纲跑到「大纲确认」卡。
|
||||
|
||||
**操作：**
|
||||
1. 在确认卡里**手动改一个小节标题**（如把 `背景` 改成 `行业背景`）→ 点「确认」。
|
||||
2. 或点「重新规划」并填修改要求（如 `给"现状分析"再加一个小节`）。
|
||||
|
||||
**预期行为：**
|
||||
- 手动改后确认：正文按改后的大纲生成。
|
||||
- 重新规划：基于**用户这份大纲**做编辑式调整（在原结构上加/改），而不是从零另起一份无关大纲。
|
||||
|
||||
**校验重点：** 自带大纲 → 确认 → 编辑 这条回路正常，不会丢掉用户原始结构。
|
||||
|
||||
---
|
||||
|
||||
### TC-11 回归：不填大纲（默认流程不受影响）
|
||||
|
||||
**操作：** 「已有大纲」**留空**，只填写作意图，正常发起。
|
||||
|
||||
**预期行为：** 与功能上线前完全一致——系统自动检索素材、**自动规划大纲**（大纲由 AI 生成，非用户提供）。
|
||||
|
||||
**校验重点：** 增量功能不污染老流程。
|
||||
|
||||
---
|
||||
|
||||
### TC-12 素材关联（仍走检索）
|
||||
|
||||
**前置：** 素材来源选「知识库」，用 TC-02 的大纲发起，等检索出素材并确认。
|
||||
|
||||
**预期行为：**
|
||||
- 走到大纲时，用户的章节会被**关联上检索到的素材**（正文里能体现素材中的事实/数据）。
|
||||
- 关联**只填素材、不改大纲标题与结构**。
|
||||
|
||||
**校验重点：** 「自带大纲」与「检索素材」能共存——大纲是用户的，正文有素材支撑。
|
||||
|
||||
---
|
||||
|
||||
## 五、缺陷记录模板
|
||||
|
||||
| 用例 | 输入大纲（摘要） | 预期 | 实际 | 是否通过 | 备注 |
|
||||
|------|------------------|------|------|----------|------|
|
||||
| TC-01 | Markdown 三级 | 标题+2章节+子要点 | | ☐ | |
|
||||
| TC-02 | 一、/（一）/1. | 中文三级 | | ☐ | |
|
||||
| TC-03 | 1./1.1/1.1.1 | 阿拉伯三级 | | ☐ | |
|
||||
| TC-04 | 缩进 bullet | 缩进三级 | | ☐ | |
|
||||
| TC-05 | 平铺 ## | 4 章节无小节 | | ☐ | |
|
||||
| TC-06 | 平铺 一、 | 3 章节无小节 | | ☐ | |
|
||||
| TC-07 | 纯文本平铺 | 4 章节 | | ☐ | |
|
||||
| TC-08 | 散文 | LLM 兜底产出大纲 | | ☐ | |
|
||||
| TC-09 | 带括号/缩写标题 | 标题逐字保留 | | ☐ | |
|
||||
| TC-10 | 确认卡编辑/重规划 | 在原大纲上调整 | | ☐ | |
|
||||
| TC-11 | 留空 | 走自动规划 | | ☐ | |
|
||||
| TC-12 | 检索+自带大纲 | 素材关联不改结构 | | ☐ | |
|
||||
|
||||
---
|
||||
|
||||
## 六、备注
|
||||
|
||||
- TC-01~07、09 的解析逻辑已有**自动化单测**覆盖（`backend/tests/test_ai_writing_outline_parser.py`，19 个用例），本文档侧重**端到端真机表现**（流式展示、确认卡渲染、正文层级、素材关联）。
|
||||
- 若某条解析结果与预期不符，先看大纲确认卡里的**结构**是否正确（定位是解析问题还是正文生成问题）。
|
||||
- 相关后端实现：`packages/harness/deerflow/agents/ai_writing/outline_parser.py`（解析）、`nodes/writer_outline.py`（自带大纲分支 + 素材关联）。
|
||||
@@ -107,7 +107,7 @@ http://localhost:5173/#/embed/chats/new?sessionId=...&message=...&username=guest
|
||||
|
||||
## 4. 公开 HTTP API（无需鉴权）
|
||||
|
||||
前缀：`/api/public/embed/`（走 Gateway，与 `VITE_BACKEND_BASE_URL` 一致，默认 `http://47.88.25.99:7001/`）。
|
||||
前缀：`/api/public/embed/`（走 Gateway，与 `VITE_BACKEND_BASE_URL` 一致，默认 `http://127.0.0.1:8001`）。
|
||||
|
||||
CSRF / Auth 中间件对 `/api/public/` 放行；请求 **不需要** Cookie。
|
||||
|
||||
@@ -315,7 +315,7 @@ document.getElementById("deerflow-embed").src =
|
||||
### 5.2 轮询直到结束
|
||||
|
||||
```javascript
|
||||
const API = "http://47.88.25.99:7001/"; // VITE_BACKEND_BASE_URL
|
||||
const API = "http://127.0.0.1:8001"; // VITE_BACKEND_BASE_URL
|
||||
|
||||
async function pollUntilFinished(sessionId) {
|
||||
const res = await fetch(
|
||||
@@ -348,7 +348,7 @@ const timer = setInterval(async () => {
|
||||
|
||||
### 5.3 跨域注意
|
||||
|
||||
父页面从浏览器 `fetch` Gateway（如 `http://47.88.25.99:7001/`）时，**Origin 必须是后端 CORS 白名单里的完整地址**（含协议、主机、端口），与 API 地址是否写 `localhost` / `127.0.0.1` 无关。
|
||||
父页面从浏览器 `fetch` Gateway（如 `http://127.0.0.1:8001`）时，**Origin 必须是后端 CORS 白名单里的完整地址**（含协议、主机、端口），与 API 地址是否写 `localhost` / `127.0.0.1` 无关。
|
||||
|
||||
**本地开发**（前端默认 `5174`）请在 Gateway 环境变量中配置：
|
||||
|
||||
|
||||
@@ -1,154 +0,0 @@
|
||||
# 公开页 · iframe 嵌入外观对接说明（主题 / 背景色）
|
||||
|
||||
本文档面向**嵌入方（宿主系统）**，说明如何在用 `<iframe>` 嵌入 DeerFlow 公开页时，从外部控制页面的**整体明暗主题（light/dark）**与**背景色 / 文字色**，使其与宿主站点的视觉风格保持一致。
|
||||
|
||||
前端实现见：`src/core/embed/background.ts`（参数解析与 postMessage 监听）、`src/App.tsx`（主题接入 `forcedTheme`）。
|
||||
|
||||
---
|
||||
|
||||
## 0. 速读
|
||||
|
||||
| 想做什么 | 怎么做 |
|
||||
|----------|--------|
|
||||
| 嵌入时指定深色主题 | iframe `src` 加 `?theme=dark` |
|
||||
| 嵌入时指定背景/文字色 | iframe `src` 加 `?bg=%23001529&text=%23ffffff`（`#` 需编码为 `%23`） |
|
||||
| 宿主切主题时同步切换 | `postMessage({ type: "deerflow:set-theme", theme: "dark" }, "*")` |
|
||||
| 宿主动态改背景色 | `postMessage({ type: "deerflow:set-bg", bg: "#001529", text: "#fff" }, "*")` |
|
||||
| 什么都不传 | 页面使用 DeerFlow 自带默认配色，互不影响 |
|
||||
|
||||
---
|
||||
|
||||
## 1. 适用页面
|
||||
|
||||
以下公开页（免登录、只读）均支持本文档的全部参数：
|
||||
|
||||
| 页面 | 路由 |
|
||||
|------|------|
|
||||
| 会话分享页 | `/#/share/{shareCode}` |
|
||||
| 圆桌分享页 | `/#/roundtable-share/{shareCode}`、`/#/roundtable-share/{shareCode}/report` |
|
||||
| 用户排行榜 | `/#/public/leaderboard` |
|
||||
| 定时任务公开页 | `/#/public/scheduled-tasks/{taskId}`（含 `/runs/{runId}/fullscreen` 全屏页） |
|
||||
|
||||
> 注意：站点使用 **hash 路由**，路径在 `#` 之后。
|
||||
> 嵌入聊天页（`/#/embed/chats/*`、`/#/embed/roundtable`）有自己的主题约定，见 `embed-iframe-chat.md`，不在本文范围内。
|
||||
|
||||
---
|
||||
|
||||
## 2. URL 查询参数（嵌入时一次性指定）
|
||||
|
||||
### 2.1 参数列表
|
||||
|
||||
| 参数 | 别名 | 作用 | 合法值 |
|
||||
|------|------|------|--------|
|
||||
| `theme` | — | 切整体明暗主题 | 见 2.2 取值映射表 |
|
||||
| `bg` | `background` | 页面根容器背景色 | 见 2.3 颜色格式 |
|
||||
| `text` | `fg` | 页面根容器文字色 | 见 2.3 颜色格式 |
|
||||
|
||||
三个参数相互独立，可任意组合；不传或传非法值时该项忽略、沿用默认。
|
||||
|
||||
### 2.2 `theme` 取值映射
|
||||
|
||||
| 传入值（不区分大小写） | 实际主题 |
|
||||
|------|------|
|
||||
| `dark`、`dark-blue`、`深蓝`、`深色` | 深色 |
|
||||
| `light`、`red`、`红白`、`浅色` | 浅色 |
|
||||
| 其他 / 不传 | 忽略，沿用站内默认（浅色） |
|
||||
|
||||
### 2.3 颜色格式（`bg` / `text`）
|
||||
|
||||
颜色值经过白名单校验，仅接受以下写法，其余一律丢弃（防注入）：
|
||||
|
||||
- **hex**：`#fff`、`#ffffffcc` 等 3/4/6/8 位（`#` 可省略；URL 里 `#` 必须编码为 `%23`）
|
||||
- **函数式**：`rgb()` / `rgba()` / `hsl()` / `hsla()`，括号内仅允许数字、逗号、百分号、空格、`/`
|
||||
- **命名色**：纯英文字母，如 `white`、`transparent`
|
||||
|
||||
### 2.4 参数位置
|
||||
|
||||
参数放在 hash 路由之后（推荐）或 `#` 之前都能识别；两处都传时**hash 路由内的优先**：
|
||||
|
||||
```
|
||||
https://<host>/#/public/leaderboard?theme=dark ← 推荐
|
||||
https://<host>/?theme=dark#/public/leaderboard ← 也支持
|
||||
```
|
||||
|
||||
### 2.5 示例
|
||||
|
||||
```html
|
||||
<!-- 深色主题 -->
|
||||
<iframe src="https://<host>/#/public/leaderboard?theme=dark"></iframe>
|
||||
|
||||
<!-- 浅色主题 + 自定义背景与文字色（# 编码为 %23） -->
|
||||
<iframe src="https://<host>/#/share/Ab3xYz?theme=light&bg=%23f5f7fa&text=%23333333"></iframe>
|
||||
|
||||
<!-- 主题与背景叠加：整页走深色变量，根背景用宿主品牌色 -->
|
||||
<iframe src="https://<host>/#/roundtable-share/Ab3xYz?theme=dark&bg=%23001529"></iframe>
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 3. postMessage（嵌入后动态切换）
|
||||
|
||||
适用于宿主站点支持用户切主题、需要 iframe 实时跟随的场景。向 iframe 的 `contentWindow` 发消息即可，**无需等待子页面回执**（监听器在应用挂载时注册）。
|
||||
|
||||
### 3.1 消息格式
|
||||
|
||||
```js
|
||||
const frame = document.getElementById("deerflow-frame");
|
||||
|
||||
// 切整体明暗主题（theme 取值同 2.2）
|
||||
frame.contentWindow.postMessage({ type: "deerflow:set-theme", theme: "dark" }, "*");
|
||||
|
||||
// 改背景色 / 文字色（字段名与 URL 参数一致，bg/background、text/fg 均可）
|
||||
frame.contentWindow.postMessage({ type: "deerflow:set-bg", bg: "#001529", text: "#fff" }, "*");
|
||||
// 兼容别名：type 也可写 "deerflow:set-background"
|
||||
```
|
||||
|
||||
说明：
|
||||
|
||||
- 两类消息相互独立，可只发其一；`set-bg` 里 `bg`、`text` 也可只传其一，未传的字段保持当前值。
|
||||
- 非法颜色值 / 不认识的 `theme` 取值会被忽略，不会清空已有设置。
|
||||
- targetOrigin 建议按需收紧（写 DeerFlow 站点的 origin 而非 `"*"`）。
|
||||
|
||||
### 3.2 宿主主题联动示例
|
||||
|
||||
```js
|
||||
function syncDeerflowTheme(isDark) {
|
||||
const win = document.getElementById("deerflow-frame").contentWindow;
|
||||
win.postMessage({ type: "deerflow:set-theme", theme: isDark ? "dark" : "light" }, "*");
|
||||
win.postMessage(
|
||||
{ type: "deerflow:set-bg", bg: isDark ? "#001529" : "#ffffff" },
|
||||
"*",
|
||||
);
|
||||
}
|
||||
|
||||
// iframe 首次加载用 URL 参数兜底，之后切换走 postMessage
|
||||
myThemeStore.subscribe((theme) => syncDeerflowTheme(theme === "dark"));
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 4. 优先级与默认行为
|
||||
|
||||
同一项配置多处传入时，按以下优先级取值（高 → 低）：
|
||||
|
||||
1. **postMessage** 动态设置的值
|
||||
2. hash 路由内的 URL 参数（`#/路由?theme=...`）
|
||||
3. `#` 之前的 URL 参数（`?theme=...#/路由`）
|
||||
4. 都没传 → DeerFlow 默认配色（浅色主题 + 页面自带 Tailwind 配色）
|
||||
|
||||
其他保证：
|
||||
|
||||
- 外部主题只影响**当前 iframe 的渲染**，不写 localStorage——不会污染用户直接访问 DeerFlow 主站时自己选择的主题。
|
||||
- `bg`/`text` 是内联样式，作用在页面根容器上，优先级高于主题变量；与 `theme` 叠加时以 `bg`/`text` 为准。
|
||||
|
||||
---
|
||||
|
||||
## 5. 常见问题排查
|
||||
|
||||
| 现象 | 排查 |
|
||||
|------|------|
|
||||
| `theme=dark` 不生效 | 确认路由属于第 1 节列出的公开页；确认参数拼在 hash 路由后（`/#/share/xx?theme=dark`），而不是写成了 `/#/share/xx#theme=dark` |
|
||||
| `bg=#fff` 不生效 | URL 里 `#` 没编码会被浏览器截断，必须写 `%23fff`；或直接省略 `#` 写 `bg=fff` |
|
||||
| postMessage 无反应 | 确认消息体是对象且 `type` 拼写为 `deerflow:set-theme` / `deerflow:set-bg`；确认在 iframe `load` 之后发送 |
|
||||
| 颜色被忽略 | 检查是否符合 2.3 白名单格式（如 `linear-gradient(...)` 等复杂值不支持） |
|
||||
| 主题切了但个别区域颜色没变 | 该区域可能使用了固定配色；把页面路由与截图反馈给 DeerFlow 维护方补充 `dark:` 样式 |
|
||||
@@ -0,0 +1,265 @@
|
||||
# 多智能体协调 API 接口文档
|
||||
|
||||
后端路由：`offline-backend-20260512/backend/app/gateway/routers/multi_agent.py`
|
||||
路由前缀：`/api/multi-agent`
|
||||
|
||||
## 概览
|
||||
|
||||
提供一个"总控 + 多个子智能体"的协调编排接口。
|
||||
|
||||
- **总控（leader）** 智能体：通过 `agent_orchestration` 技能给子智能体派活
|
||||
- **子智能体（special）**：执行被派到的具体任务，完成后把交付内容广播给其他线程
|
||||
|
||||
调用方按下面顺序使用：
|
||||
|
||||
1. 调 `POST /api/multi-agent/init` 一次性创建协调智能体 + 给每个子智能体建 `thread`，拿到 `thread_ids` 映射
|
||||
2. 客户端把 `thread_ids` 保管好，后续每次调 `POST /api/multi-agent/run/stream` 时回传
|
||||
3. 先调用 `agent_type="leader"` 拿派活清单，再针对清单里每个子智能体调用 `agent_type="special"`
|
||||
|
||||
|
||||
## 1. 初始化多智能体集群
|
||||
|
||||
### `POST /api/multi-agent/init`
|
||||
|
||||
创建协调智能体并为列表中的每个子智能体建立独立的 `thread`，同时把"职责说明"广播到所有子智能体线程作为初始上下文。
|
||||
|
||||
#### 请求体
|
||||
|
||||
```json
|
||||
{
|
||||
"agent_names": ["taiwan-intelligence-gathering-agent", "us-intelligence-gathering-agent"],
|
||||
"main_agent_name": "main-controller",
|
||||
"model": "deepseek-chat"
|
||||
}
|
||||
```
|
||||
|
||||
| 字段 | 类型 | 必填 | 说明 |
|
||||
|---|---|---|---|
|
||||
| `agent_names` | `string[]` | 是 | 已有子智能体的 id 列表（kebab-case，区分大小写不敏感，自动转小写） |
|
||||
| `main_agent_name` | `string` | 否 | 要创建的协调智能体名，默认 `main-agent`。重名时会复用已存在的同名智能体 |
|
||||
| `model` | `string` | 否 | 协调智能体使用的模型名。不传则用后端默认模型 |
|
||||
|
||||
#### 响应（200）
|
||||
|
||||
```json
|
||||
{
|
||||
"status": "success",
|
||||
"main_agent_name": "main-controller",
|
||||
"thread_ids": {
|
||||
"main-controller": "thread-uuid-xxx",
|
||||
"taiwan-intelligence-gathering-agent": "thread-uuid-yyy",
|
||||
"us-intelligence-gathering-agent": "thread-uuid-zzz"
|
||||
}
|
||||
}
|
||||
```
|
||||
|
||||
`thread_ids` 是一个 `{agent_name: thread_id}` 字典。**前端必须把这个对象完整保管下来**，后续调用 stream 接口需要原样回传。
|
||||
|
||||
#### 错误
|
||||
|
||||
| 状态码 | 触发条件 |
|
||||
|---|---|
|
||||
| `400` | `agent_names` 为空 |
|
||||
| `401` | 未携带 `Authorization` 头 |
|
||||
| `500` | 协调智能体创建失败且不是"已存在"错误 |
|
||||
|
||||
---
|
||||
|
||||
## 2. 流式调度
|
||||
|
||||
### `POST /api/multi-agent/run/stream`
|
||||
|
||||
通过 SSE（Server-Sent Events）流式驱动一个 leader 或 special 智能体跑一轮对话。
|
||||
|
||||
#### 请求体
|
||||
|
||||
```json
|
||||
{
|
||||
"agent_type": "leader",
|
||||
"d_agent_thread_id": {
|
||||
"main-controller": "thread-uuid-xxx",
|
||||
"taiwan-intelligence-gathering-agent": "thread-uuid-yyy"
|
||||
},
|
||||
"agent_name": "main-controller",
|
||||
"new_message": "请分析特朗普的支持率",
|
||||
"skill_stop_names": [],
|
||||
"model": "deepseek-chat"
|
||||
}
|
||||
```
|
||||
|
||||
| 字段 | 类型 | 必填 | 说明 |
|
||||
|---|---|---|---|
|
||||
| `agent_type` | `"leader"` \| `"special"` | 否 | 默认 `leader`。`leader` 调用协调智能体，`special` 调用子智能体 |
|
||||
| `d_agent_thread_id` | `object` | 是 | 由 `/init` 接口返回的 `thread_ids` 映射，原样回传 |
|
||||
| `agent_name` | `string` | 是 | 本轮要驱动的智能体名（必须存在于 `d_agent_thread_id` 中） |
|
||||
| `new_message` | `string` | 是 | 这轮发给智能体的用户消息 |
|
||||
| `skill_stop_names` | `string[]` | 否 | 触发后中断执行的技能名列表（默认空） |
|
||||
| `model` | `string` | 否 | 本轮 LLM 模型名，默认 `deepseek-chat` |
|
||||
|
||||
#### 响应
|
||||
|
||||
`Content-Type: text/event-stream` 的 SSE 长连接，按到达顺序产出多种数据：
|
||||
|
||||
##### a. 透传上游事件
|
||||
|
||||
原样转发底层 `POST /api/threads/{thread_id}/runs/stream` 的事件帧（LangGraph 协议），包含 `event:`、`data:` 等行。前端按 LangGraph SDK 既有方式渲染即可。
|
||||
|
||||
##### b. 最后一帧：调度状态
|
||||
|
||||
流结束前会额外产出一帧 `data:` 行，结构因 `agent_type` 而异：
|
||||
|
||||
**`agent_type == "leader"`：**
|
||||
|
||||
```
|
||||
data: {"status": [["us-intelligence-gathering-agent", "收集美国民众对特朗普的看法"], ["taiwan-intelligence-gathering-agent", "..."]]}
|
||||
```
|
||||
|
||||
`status` 是一个 `[agent_name, task_text]` 二元组数组，代表协调智能体本轮派出去的任务清单。**前端拿到后应当对每个元素再调用一次本接口（`agent_type="special"`）来执行子智能体**。如果数组为空，说明 leader 觉得已有信息足够，不再派活，可以直接展示 leader 的最终文本回答。
|
||||
|
||||
**`agent_type == "special"`：**
|
||||
|
||||
```
|
||||
data: {"status": "子智能体xxx完成yyy工作"}
|
||||
```
|
||||
|
||||
代表子智能体完成本轮工作并已把交付内容广播给其他线程。
|
||||
|
||||
#### 副作用（重要）
|
||||
|
||||
- **leader 调用**：自动把"总控智能体收到用户消息: …"广播到所有非自己的线程；派活后再把"总控智能体给 X 派活, 内容为: …"广播到所有非派活双方的线程
|
||||
- **special 调用**：完成后把"子智能体 X 完成 Y 工作, 交付内容为: …"广播到所有非自己的线程
|
||||
|
||||
这意味着每个线程的对话历史里会逐步累积出全局上下文，无需前端再做合并。
|
||||
|
||||
#### 错误
|
||||
|
||||
| 状态码 | 触发条件 |
|
||||
|---|---|
|
||||
| `400` | `agent_name` 缺失或未在 `d_agent_thread_id` 中 |
|
||||
| `401` | 未携带 `Authorization` 头 |
|
||||
| `502` | 上游 `runs/stream` 返回非 200 / SSE 解析失败 |
|
||||
|
||||
---
|
||||
|
||||
## 3. 前端调用流程示例
|
||||
|
||||
```ts
|
||||
// 1. 拿 token
|
||||
const loginResp = await fetch("/api/v1/auth/login/username", {
|
||||
method: "POST",
|
||||
headers: { "Content-Type": "application/json" },
|
||||
body: JSON.stringify({ username: "yourname" }),
|
||||
});
|
||||
const { access_token } = await loginResp.json();
|
||||
const authHeaders = {
|
||||
Authorization: `Bearer ${access_token}`,
|
||||
"Content-Type": "application/json",
|
||||
};
|
||||
|
||||
// 2. 初始化
|
||||
const initResp = await fetch("/api/multi-agent/init", {
|
||||
method: "POST",
|
||||
headers: authHeaders,
|
||||
body: JSON.stringify({
|
||||
agent_names: [
|
||||
"taiwan-intelligence-gathering-agent",
|
||||
"us-intelligence-gathering-agent",
|
||||
],
|
||||
main_agent_name: "main-controller",
|
||||
model: "deepseek-chat",
|
||||
}),
|
||||
});
|
||||
const { thread_ids } = await initResp.json();
|
||||
|
||||
// 3. 流式调用 leader
|
||||
async function streamLeader(message: string, model = "deepseek-chat") {
|
||||
const resp = await fetch("/api/multi-agent/run/stream", {
|
||||
method: "POST",
|
||||
headers: authHeaders,
|
||||
body: JSON.stringify({
|
||||
agent_type: "leader",
|
||||
d_agent_thread_id: thread_ids,
|
||||
agent_name: "main-controller",
|
||||
new_message: message,
|
||||
skill_stop_names: [],
|
||||
model,
|
||||
}),
|
||||
});
|
||||
const reader = resp.body!.getReader();
|
||||
const decoder = new TextDecoder();
|
||||
let buf = "";
|
||||
let dispatched: [string, string][] = [];
|
||||
|
||||
while (true) {
|
||||
const { done, value } = await reader.read();
|
||||
if (done) break;
|
||||
buf += decoder.decode(value, { stream: true });
|
||||
const lines = buf.split("\n");
|
||||
buf = lines.pop() ?? "";
|
||||
for (const line of lines) {
|
||||
if (!line.startsWith("data:")) continue;
|
||||
try {
|
||||
const payload = JSON.parse(line.slice(5).trim());
|
||||
if (Array.isArray(payload.status)) {
|
||||
dispatched = payload.status;
|
||||
}
|
||||
// 其余事件按 LangGraph SDK 协议自行渲染
|
||||
} catch {
|
||||
/* 非 JSON 帧，可忽略或交给现有渲染器 */
|
||||
}
|
||||
}
|
||||
}
|
||||
return dispatched;
|
||||
}
|
||||
|
||||
// 4. 对派活清单逐个调用 special
|
||||
async function streamSpecial(agentName: string, task: string, model = "deepseek-chat") {
|
||||
return fetch("/api/multi-agent/run/stream", {
|
||||
method: "POST",
|
||||
headers: authHeaders,
|
||||
body: JSON.stringify({
|
||||
agent_type: "special",
|
||||
d_agent_thread_id: thread_ids,
|
||||
agent_name: agentName,
|
||||
new_message: task,
|
||||
skill_stop_names: [],
|
||||
model,
|
||||
}),
|
||||
});
|
||||
// 同样按 SSE 解析
|
||||
}
|
||||
|
||||
// 5. 编排：先 leader 派活 → 依次执行 special → 再让 leader 汇总
|
||||
const dispatched = await streamLeader("请分析特朗普的支持率");
|
||||
for (const [agentName, task] of dispatched) {
|
||||
await streamSpecial(agentName, task);
|
||||
}
|
||||
const next = await streamLeader("根据上面收集到的信息出最终结论");
|
||||
// 当 leader 返回 status: [] 时，意味着汇报完毕，可结束循环
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 4. 注意事项
|
||||
|
||||
- **`model` 字段的取值**：必须是后端 `config.yaml` 中 `models[]` 已配置的模型名（如 `deepseek-chat`、`glm-4`、`gpt-4o` 等）。传错会导致上游 502
|
||||
- **`thread_ids` 必须每次回传**：后端不存储这个映射，是无状态接口
|
||||
- **协调智能体必须装有 `agent_orchestration` 技能**：`/init` 创建时会自动指定 `skills=["agent_orchestration"]`。后端 `skills/` 目录里必须存在该技能定义，否则 leader 无法派活
|
||||
- **同名协调智能体冲突**：`main_agent_name` 重名时后端不会报错（"already exists" 走兼容分支），会复用既有智能体——所以同一个用户多次 init 用同一个名字是安全的，但 SOUL 不会被覆盖
|
||||
- **SSE 断连**：本接口底层 `on_disconnect: "continue"`，断连后上游 run 会继续跑到结束（不会取消）
|
||||
- **并发广播**：广播消息使用 `asyncio.gather` 并发更新所有线程 state，性能上不会随子智能体数线性变慢
|
||||
|
||||
---
|
||||
|
||||
## 5. 相关接口（底层依赖）
|
||||
|
||||
本接口内部转发以下 Gateway 原生接口，了解细节可查：
|
||||
|
||||
| 接口 | 路由 | 用途 |
|
||||
|---|---|---|
|
||||
| 创建 thread | `POST /api/threads` | 为每个智能体建立独立会话 |
|
||||
| 读 thread state | `GET /api/threads/{id}/state` | 广播前取当前消息列表 |
|
||||
| 写 thread state | `POST /api/threads/{id}/state` | 把广播消息追加到 messages |
|
||||
| 取 agent 详情 | `GET /api/agents/{id}` | init 时读取子智能体的 name/description |
|
||||
| 建 agent | `POST /api/agents` | init 时创建协调智能体 |
|
||||
| 流式 run | `POST /api/threads/{id}/runs/stream` | leader/special 的实际执行 |
|
||||
@@ -1,514 +0,0 @@
|
||||
# 多智能体规划 · 后台挂起执行 + 进度可视化 —— 分步实现方案
|
||||
|
||||
> 目标：在「多智能体规划」第二步增加**后台挂起**能力——把当前任务挂到后台，自动一路跑到
|
||||
> 最后一步（Step3 结果绘制完成）为止；历史记录下拉框里进行中的任务显示加载动画；点击进行中
|
||||
> 的任务弹出一个**大弹窗**，里面是**横向的智能体派活链路流程图**（节点 = 各席位智能体，图标 +
|
||||
> 名称，不同颜色，正在发言的节点和它的连接线带加载动画）；弹窗里有「查看」按钮，点击进入正在
|
||||
> 进行中的任务实时观察。
|
||||
|
||||
---
|
||||
|
||||
## 0. 现状与架构事实（动手前必读）
|
||||
|
||||
研究自现有代码，决定了整套方案的形态。**关键结论：当前编排循环跑在前端，离开页面即中断。**
|
||||
|
||||
### 0.1 前端现状
|
||||
|
||||
| 事实 | 位置 |
|
||||
|------|------|
|
||||
| Step2 编排循环（leader → 顺序派 special → leader … 直到 `dispatched===[]` 共识 / `MAX_CYCLES=8`）**完全由前端 hook 驱动** | `src/roundtable-planning/hooks/useStep2Orchestration.ts:992-1204`（recommend）、`1228-1439`（chain） |
|
||||
| 循环由 `currentStep===2` 的 `useEffect` 启停；离开 Step2 的 cleanup 会 `orchestrationCancelRef.current=true` + `abortAllInFlight()` 中断 | `useStep2Orchestration.ts:1468-1541`（尤其 cleanup `1534-1537`、`1471-1477`） |
|
||||
| 共识判定：`leaderRes.dispatched.length===0` → `setHasConsensus(true)` | `useStep2Orchestration.ts:1091-1099` |
|
||||
| 进入 Step3：手动点按钮，前置 `hasConsensus===true` | `Step2Panel.tsx:370-381`、`RoundtablePlanningPage.tsx:586-595` |
|
||||
| Step3「结果绘制」：前端流式调 `roundtable-report` 智能体，写出 `outputs/方案总览.html`，`phase→"done"` 即完成 | `src/roundtable-planning/hooks/useStep3Report.ts:36,51,165,560-596,636-652` |
|
||||
| 草稿持久化：`/api/roundtable-drafts`，含 step1/step2/step3 三段快照 + `furthestStep(1|2|3)`，**无显式 status 字段** | `src/roundtable-planning/api/drafts.ts:25-110,112,184-225` |
|
||||
| 历史下拉 UI（触发按钮 + 列表 + 重命名/删除） | `RoundtablePlanningPage.tsx:771-892` |
|
||||
| 自动保存触发点（Step1 回复完 / Step2 每轮气泡收尾 / Step3 报告保存 / 步骤前进） | `useDraftPersistence.ts:248-316` |
|
||||
| 加载草稿恢复：`handleLoadDraft` → `getDraft` → `cancelAllStreams` → hydrate → `setCurrentStep(furthestStep)` | `useDraftPersistence.ts:308-368` |
|
||||
|
||||
### 0.2 后端现状
|
||||
|
||||
| 事实 | 位置 |
|
||||
|------|------|
|
||||
| `/api/multi-agent/init`：为各席位并行建 thread + 单例总控 `roundtable-coordinator`，SSE 返回 `seat_ready`/`coordinator_ready`/`init_done(thread_ids)` | `app/gateway/routers/multi_agent.py:647-823` |
|
||||
| `/api/multi-agent/run/stream`：驱动**单轮** leader 或 special 的 LLM 执行，末尾 JSON 帧 `status`（派活列表 / `[]` 共识 / `clarification` / `error`） | `multi_agent.py:1076-1142` |
|
||||
| `/api/multi-agent/report/init`：为 Step3 报告建 thread | `multi_agent.py`（report/init） |
|
||||
| **后端不存在编排循环**——leader 派活由前端解释后再逐个发 `run/stream`；后端只做单轮执行 + 把派活/完成「广播」追加进各 thread 历史消息 | `multi_agent.py:24-29,487-502,1108-1142` |
|
||||
| 派活格式：`agent_orchestration` 工具调用 `{agent_name, task}`（旧式 bash 解析兼容） | `multi_agent.py:525-552` |
|
||||
|
||||
### 0.3 可复用模板：AI 写作的「后台多阶段流程」
|
||||
|
||||
AI 写作已经把「后台跑多阶段流程 + 前端 SSE 订阅 + 可 resume + 定时清理」跑通了，**圆桌后台作业直接照搬这套**：
|
||||
|
||||
| 模板要素 | 位置 |
|
||||
|------|------|
|
||||
| LangGraph 图驱动多阶段（researching/writing_outline/writing_draft/reviewing/…），条件路由推进，`interrupt()` 暂停点等用户输入 | `packages/harness/deerflow/agents/ai_writing/graph.py`、`nodes/*`（`pause_material_confirm` 等 `45-290`） |
|
||||
| Session 持久化表 `ai_writing_sessions`：`id/user_id/title/status/draft_*/transcript(LONGTEXT)/...`，`status∈{in_progress,done,error}` | `packages/harness/deerflow/persistence/ai_writing_sessions/model.py:14-39` |
|
||||
| `transcript` 用 `PortableLongText`（MySQL LONGTEXT，避开 64KB），时间用 `BeijingDateTime` | 同上 `27-38` |
|
||||
| 进度走 LangGraph Server 标准流 `stream_mode=["messages-tuple","values"]`，**graph 在 langgraph runtime 里跑，任意 worker 可 join stream**（这是后台健壮性的关键） | `ai_writing.py:1-17`、`langgraph.json:16-18` |
|
||||
| Checkpointer 共享 `AsyncSqliteSaver`，落 `.deer-flow/data/checkpoints.db`（需挂数据卷） | `runtime/checkpointer/async_provider.py` |
|
||||
| Cleanup cron（leader 文件锁、活跃保护、先删 CP 再删 DB、admin 手动触发） | `app/gateway/ai_writing_cleanup.py:26,34-59,111-135,193-215,307-367` |
|
||||
| **多 worker 坑**：模块级内存 dict（`_event_queues`）每 worker 一份会丢会话；SSE 进程内直调会整段 buffer。解法：状态进 DB+checkpointer、SSE 走真 TCP loopback `127.0.0.1`（`trust_env=False` 禁代理） | `multi_agent.py:271-293`、`AIWritingPage.tsx` 文件头注释 §3/§5 |
|
||||
|
||||
### 0.4 决策（已与产品确认）
|
||||
|
||||
1. **后端真后台执行**：把编排循环搬到后端 run，刷新 / 关页 / 换设备都不中断。
|
||||
2. **流程图节点 = 智能体派活链路**：横向排列的席位节点，正在发言的节点 + 连接线加载动画。
|
||||
|
||||
> ⚠️ **本方案最大的工作量在 Phase 2**：把目前跑在前端 TS 里的编排循环（`runOrchestration`/`runChainOrchestration`）
|
||||
> 移植成一个后端 **LangGraph「圆桌总控编排图」**。这是与 AI 写作一致、可在 LangGraph runtime 里
|
||||
> 后台跑且天然 checkpoint 的健壮路径。下文每个阶段都给出验收标准，可独立提测。
|
||||
|
||||
---
|
||||
|
||||
## 1. 总体架构
|
||||
|
||||
```
|
||||
┌─────────────────────────────── 前端 (frontend-web) ───────────────────────────────┐
|
||||
│ Step2Panel ──「后台挂起」按钮──▶ POST /api/roundtable-jobs (start) │
|
||||
│ HistoryDropdown ── 进行中任务显示 spinner ──▶ 点击 ──▶ DispatchChainModal(大弹窗) │
|
||||
│ DispatchChainModal ── 订阅 GET /api/roundtable-jobs/{id}/stream(SSE) │
|
||||
│ 横向派活链路流程图(节点=席位, 活动节点+连线动画) ──「查看」──▶ 进入任务实时观察 │
|
||||
└───────────────────────────────────────────────────────────────────────────────────┘
|
||||
│ SSE 进度 / 快照
|
||||
┌─────────────────────────────── 后端 (backend) ─────────────────────────────────────┐
|
||||
│ /api/roundtable-jobs 路由层(start/get/stream/resume/cancel) ── 仿 ai_writing.py │
|
||||
│ │ 启动后台 run │
|
||||
│ LangGraph「圆桌总控编排图」 roundtable_orchestrator ── 仿 ai_writing graph │
|
||||
│ init → leader_call → parse_dispatch → dispatch_seats(顺序) → loop ↺ │
|
||||
│ → consensus → report_draw(写 HTML) → done （clarification 点 interrupt()) │
|
||||
│ │ 每次状态转移持久化进度 │
|
||||
│ roundtable_jobs 表(status/phase/dispatch_chain/active_speaker/...) + 共享 checkpointer│
|
||||
│ cleanup cron (仿 ai_writing_cleanup.py) │
|
||||
└───────────────────────────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
**核心思路**：后台作业 = 一个 LangGraph 图的一次 run。图节点内部复用现有的单轮执行能力
|
||||
（`multi_agent.py` 里 leader/special 的 run 逻辑），把「解释派活 → 顺序派活 → 回到 leader」的
|
||||
循环从前端搬进图节点。前端从「驱动者」退化为「观察者」：通过 SSE 订阅进度渲染。
|
||||
|
||||
---
|
||||
|
||||
## 2. 数据契约（前后端共享，先定义后实现）
|
||||
|
||||
### 2.1 作业状态机 `JobStatus`
|
||||
|
||||
```
|
||||
queued # 已创建，等待 worker 拉起
|
||||
running # 编排进行中（Step2 派活循环 / Step3 绘制）
|
||||
awaiting_input # 命中 clarification，等用户回答（interrupt 挂起）
|
||||
done # 已跑到 Step3 绘制完成
|
||||
error # 出错终止
|
||||
cancelled # 用户取消
|
||||
```
|
||||
|
||||
### 2.2 作业阶段 `JobPhase`（细化给流程图/进度条用）
|
||||
|
||||
```
|
||||
initializing # 建 thread / 总控就绪
|
||||
leader_thinking # 总控在决策派活
|
||||
dispatching # 正在顺序派活给席位（配合 active_speaker）
|
||||
consensus # 共识达成（dispatched=[]）
|
||||
report_drawing # Step3 报告绘制中
|
||||
report_done # 绘制完成
|
||||
```
|
||||
|
||||
### 2.3 派活链路节点 `DispatchChainNode`（流程图核心数据）
|
||||
|
||||
```ts
|
||||
interface DispatchChainNode {
|
||||
agentId: string; // 席位 agent_id（小写）
|
||||
name: string; // 显示名
|
||||
avatarType: string; // 复用 getAvatarTypeFor(agentId, index) → 决定颜色
|
||||
order: number; // 横向排列次序（chain 模式=链条顺序；recommend 模式=席位/发言顺序）
|
||||
state: 'pending' | 'active' | 'done'; // 节点状态（驱动颜色/动画）
|
||||
lastSpokeCycle?: number; // 最近发言轮次（可选，做 tooltip）
|
||||
}
|
||||
```
|
||||
|
||||
### 2.4 进度事件 `JobProgressEvent`（SSE 推送的归一化事件）
|
||||
|
||||
```ts
|
||||
type JobProgressEvent =
|
||||
| { type: 'status'; status: JobStatus }
|
||||
| { type: 'phase'; phase: JobPhase; cycle: number }
|
||||
| { type: 'chain'; nodes: DispatchChainNode[] } // 派活链路全量/增量
|
||||
| { type: 'active'; agentId: string | null } // 当前发言席位（null=总控/无）
|
||||
| { type: 'consensus'; percentage: number }
|
||||
| { type: 'dialogue'; bubble: Step2DialogueLike } // 新增/收尾的对话气泡（供「查看」实时渲染）
|
||||
| { type: 'clarification'; question: string } // 命中澄清
|
||||
| { type: 'done'; step3: DraftStep3Snapshot } // 终态：最终报告快照
|
||||
| { type: 'error'; detail: string };
|
||||
```
|
||||
|
||||
> SSE 通道可直接复用 LangGraph Server 的 `values` 流（图状态快照），由前端 hook 把 `values`
|
||||
> 映射成上述归一化事件；也可在路由层做映射后再下发。**推荐前端映射**（少一层后端耦合）。
|
||||
|
||||
### 2.5 作业记录 `RoundtableJob`（持久化 + 列表展示）
|
||||
|
||||
```ts
|
||||
interface RoundtableJob {
|
||||
id: string;
|
||||
draftId: string; // 关联的草稿
|
||||
userId: string;
|
||||
status: JobStatus;
|
||||
phase: JobPhase;
|
||||
cycle: number; // 当前轮次（0..MAX_CYCLES）
|
||||
dispatchChain: DispatchChainNode[];
|
||||
activeAgentId: string | null;
|
||||
consensusPercentage: number;
|
||||
pendingClarification: string | null;
|
||||
threadIds: Record<string,string> | null; // 复用 ThreadIdMap
|
||||
coordinatorName: string | null;
|
||||
orchestrationMode: 'recommend' | 'chain';
|
||||
chain?: { id: string; title: string } | null;
|
||||
error: string | null;
|
||||
createdAt: string;
|
||||
updatedAt: string;
|
||||
}
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## Phase 1 —— 后端：圆桌作业持久化模型
|
||||
|
||||
**目标**：建一张 `roundtable_jobs` 表（仿 `ai_writing_sessions`），存作业状态/阶段/派活链路，
|
||||
并在 `roundtable_drafts` 列表里能 join 出当前作业状态。
|
||||
|
||||
**涉及文件（新增/改）**
|
||||
- 新增 `packages/harness/deerflow/persistence/roundtable_jobs/model.py`（ORM 模型，仿 `ai_writing_sessions/model.py:14-39`）
|
||||
- 新增 `.../roundtable_jobs/sql.py`（Repository：`create / get / update_progress / list_by_user / list_older_than / delete`，仿 `ai_writing_sessions/sql.py`）
|
||||
- 新增 alembic 迁移 `persistence/migrations/versions/<date>_create_roundtable_jobs.py`
|
||||
- 改 `roundtable_drafts` 的列表查询：返回里带上 `status`/`jobId`（LEFT JOIN roundtable_jobs，或在 draft 行上冗余一个 `status` 字段，二选一，见下）
|
||||
|
||||
**步骤**
|
||||
1. 字段设计（对齐 §2.5）：`id(PK)`、`draft_id`、`user_id`、`status`、`phase`、`cycle(Int)`、
|
||||
`dispatch_chain(PortableLongText/JSON)`、`active_agent_id`、`consensus_percentage(Int)`、
|
||||
`pending_clarification(Text)`、`thread_ids(JSON)`、`coordinator_name`、`orchestration_mode`、
|
||||
`chain(JSON)`、`error(Text)`、`created_at/updated_at(BeijingDateTime)`。
|
||||
2. `dispatch_chain` / `thread_ids` / `chain` 用 `PortableLongText` 存 JSON 字符串，repo 层序列化。
|
||||
3. 列表 join：给 `DraftMeta` 增加 `status?: JobStatus` 与 `jobId?: string`。
|
||||
**推荐方案**：草稿与作业 1:1（一个草稿最多一个活跃作业），在 draft 列表查询 LEFT JOIN
|
||||
`roundtable_jobs` 取最新作业的 `status`。
|
||||
4. 索引：`(user_id, updated_at)`、`(draft_id)`、`(status)`（cleanup/列表用）。
|
||||
|
||||
**验收**
|
||||
- 迁移可 `alembic upgrade head` 跑通，MySQL 上 `dispatch_chain` 为 `longtext`。
|
||||
- 单测：create→update_progress→get 往返；`list_by_user` 按 `updated_at` 倒序；`list_older_than` 分批。
|
||||
|
||||
---
|
||||
|
||||
## Phase 2 —— 后端：圆桌总控编排 LangGraph 图（核心 / 最大工作量）
|
||||
|
||||
**目标**：把前端的编排循环移植成一个后端 LangGraph 图，可在 LangGraph runtime 里后台跑、天然
|
||||
checkpoint，一路跑到 Step3 绘制完成。
|
||||
|
||||
**涉及文件（新增）**
|
||||
- `packages/harness/deerflow/agents/roundtable_orchestrator/graph.py`（图装配，仿 `ai_writing/graph.py`）
|
||||
- `.../roundtable_orchestrator/state.py`（图状态 schema：intent、selectedAgents、threadIds、coordinatorName、
|
||||
mode/chain、dialogues、dispatchChain、activeAgentId、cycle、consensus、phase、step3 等）
|
||||
- `.../roundtable_orchestrator/nodes/*.py`（各节点）
|
||||
- 在 `langgraph.json` 注册新 assistant `roundtable_orchestrator`（仿 ai_writing assistant 注册，参考 `langgraph.json:16-18`）
|
||||
|
||||
**图节点设计（把前端循环逐段翻译过来）**
|
||||
|
||||
| 节点 | 职责 | 对应前端逻辑 |
|
||||
|------|------|------|
|
||||
| `init` | 复用 `_create_thread_direct` 批量建席位 thread + 总控就绪（或接收前端已建好的 `threadIds` 直接用） | `useStep2Orchestration.ts:1485-1531` / `multi_agent.py:647-823` |
|
||||
| `leader_call` | 跑一轮总控（agent_type=leader），拿 `status`（派活列表 / `[]` / `clarification`） | 单轮逻辑 `multi_agent.py:1076-1142` |
|
||||
| `route_after_leader`（条件边） | `clarification`→`interrupt`；有派活→`dispatch_seats`；`[]`→`consensus` | `useStep2Orchestration.ts:1091-1181` |
|
||||
| `dispatch_seats` | **顺序**派活：对 dispatched 里每个席位依次跑 special run；每开始一个就更新 `activeAgentId` + 进度持久化 | `useStep2Orchestration.ts:1101-1181`（务必串行，勿并行，详见前端 §5.4 注释） |
|
||||
| `clarification_pause` | `interrupt()` 挂起，`status→awaiting_input`，写 `pending_clarification` | 仿 ai_writing `pause_*` `45-290` |
|
||||
| `consensus` | `setHasConsensus`、`consensus=100`，落库后进入 `report_draw` | `1091-1099` |
|
||||
| `report_draw` | 跑 `roundtable-report`（先 report/init 建 thread），流式写 `outputs/方案总览.html`，产出 `DraftStep3Snapshot` | `useStep3Report.ts:91-149,636-652,560-596` |
|
||||
| `done` | `status→done`、写 `step3` 进作业 + 回写草稿 step3 快照 | `RoundtablePlanningPage.tsx:613-617` |
|
||||
|
||||
**关键实现要点**
|
||||
1. **复用单轮执行**：节点内部不要重写 LLM 调用，抽出 `multi_agent.py` 里 leader/special/report 的
|
||||
单轮 run 为可被图节点直接 `await` 的内部函数（去掉 HTTP 层，进程内调 LangGraph）。
|
||||
2. **派活广播**：保留现有把「总控给 X 派了活 / X 完成」追加进各 thread 历史的广播
|
||||
（`multi_agent.py:487-502`），让席位读到全局上下文。
|
||||
3. **MAX_CYCLES 兜底**：图里维护 `cycle`，≥8 强制收敛（对齐 `roundtable-constants.ts:13`）。
|
||||
4. **chain 模式**：`orchestration_mode==='chain'` 时按 `chain.seats` 固定顺序派活，跳过总控自由决策的
|
||||
部分（对齐 `runChainOrchestration` `1228-1439`）。
|
||||
5. **进度落库**：每个节点结束 `update_progress(jobId, {phase, cycle, dispatchChain, activeAgentId,
|
||||
consensus, dialogues_delta})`，前端 SSE 才有实时数据。
|
||||
6. **dialogues 累积**：把每轮气泡写进图状态 `dialogues`，既供「查看」实时渲染，也供作业结束回写草稿 step2。
|
||||
|
||||
**验收**
|
||||
- 用脚本（仿 `scripts/probe_ai_writing_assistant.py`）直接对图发起一次 run，不开前端，能从 intent +
|
||||
selectedAgents 一路跑到写出 `方案总览.html`，作业 `status` 终态 `done`。
|
||||
- 命中 clarification 时图 `interrupt`，`status=awaiting_input`，`pending_clarification` 有值。
|
||||
- chain 模式按链条顺序派活，recommend 模式由总控自由派活，两条路径都能收敛。
|
||||
|
||||
---
|
||||
|
||||
## Phase 3 —— 后端：作业控制 + 进度 SSE 接口
|
||||
|
||||
**目标**：暴露 start/get/stream/resume/cancel，后台 run 走 LangGraph Server（任意 worker 可 join）。
|
||||
|
||||
**涉及文件（新增/改）**
|
||||
- 新增 `app/gateway/routers/roundtable_jobs.py`（仿 `ai_writing.py` 的 CRUD + 控制壳）
|
||||
- 在 gateway `app.py` 注册路由
|
||||
- 新增 `app/gateway/roundtable_jobs_cleanup.py`（仿 `ai_writing_cleanup.py`）+ 在启动处挂 cron
|
||||
|
||||
**接口契约**
|
||||
|
||||
| 方法 | 路径 | 说明 |
|
||||
|------|------|------|
|
||||
| `POST` | `/api/roundtable-jobs` | 入参：`{draftId, intent, selectedAgents, threadIds?, coordinatorName?, model, orchestrationMode, chain?}`。建作业记录(`queued`) + 启动图 run（LangGraph Server thread），返回 `RoundtableJob` |
|
||||
| `GET` | `/api/roundtable-jobs/{id}` | 作业快照（轮询兜底 / 弹窗首帧） |
|
||||
| `GET` | `/api/roundtable-jobs/{id}/stream` | SSE 进度流：透传图 `values`/`messages-tuple`，或映射成 §2.4 事件下发。**走真 TCP loopback**（`multi_agent.py:271-293`），禁代理 |
|
||||
| `POST` | `/api/roundtable-jobs/{id}/resume` | 入参 `{answer}`，回填 clarification → 图 resume，`status→running` |
|
||||
| `POST` | `/api/roundtable-jobs/{id}/cancel` | 取消图 run，`status→cancelled` |
|
||||
| `GET` | `/api/roundtable-jobs?draftId=` | 按草稿查活跃作业（页面恢复用） |
|
||||
| `POST` | `/api/roundtable-jobs/admin/cleanup` | 管理员手动清理（仿 ai_writing admin cleanup `307-367`） |
|
||||
|
||||
**多 worker 健壮性（务必遵守，照搬 ai_writing 教训）**
|
||||
- 作业状态只存 **DB + checkpointer**，**不要**用模块级内存 dict 存 per-session 状态。
|
||||
- SSE 必须经真 TCP `127.0.0.1` loopback，不能 `httpx.ASGITransport(app=...)` 进程内直调（会整段 buffer）。
|
||||
- 图 run 跑在 LangGraph runtime（独立于处理 HTTP 的 worker），刷新/换 worker 都能重新 join stream。
|
||||
- 数据卷：`.deer-flow/data/checkpoints.db` 必须挂出，否则容器重启丢状态。
|
||||
|
||||
**验收**
|
||||
- `POST /start` 后立刻 `GET /{id}` 能拿到 `queued/running`；关掉 SSE 再重连能继续收进度。
|
||||
- 杀掉发起 start 的 worker（或多 worker 轮询），`resume`/`stream` 仍命中同一作业。
|
||||
- cleanup 干跑：活跃作业永不删，先删 CP 再删 DB。
|
||||
|
||||
---
|
||||
|
||||
## Phase 4 —— 前端：作业 API 客户端 + 数据模型扩展
|
||||
|
||||
**目标**：前端有一层干净的作业 API 与类型，历史列表能拿到 status。
|
||||
|
||||
**涉及文件（新增/改）**
|
||||
- 新增 `src/roundtable-planning/api/roundtable-jobs.ts`：
|
||||
```ts
|
||||
export async function startJob(payload: StartJobPayload): Promise<RoundtableJob>
|
||||
export async function getJob(id: string): Promise<RoundtableJob>
|
||||
export function streamJob(id: string, on: (e: JobProgressEvent) => void): () => void // 返回 unsubscribe
|
||||
export async function resumeJob(id: string, answer: string): Promise<void>
|
||||
export async function cancelJob(id: string): Promise<void>
|
||||
```
|
||||
- `streamJob` 复用项目现有 SSE/`streamMultiAgent` 的封装方式（`api/multi-agent.ts:403+`），把
|
||||
LangGraph `values` 映射成 §2.4 的归一化事件。
|
||||
- 改 `src/roundtable-planning/api/drafts.ts`：`DraftMeta` 增 `status?: JobStatus; jobId?: string`
|
||||
（`drafts.ts:76-83` + 字段映射 `122-178` 加 snake_case 转换）。
|
||||
- 新增 `src/roundtable-planning/lib/job-types.ts`：把 §2.1–2.5 的类型集中导出。
|
||||
|
||||
**验收**：`pnpm typecheck` 通过；`getJob`/`streamJob` 能拉到 Phase 3 的数据。
|
||||
|
||||
---
|
||||
|
||||
## Phase 5 —— 前端：Step2「后台挂起」按钮
|
||||
|
||||
**目标**：Step2 顶栏/底栏加按钮，点击把当前任务交给后端后台跑，前端停止本地循环、转为观察者。
|
||||
|
||||
**涉及文件（改）**
|
||||
- `src/roundtable-planning/components/Step2Panel.tsx`（按钮放在现有「进入第三步」附近 `370-381`）
|
||||
- `src/roundtable-planning/pages/RoundtablePlanningPage.tsx`（接线 + 观察者模式）
|
||||
- `src/roundtable-planning/hooks/useStep2Orchestration.ts`（暴露「停止本地循环并交后台」的方法）
|
||||
|
||||
**步骤**
|
||||
1. Step2Panel 加「⏼ 后台挂起」按钮，`onBackgroundSuspend` 回调。
|
||||
2. 页面 `handleBackgroundSuspend`：
|
||||
- `ensureDraft(2)` 拿到 `draftId`（`useDraftPersistence.ts` ensureDraft）。
|
||||
- 收集当前状态：`intentReady`、`selectedAgents`、`threadIds`、`coordinatorName`、`model`、
|
||||
`orchestrationMode`、`chain`（全部来自 `step2`/`step1` 现有状态）。
|
||||
- `startJob(...)` → 拿到 `jobId`。
|
||||
- **停止前端本地编排**：调用 `step2.stopOrchestration()` + 置一个 `backgroundedJobId` 标记，
|
||||
让 `currentStep===2` 的 useEffect **不再自启本地循环**（在 `useStep2Orchestration.ts:1468-1541`
|
||||
的启动条件里加 `&& !backgrounded`）。
|
||||
- toast「已转入后台，可离开页面，任务会自动跑到结果绘制完成」。
|
||||
- 刷新历史列表（`drafts` 重新拉，进行中项即出现 spinner）。
|
||||
3. **观察者模式**：若当前停留在 Step2 且该 draft 有 running 作业，订阅 `streamJob`，把进度事件渲染成
|
||||
对话气泡/活动席位（复用现有 Step2 渲染），而不是本地驱动。
|
||||
|
||||
**注意：避免「双跑」**——后台已接管后，前端绝不能再并行驱动同一批 thread。用 `backgroundedJobId`
|
||||
互斥；加载草稿时若发现有 running 作业，直接进观察者模式。
|
||||
|
||||
**验收**：点「后台挂起」→ 可立即切到别的任务/路由；回来仍在跑；不会出现前后端同时派活。
|
||||
|
||||
---
|
||||
|
||||
## Phase 6 —— 前端：历史下拉框「进行中」加载动画
|
||||
|
||||
**目标**：历史下拉里 `status==='running'`（或 `awaiting_input`）的草稿条目显示加载动画；点击进行中
|
||||
条目不直接 load，而是打开流程图弹窗。
|
||||
|
||||
**涉及文件（改）**
|
||||
- `src/roundtable-planning/pages/RoundtablePlanningPage.tsx` 的 `HistoryDropdown`（`771-892`）
|
||||
|
||||
**步骤**
|
||||
1. 列表项左侧名称前/后，按 `d.status` 渲染状态标记：
|
||||
- `running` → 旋转的 `Loader2`（`animate-spin`）+「进行中」。
|
||||
- `awaiting_input` → 呼吸点 +「待澄清」（琥珀色）。
|
||||
- `done`/无 → 维持现状（显示 `Step {furthestStep}`）。
|
||||
- `error` → 红点 +「出错」。
|
||||
2. 点击行为分流：
|
||||
- `running`/`awaiting_input` → `openDispatchChainModal(d.jobId)`（Phase 7），**不**直接 `onLoad`。
|
||||
- 其它 → 维持 `onLoad(id)`。
|
||||
3. 列表需要 `status`，确保 Phase 4 的 `DraftMeta.status` 已就位；下拉打开时 `refresh()` 拉最新。
|
||||
|
||||
**验收**：后台挂起后，历史下拉对应条目转圈；点击弹出流程图弹窗。
|
||||
|
||||
---
|
||||
|
||||
## Phase 7 —— 前端：派活链路流程图弹窗（横向 / 大弹窗 / 动画）
|
||||
|
||||
**目标**：大弹窗，横向排列的智能体派活链路。节点 = 席位（图标 + 名称），不同颜色；正在发言的
|
||||
节点和它的连接线带加载动画；含「查看」按钮。
|
||||
|
||||
**涉及文件（新增）**
|
||||
- `src/roundtable-planning/components/DispatchChainModal.tsx`
|
||||
|
||||
**实现选型**：链路是「总控 → 席位序列」的线性结构 + 需要对**单个节点和单条连线**做精确动画，
|
||||
**推荐手写 flex 横向布局 + CSS 动画连接线**（轻、可控）。项目已装 `@xyflow/react@12`（带 `animated`
|
||||
边）作为备选，但手写对本场景的「节点 loading + 连线流动」更直接。
|
||||
|
||||
**结构（手写方案草图）**
|
||||
```tsx
|
||||
<Dialog open={open} onOpenChange={onOpenChange}>
|
||||
<DialogContent className="flex max-h-[88vh] flex-col gap-0 sm:max-w-5xl"> {/* 大弹窗 */}
|
||||
<DialogHeader>… 任务标题 + 阶段标签（leader_thinking/dispatching/report_drawing…） …</DialogHeader>
|
||||
|
||||
{/* 横向链路：总控 ──▶ 席位1 ──▶ 席位2 ──▶ … ──▶ [共识] ──▶ [报告绘制] */}
|
||||
<div className="custom-scrollbar overflow-x-auto py-8">
|
||||
<div className="flex items-center gap-0 min-w-max px-6">
|
||||
<CoordinatorNode active={phase==='leader_thinking'} />
|
||||
{nodes.map((n,i)=>(
|
||||
<Fragment key={n.agentId}>
|
||||
<Connector active={n.state==='active'} /> {/* 活动连线：流动虚线动画 */}
|
||||
<ChainNode node={n} /> {/* 活动节点：彩色+脉冲光环 */}
|
||||
</Fragment>
|
||||
))}
|
||||
<Connector active={phase==='consensus'} />
|
||||
<MilestoneNode label="共识" done={consensus===100} />
|
||||
<Connector active={phase==='report_drawing'} />
|
||||
<MilestoneNode label="报告绘制" active={phase==='report_drawing'} done={phase==='report_done'} />
|
||||
</div>
|
||||
</div>
|
||||
|
||||
<DialogFooter className="border-t pt-3 sm:justify-between">
|
||||
<span className="text-xs text-muted-foreground">轮次 {cycle}/8 · {statusLabel}</span>
|
||||
<Button onClick={onView}>查看</Button> {/* Phase 8 */}
|
||||
</DialogFooter>
|
||||
</DialogContent>
|
||||
</Dialog>
|
||||
```
|
||||
|
||||
**节点颜色**：`avatarType = getAvatarTypeFor(agentId, order)` → `chainTagClasses(avatarType)`
|
||||
（复用 `lib/business-chain.ts:56-72` + `lib/roundtable-constants.ts`），与业务链条卡片同一套色。
|
||||
|
||||
**节点状态样式**
|
||||
- `pending`：灰色描边、低透明度。
|
||||
- `active`：席位色填充 + 外圈 `animate-ping`/脉冲光环 + 图标旋转或呼吸。
|
||||
- `done`：席位色实心 + 右上角 ✓。
|
||||
|
||||
**连接线加载动画（active 连线）**：CSS 流动虚线（marching-ants）或渐变扫光：
|
||||
```css
|
||||
/* roundtable-planning.css 追加 */
|
||||
@keyframes chain-flow { to { background-position: 20px 0; } }
|
||||
.chain-connector--active {
|
||||
background-image: linear-gradient(90deg, var(--brand-500) 50%, transparent 50%);
|
||||
background-size: 20px 100%;
|
||||
animation: chain-flow 0.6s linear infinite;
|
||||
}
|
||||
```
|
||||
|
||||
**数据来源**：弹窗 mount 时 `getJob(jobId)` 取首帧，随后 `streamJob(jobId, onEvent)` 实时更新
|
||||
`nodes`/`activeAgentId`/`phase`/`consensus`/`cycle`；`done` 事件后停止动画、全节点置 `done`。
|
||||
弹窗关闭时 `unsubscribe`。
|
||||
|
||||
**验收**：弹窗够大、横向滚动；活动席位节点和它左侧连线在动；阶段推进时高亮顺移；`done` 后静止。
|
||||
|
||||
---
|
||||
|
||||
## Phase 8 —— 前端：「查看」进入进行中任务并实时观察
|
||||
|
||||
**目标**：弹窗「查看」→ 进入该任务，停在 Step2 实时看后台进度（气泡/活动席位随 SSE 刷新），
|
||||
跑完自动呈现 Step3 结果。
|
||||
|
||||
**涉及文件（改）**
|
||||
- `RoundtablePlanningPage.tsx`（接 `onView`）
|
||||
- `useDraftPersistence.ts`（`handleLoadDraft` 增「带作业」分支）
|
||||
- `useStep2Orchestration.ts`（观察者渲染）
|
||||
|
||||
**步骤**
|
||||
1. `onView`：关弹窗 → `handleLoadDraft(draftId)` 恢复该草稿到 Step2（`useDraftPersistence.ts:308-368`），
|
||||
但**跳过本地编排自启**（识别到有 running 作业 → 进观察者模式，对齐 Phase 5 的互斥标记）。
|
||||
2. 订阅 `streamJob(jobId)`，把 `dialogue` 事件 append 进 `step2RoundtableDialogues`、`active` 事件
|
||||
驱动 `activeSpeakerId` 高亮、`consensus` 更新百分比。
|
||||
3. 命中 `awaiting_input`：在 Step2 弹出澄清输入（复用现有 `pendingClarification` UI），用户回答 →
|
||||
`resumeJob(jobId, answer)`（而非本地 `submitClarification`）。
|
||||
4. `done` 事件：用返回的 `DraftStep3Snapshot` 走 `hydrateStep3` + 允许切到 Step3 查看报告
|
||||
（`useStep3Report.hydrateSaved` `708-767`）；作业列表该项变「已完成」。
|
||||
|
||||
**验收**：从历史「查看」进入能实时看到后台在跑；澄清能在前端回答并续跑；跑完能看到最终报告。
|
||||
|
||||
---
|
||||
|
||||
## Phase 9 —— 边界、健壮性与收尾
|
||||
|
||||
1. **双跑互斥**：任一 draft 同一时刻只允许一个执行者（前端本地 or 后台作业）。加载草稿/进入 Step2
|
||||
前先 `GET /api/roundtable-jobs?draftId=`，有 running 就进观察者模式。
|
||||
2. **后台命中澄清**：`awaiting_input` 时历史下拉显示「待澄清」徽标；流程图弹窗显示「等待澄清」；
|
||||
只有「查看」进去才能回答（或在弹窗里直接给一个回答框，调用 `resumeJob`）。
|
||||
3. **取消**：流程图弹窗/Step2 提供「停止后台」→ `cancelJob`，`status→cancelled`，前端可恢复本地编辑。
|
||||
4. **错误恢复**：`error` 事件展示 detail；允许「重试」= 重新 `startJob`（沿用已建 threadIds）。
|
||||
5. **MAX_CYCLES / 预算**：后台图内兜底（对齐前端 `MAX_CYCLES=8`），到顶强制收敛并照常进 Step3。
|
||||
6. **多 worker**：严格遵守 Phase 3 的「状态进 DB+checkpointer、SSE 走 loopback」；部署若多 worker
|
||||
需 sticky 或全靠 DB 态（推荐后者）。
|
||||
7. **清理**：`roundtable_jobs` cleanup cron（活跃保护、先删 CP 再删 DB），与 draft 生命周期对齐。
|
||||
8. **历史列表性能**：draft 列表 join 作业 status 时注意 N+1；用一次 LEFT JOIN 取最新作业。
|
||||
|
||||
---
|
||||
|
||||
## 端到端验证清单（提测前逐条过）
|
||||
|
||||
- [ ] Step2 点「后台挂起」→ 立刻可离开页面 / 切到别的任务，后台继续跑。
|
||||
- [ ] 历史下拉里该任务显示加载动画（转圈）。
|
||||
- [ ] 点击进行中任务 → 大弹窗、横向派活链路、活动节点+连线在动、节点不同颜色。
|
||||
- [ ] 阶段推进（leader→派活→共识→报告绘制）时弹窗高亮顺移。
|
||||
- [ ] 弹窗「查看」→ 进入任务实时观察气泡/活动席位。
|
||||
- [ ] 后台命中澄清 → 列表「待澄清」+ 弹窗提示 + 可回答续跑（`resumeJob`）。
|
||||
- [ ] 一路自动跑到 Step3「绘制完成」，最终报告可在 Step3 查看，作业 `done`。
|
||||
- [ ] 浏览器刷新 / 关页重开后，任务仍在后台跑、状态可恢复（真后台验证）。
|
||||
- [ ] 取消后台 → `cancelled`，无残留双跑。
|
||||
- [ ] 多 worker（或杀 worker）下 stream/resume 不丢会话。
|
||||
- [ ] cleanup 干跑不误删活跃作业。
|
||||
|
||||
---
|
||||
|
||||
## 工作量与排期建议
|
||||
|
||||
| 阶段 | 内容 | 粗估 | 依赖 |
|
||||
|------|------|------|------|
|
||||
| P0 | 数据契约定型（本文 §2） | 0.5d | — |
|
||||
| P1 | 后端作业表 + repo + 迁移 | 1d | P0 |
|
||||
| **P2** | **后端编排图（移植循环）** | **3–5d（核心难点）** | P0,P1 |
|
||||
| P3 | 后端控制 + SSE + cleanup | 1.5d | P1,P2 |
|
||||
| P4 | 前端作业 API + 类型 + draft status | 0.5d | P3 |
|
||||
| P5 | Step2 后台挂起按钮 + 互斥 | 1d | P4 |
|
||||
| P6 | 历史下拉 loading 动画 | 0.5d | P4 |
|
||||
| P7 | 派活链路流程图弹窗 | 1.5d | P4 |
|
||||
| P8 | 「查看」实时观察 + 澄清续跑 | 1.5d | P5,P7 |
|
||||
| P9 | 边界/健壮性/清理 | 1d | 全部 |
|
||||
|
||||
> 建议先打通 **P1→P2→P3** 的后端最小闭环（脚本验证图能后台跑到出报告），再做前端 P4–P8。
|
||||
> P2 是整条链路成败的关键，务必先用脱离前端的探针脚本（仿 `scripts/probe_ai_writing_assistant.py`）跑通。
|
||||
|
||||
---
|
||||
|
||||
## 附：关键参考代码索引
|
||||
|
||||
- 前端编排循环（待移植到后端）：`useStep2Orchestration.ts:992-1204,1228-1439,1468-1541`
|
||||
- 共识判定 / 进 Step3：`useStep2Orchestration.ts:1091-1099`、`Step2Panel.tsx:370-381`、`RoundtablePlanningPage.tsx:586-595`
|
||||
- Step3 绘制完成标志：`useStep3Report.ts:51,165,560-596,636-652`
|
||||
- 草稿模型 / API / 自动保存 / 加载恢复：`api/drafts.ts:25-110,184-225`、`useDraftPersistence.ts:248-316,308-368`
|
||||
- 历史下拉 UI：`RoundtablePlanningPage.tsx:771-892`
|
||||
- 后端单轮执行 / 派活 / 广播：`multi_agent.py:647-823,1076-1142,525-552,487-502`、loopback `271-293`
|
||||
- AI 写作模板：`agents/ai_writing/graph.py`、`ai_writing_sessions/model.py:14-39`、`ai_writing.py`、`ai_writing_cleanup.py:307-367`、`AIWritingPage.tsx` 文件头注释
|
||||
- 颜色体系复用：`lib/business-chain.ts:56-72`、`lib/roundtable-constants.ts`（`getAvatarTypeFor`）
|
||||
- 流程图备选库：`@xyflow/react@12`（已安装）
|
||||
```
|
||||
@@ -1,404 +0,0 @@
|
||||
# 多智能体会商 Step2 多结果线方案
|
||||
|
||||
## 背景
|
||||
|
||||
当前「多智能体会商 / 多智能体规划」里，一个历史草稿只保存一份 `step2` 快照。用户从历史记录加载草稿后，如果再次点击「智能体推荐」并在弹窗里点击「进入研讨」，新的 Step2 研讨会继续写入同一份 `draft.step2`，结果是第一次生成内容会被第二次生成内容覆盖，或者页面只能展示最新一次结果。
|
||||
|
||||
产品期望是：
|
||||
|
||||
- 用户从历史记录进入已有草稿时，进入第二步后，左下角能看到第一次生成结果。
|
||||
- 用户再次点击「智能体推荐」并点击「进入研讨」时，要开启一条新的研讨结果线。
|
||||
- 第二次生成结果显示在左下角，但不能影响第一次生成结果。
|
||||
- 用户之后再次查看同一个历史草稿时，可以分别查看第一次生成结果和第二次生成结果。
|
||||
|
||||
因此，这个需求的核心不是简单增加一个展示区域，而是把 Step2 从「单结果」升级为「一个草稿下多条研讨 run」。
|
||||
|
||||
## 现状
|
||||
|
||||
相关位置：
|
||||
|
||||
- `src/roundtable-planning/api/drafts.ts`
|
||||
- `src/roundtable-planning/pages/RoundtablePlanningPage.tsx`
|
||||
- `src/roundtable-planning/hooks/useStep2Orchestration.ts`
|
||||
- `src/roundtable-planning/components/Step2Panel.tsx`
|
||||
- `src/roundtable-planning/components/RecommendAgentsDialog.tsx`
|
||||
|
||||
当前 `DraftStep2Snapshot` 结构大致是：
|
||||
|
||||
```ts
|
||||
export interface DraftStep2Snapshot {
|
||||
selectedAgents?: Array<{ agent_id: string; name: string }>;
|
||||
threadIds: ThreadIdMap | null;
|
||||
coordinatorName: string | null;
|
||||
step2RoundtableDialogues: unknown[];
|
||||
lastLeaderContent: string;
|
||||
hasConsensus: boolean;
|
||||
consensusPercentage: number;
|
||||
budgetLimit: number;
|
||||
seatStubMessages?: Record<string, unknown[]>;
|
||||
orchestrationMode?: "recommend" | "chain";
|
||||
chain?: { id: string; title: string } | null;
|
||||
}
|
||||
```
|
||||
|
||||
这意味着一个草稿只有一份：
|
||||
|
||||
- 参会智能体
|
||||
- 多智能体 thread 映射
|
||||
- 研讨气泡
|
||||
- 总控最终内容
|
||||
- 共识状态
|
||||
- 沙箱产物 stub
|
||||
- 编排模式和业务链条
|
||||
|
||||
当前 `RoundtablePlanningPage.tsx` 里的 `handleConfirmAgents` 在用户点击「进入研讨」后会：
|
||||
|
||||
1. 规范化选中的 agents。
|
||||
2. 写入 `step2.selectedAgents`。
|
||||
3. 调用 `step2.resetInitGate()`。
|
||||
4. 设置 `orchestrationMode` / `chainMeta`。
|
||||
5. 切到 `currentStep = 2`。
|
||||
|
||||
它没有先把当前 Step2 结果归档成「第一次生成结果」，也没有为新的研讨创建独立 run。因此第二次研讨天然会复用同一个 `draft.step2` 写入位置。
|
||||
|
||||
## 目标模型
|
||||
|
||||
建议引入 Step2 run 概念：一个草稿可以有多条 Step2 结果线，每条结果线保存一份完整的 Step2 快照。
|
||||
|
||||
```ts
|
||||
interface Step2RunSnapshot {
|
||||
id: string;
|
||||
title: string;
|
||||
createdAt: string;
|
||||
source: "initial" | "recommend" | "chain" | "manual";
|
||||
selectedAgents?: Array<{ agent_id: string; name: string }>;
|
||||
threadIds: ThreadIdMap | null;
|
||||
coordinatorName: string | null;
|
||||
step2RoundtableDialogues: unknown[];
|
||||
lastLeaderContent: string;
|
||||
hasConsensus: boolean;
|
||||
consensusPercentage: number;
|
||||
budgetLimit: number;
|
||||
seatStubMessages?: Record<string, unknown[]>;
|
||||
orchestrationMode?: "recommend" | "chain";
|
||||
chain?: { id: string; title: string } | null;
|
||||
step3?: DraftStep3Snapshot | null;
|
||||
}
|
||||
```
|
||||
|
||||
`DraftStep2Snapshot` 保留现有字段作为当前激活 run 的镜像，同时新增 `runs` 和 `activeRunId`：
|
||||
|
||||
```ts
|
||||
export interface DraftStep2Snapshot {
|
||||
activeRunId?: string;
|
||||
runs?: Step2RunSnapshot[];
|
||||
|
||||
selectedAgents?: Array<{ agent_id: string; name: string }>;
|
||||
threadIds: ThreadIdMap | null;
|
||||
coordinatorName: string | null;
|
||||
step2RoundtableDialogues: unknown[];
|
||||
lastLeaderContent: string;
|
||||
hasConsensus: boolean;
|
||||
consensusPercentage: number;
|
||||
budgetLimit: number;
|
||||
seatStubMessages?: Record<string, unknown[]>;
|
||||
orchestrationMode?: "recommend" | "chain";
|
||||
chain?: { id: string; title: string } | null;
|
||||
}
|
||||
```
|
||||
|
||||
这样做的好处：
|
||||
|
||||
- 老代码仍然可以从 `draft.step2.step2RoundtableDialogues` 读取当前结果。
|
||||
- 新 UI 可以从 `draft.step2.runs` 展示第一次、第二次、第三次生成结果。
|
||||
- 老草稿没有 `runs` 时，可以在 hydrate 时把原来的 `step2` 包装成一条默认 run。
|
||||
- 暂时不需要改数据库表结构，因为 `step2` 本身就是 JSON 快照。
|
||||
|
||||
## 行为设计
|
||||
|
||||
### 1. 历史记录加载
|
||||
|
||||
用户点击历史记录加载草稿时：
|
||||
|
||||
1. 拉取完整草稿。
|
||||
2. 如果 `draft.step2.runs` 存在，使用 `activeRunId` 找到当前结果线。
|
||||
3. 如果 `draft.step2.runs` 不存在，但旧字段里有 `step2RoundtableDialogues`，自动包装成一条 run：
|
||||
|
||||
```ts
|
||||
{
|
||||
id: "legacy-run",
|
||||
title: "第一次生成结果",
|
||||
createdAt: draft.updatedAt,
|
||||
source: "initial",
|
||||
...draft.step2
|
||||
}
|
||||
```
|
||||
|
||||
4. hydrate 当前 active run 到 Step2 runtime。
|
||||
5. 左下角展示 run 列表。
|
||||
|
||||
### 2. 再次点击智能体推荐进入研讨
|
||||
|
||||
用户在已有历史草稿中再次点击「智能体推荐」并确认进入研讨时：
|
||||
|
||||
1. 先把当前页面里的 Step2 状态保存为当前 active run。
|
||||
2. 创建一条新的 run：
|
||||
|
||||
```ts
|
||||
{
|
||||
id: crypto.randomUUID(),
|
||||
title: `第 ${runs.length + 1} 次生成结果`,
|
||||
createdAt: new Date().toISOString(),
|
||||
source: result.mode,
|
||||
selectedAgents: normalizedAgents,
|
||||
threadIds: null,
|
||||
coordinatorName: null,
|
||||
step2RoundtableDialogues: [],
|
||||
lastLeaderContent: "",
|
||||
hasConsensus: false,
|
||||
consensusPercentage: 0,
|
||||
budgetLimit: DEFAULT_BUDGET,
|
||||
seatStubMessages: {},
|
||||
orchestrationMode: result.mode,
|
||||
chain: result.mode === "chain" ? result.chain ?? null : null
|
||||
}
|
||||
```
|
||||
|
||||
3. 将 `activeRunId` 切到新 run。
|
||||
4. 清空 Step2 runtime。
|
||||
5. 设置新选中的 agents / mode / chain。
|
||||
6. `resetInitGate()`，进入 Step2 后重新创建一组独立 thread。
|
||||
7. 自动保存草稿。
|
||||
|
||||
关键点：第二次 run 必须使用新的 `threadIds`，不能沿用第一次 run 的 `threadIds`，否则后端 thread 历史会混在一起。
|
||||
|
||||
### 3. 左下角结果列表
|
||||
|
||||
左下角建议展示一个「生成结果」区域：
|
||||
|
||||
- 第一次生成结果
|
||||
- 第二次生成结果
|
||||
- 第三次生成结果
|
||||
|
||||
每条展示：
|
||||
|
||||
- 标题
|
||||
- 创建时间
|
||||
- 状态：研讨中 / 已达成共识 / 未完成 / 已生成报告
|
||||
- 使用模式：智能体推荐 / 业务链条
|
||||
|
||||
点击某条结果时：
|
||||
|
||||
1. 保存当前 active run。
|
||||
2. 切换 `activeRunId`。
|
||||
3. hydrate 被点击的 run。
|
||||
4. 中间 Step2 对话区显示该 run 的研讨气泡。
|
||||
5. 如果该 run 已有 `step3`，进入 Step3 时展示对应报告。
|
||||
|
||||
### 4. Step3 结果归属
|
||||
|
||||
目前 `draft.step3` 是草稿级别的一份报告。如果只保留这一份，第二次结果生成报告后仍会覆盖第一次报告。
|
||||
|
||||
建议：
|
||||
|
||||
- `draft.step3` 继续作为当前 active run 的镜像，兼容现有 Step3 页面。
|
||||
- 每条 `Step2RunSnapshot.step3` 保存自己的报告。
|
||||
- `onReportSaved` 时同时写入当前 active run 的 `step3`。
|
||||
|
||||
这样用户切换到第一次生成结果时，可以看到第一次的 Step3 报告；切换到第二次生成结果时，可以看到第二次的 Step3 报告。
|
||||
|
||||
## 建议改动点
|
||||
|
||||
### 1. `api/drafts.ts`
|
||||
|
||||
扩展类型：
|
||||
|
||||
- 新增 `Step2RunSnapshot`。
|
||||
- `DraftStep2Snapshot` 增加 `activeRunId?: string` 和 `runs?: Step2RunSnapshot[]`。
|
||||
|
||||
同时保留旧字段，避免影响现有保存、加载和 Step3。
|
||||
|
||||
### 2. `useStep2Orchestration.ts`
|
||||
|
||||
新增或暴露两个能力：
|
||||
|
||||
- `getCurrentRunSnapshot()`：从当前 Step2 runtime 生成一条 run 快照。
|
||||
- `hydrateRunSnapshot(run)`：把某条 run 恢复到 Step2 runtime。
|
||||
|
||||
`hydrateFromDraft` 继续存在，但内部可以优先选择 active run 来 hydrate。
|
||||
|
||||
### 3. `RoundtablePlanningPage.tsx`
|
||||
|
||||
新增页面级状态：
|
||||
|
||||
```ts
|
||||
const [step2Runs, setStep2Runs] = useState<Step2RunSnapshot[]>([]);
|
||||
const [activeStep2RunId, setActiveStep2RunId] = useState<string | null>(null);
|
||||
```
|
||||
|
||||
调整几个流程：
|
||||
|
||||
- `getStep2Snapshot` 输出 `runs` 和 `activeRunId`。
|
||||
- `hydrateStep2` 兼容旧草稿并恢复 runs。
|
||||
- `handleConfirmAgents` 在新建研讨前先归档当前 active run，再创建新 run。
|
||||
- `onReportSaved` 把 Step3 报告写回当前 active run。
|
||||
|
||||
### 4. `Step2Panel.tsx`
|
||||
|
||||
新增左下角结果列表展示。建议作为一个独立组件，避免 Step2Panel 继续变大：
|
||||
|
||||
```tsx
|
||||
<Step2RunSwitcher
|
||||
runs={step2Runs}
|
||||
activeRunId={activeStep2RunId}
|
||||
onSelect={handleSelectStep2Run}
|
||||
/>
|
||||
```
|
||||
|
||||
组件职责只做展示和选择，不直接读写 draft。
|
||||
|
||||
### 5. `Step3Panel.tsx`
|
||||
|
||||
传入当前 active run 的报告：
|
||||
|
||||
- 如果当前 run 有 `step3`，展示它。
|
||||
- 如果没有，按现有逻辑生成新报告。
|
||||
- 生成完成后通过 `onReportSaved` 写回 run。
|
||||
|
||||
## 兼容策略
|
||||
|
||||
老草稿没有 `runs` 字段时，不做数据迁移，前端加载时即时包装。
|
||||
|
||||
判断规则：
|
||||
|
||||
```ts
|
||||
function normalizeStep2Runs(step2: DraftStep2Snapshot): {
|
||||
runs: Step2RunSnapshot[];
|
||||
activeRunId: string | null;
|
||||
} {
|
||||
if (step2.runs?.length) {
|
||||
return {
|
||||
runs: step2.runs,
|
||||
activeRunId: step2.activeRunId ?? step2.runs[0].id,
|
||||
};
|
||||
}
|
||||
|
||||
const hasLegacyContent =
|
||||
step2.step2RoundtableDialogues?.length > 0 ||
|
||||
Boolean(step2.threadIds) ||
|
||||
Boolean(step2.lastLeaderContent);
|
||||
|
||||
if (!hasLegacyContent) {
|
||||
return { runs: [], activeRunId: null };
|
||||
}
|
||||
|
||||
const legacyRun: Step2RunSnapshot = {
|
||||
id: "legacy-run",
|
||||
title: "第一次生成结果",
|
||||
createdAt: new Date().toISOString(),
|
||||
source: step2.orchestrationMode ?? "initial",
|
||||
selectedAgents: step2.selectedAgents,
|
||||
threadIds: step2.threadIds,
|
||||
coordinatorName: step2.coordinatorName,
|
||||
step2RoundtableDialogues: step2.step2RoundtableDialogues,
|
||||
lastLeaderContent: step2.lastLeaderContent,
|
||||
hasConsensus: step2.hasConsensus,
|
||||
consensusPercentage: step2.consensusPercentage,
|
||||
budgetLimit: step2.budgetLimit,
|
||||
seatStubMessages: step2.seatStubMessages,
|
||||
orchestrationMode: step2.orchestrationMode,
|
||||
chain: step2.chain,
|
||||
};
|
||||
|
||||
return { runs: [legacyRun], activeRunId: legacyRun.id };
|
||||
}
|
||||
```
|
||||
|
||||
## 风险点
|
||||
|
||||
### 1. 自动保存覆盖
|
||||
|
||||
Step2 现在会在每个气泡收尾时自动保存。如果引入多 run，自动保存必须先把当前 runtime 合并回 active run，再写入 `draft.step2.runs`。
|
||||
|
||||
否则会出现 UI 切到第二次结果，但保存时仍更新第一次 run 的情况。
|
||||
|
||||
### 2. threadIds 混用
|
||||
|
||||
每条 run 必须拥有自己的 `threadIds`。新建 run 时要清空 `threadIds` 并让 Step2 重新 init。
|
||||
|
||||
如果复用旧 thread，会导致两次研讨的后端上下文互相污染。
|
||||
|
||||
### 3. Step3 报告覆盖
|
||||
|
||||
如果只保存草稿级 `step3`，多 run 的报告无法分别查看。
|
||||
|
||||
建议把报告写入当前 run，同时保留草稿级 `step3` 作为 active run 镜像。
|
||||
|
||||
### 4. 正在运行中的 run
|
||||
|
||||
如果配合后台任务能力，run 还需要带 job 信息：
|
||||
|
||||
```ts
|
||||
jobId?: string | null;
|
||||
jobStatus?: JobStatus | null;
|
||||
```
|
||||
|
||||
这样左下角可以显示「第二次生成结果正在运行」，并打开派活链路弹窗。
|
||||
|
||||
## 分阶段实现建议
|
||||
|
||||
### P0：纯前端多 run 快照
|
||||
|
||||
目标：先解决“第一次 / 第二次结果分别可查看”。
|
||||
|
||||
- 扩展 `DraftStep2Snapshot` 类型。
|
||||
- 页面维护 `step2Runs` / `activeStep2RunId`。
|
||||
- `handleConfirmAgents` 新建 run。
|
||||
- 左下角展示 run 列表。
|
||||
- 切换 run 能恢复不同 Step2 对话。
|
||||
|
||||
不改后端表结构。
|
||||
|
||||
### P1：Step3 绑定 run
|
||||
|
||||
目标：每次研讨结果对应自己的报告。
|
||||
|
||||
- `Step2RunSnapshot.step3` 保存报告。
|
||||
- `onReportSaved` 写回 active run。
|
||||
- 切换 run 时同步 Step3 显示。
|
||||
|
||||
### P2：后台任务绑定 run
|
||||
|
||||
目标：和后台挂起执行方案打通。
|
||||
|
||||
- run 增加 `jobId` / `jobStatus`。
|
||||
- 历史列表和左下角都能显示某条 run 正在执行。
|
||||
- 点击正在执行的 run 打开派活链路弹窗。
|
||||
|
||||
### P3：后端 run 表
|
||||
|
||||
如果后续一个草稿下 run 数量很多，或者需要服务端按 run 查询、删除、重命名，再考虑新增后端表：
|
||||
|
||||
- `roundtable_runs`
|
||||
- `draft_id`
|
||||
- `run_id`
|
||||
- `title`
|
||||
- `status`
|
||||
- `step2_snapshot`
|
||||
- `step3_snapshot`
|
||||
- `created_at`
|
||||
- `updated_at`
|
||||
|
||||
短期不建议一开始就上表，先用现有 `step2` JSON 快照落地更快。
|
||||
|
||||
## 验收清单
|
||||
|
||||
- 从历史记录加载旧草稿时，左下角出现「第一次生成结果」。
|
||||
- 点击「第一次生成结果」能看到原来的 Step2 研讨气泡。
|
||||
- 在历史草稿中重新点击「智能体推荐」并「进入研讨」后，左下角新增「第二次生成结果」。
|
||||
- 第二次研讨开始时，Step2 对话区为空或只显示新 run 初始化内容。
|
||||
- 第二次 run 使用新的 `threadIds`，不会复用第一次 run 的后端线程。
|
||||
- 切回「第一次生成结果」时，第一次的对话、共识状态、沙箱产物仍可查看。
|
||||
- 切回「第二次生成结果」时，第二次的对话继续显示。
|
||||
- 刷新页面后再次加载该历史草稿，两条结果线都还在。
|
||||
- 如果两次都生成了 Step3 报告，切换不同结果线时能看到各自对应的报告。
|
||||
@@ -1,198 +0,0 @@
|
||||
# 后台挂起「在当前进度继续」实现方案
|
||||
|
||||
> 状态：**待实现**（仅方案，未改代码）
|
||||
> 关联：[[multi-agent-background-run-dev.md]]、[[multi-agent-frontend-dev.md]]、memory `project-roundtable-background`
|
||||
> 目标：点「后台挂起」后，后端作业**在前端当前研讨进度上接着跑**，而不是新建空线程从头重跑。
|
||||
|
||||
---
|
||||
|
||||
## 1. 现象与根因
|
||||
|
||||
### 现象
|
||||
在第二步研讨进行中点「后台挂起」，转入后台的作业是**从头重新开始执行**（重新初始化席位、从原始 intent 重新派活），而不是接着前端已经讨论到的进度往下走。
|
||||
|
||||
### 根因
|
||||
后端「起作业」路由 `start_job` **从不设 `is_resume`**（默认 `False`），而执行器据此**新建一套全新线程**：
|
||||
|
||||
```python
|
||||
# offline-backend-20260512/backend/app/gateway/roundtable_job_executor.py (_run, 约 215 行)
|
||||
init = getattr(gateway, "init_threads", None)
|
||||
if init is not None and not params.is_resume:
|
||||
thread_ids, coordinator_name = await init() # ← 建全新空线程
|
||||
```
|
||||
|
||||
加上前端 `startBackgroundJob` **故意不传 `threadIds`**、用 `buildInitialIntent()` 作第一轮 leader prompt，于是后台作业 = 空线程 + 原始意图 = 从头重跑。
|
||||
|
||||
> 这是**原设计的有意取舍**。`startBackgroundJob` 注释写明：不复用前端线程，因为「那些线程可能有活跃 run，复用会撞 409」。本方案就是要在可控前提下打开「复用前端线程续跑」这条路。
|
||||
|
||||
### 后端其实已支持续跑
|
||||
澄清续跑路由 `resume_job`（`roundtable_jobs.py` 约 293 行）已经在做「在原线程上接着派活」：
|
||||
|
||||
```python
|
||||
executor.start_job(StartParams(
|
||||
...,
|
||||
seed_message=_RESUME_PREFIX + answer, # 续跑式提示，而非原始 intent
|
||||
thread_ids=row.get("thread_ids"), # 复用作业自己的线程
|
||||
coordinator_name=row.get("coordinator_name") or "roundtable-coordinator",
|
||||
is_resume=True, # ← 关键：跳过 init_threads，直接复用
|
||||
...
|
||||
))
|
||||
```
|
||||
|
||||
执行器在 `is_resume=True` 时直接用 `params.thread_ids` + `params.coordinator_name`（不调 `init_threads`），交给 `run_orchestration` 续跑。**所以能力已存在，`start_job` 只是没走这条路。**
|
||||
|
||||
---
|
||||
|
||||
## 2. 设计目标与关键事实
|
||||
|
||||
- 「同进度续跑」= 后台作业复用**前端当前会话的真实线程**（总控 + 各席位），这些线程的 checkpoint 里已存着到目前为止的完整研讨上下文，续跑 = 在原线程上 post 一条「继续」消息让总控接着派活。
|
||||
- 前端编排是**逐轮**的：每轮一个 `POST /api/multi-agent/run/stream`（leader 或某席位），**轮与轮之间线程空闲**。只有「某一轮在飞时」线程才被占用。
|
||||
- 后端复用现成线程是**已验证路径**（`resume_job` 天天在用），唯一新增的不同点是：这次复用的是「前端会话线程」而非「上一个后台作业的线程」——但线程同属一个后端线程库，复用方式完全一致。
|
||||
|
||||
---
|
||||
|
||||
## 3. 改造点（最小集）
|
||||
|
||||
### 3.1 后端：让 `start_job` 支持复用线程
|
||||
|
||||
文件：`offline-backend-20260512/backend/app/gateway/routers/roundtable_jobs.py`
|
||||
|
||||
1. `StartJobRequest` 增加字段：
|
||||
```python
|
||||
resume: bool = False # true + 携带 threadIds ⇒ 复用这些线程续跑，不新建
|
||||
```
|
||||
2. `start_job` 里把它透传给 `StartParams`：
|
||||
```python
|
||||
is_resume = bool(body.resume and body.threadIds)
|
||||
executor.start_job(StartParams(
|
||||
...,
|
||||
thread_ids=body.threadIds,
|
||||
coordinator_name=body.coordinatorName,
|
||||
is_resume=is_resume, # ← 新增
|
||||
...
|
||||
))
|
||||
```
|
||||
其余执行器逻辑**不动**（它在 `is_resume=True` 时已能复用 `params.thread_ids`）。
|
||||
|
||||
> 幂等闸门（同草稿未终态作业直接返回）保持不变；续跑挂起前前端应保证该草稿没有在跑的旧作业。
|
||||
|
||||
### 3.2 前端 API：补 `resume` 入参
|
||||
|
||||
文件：`frontend-web/src/roundtable-planning/lib/job-types.ts`
|
||||
- `StartJobPayload` 增加 `resume?: boolean`。
|
||||
|
||||
文件：`frontend-web/src/roundtable-planning/api/roundtable-jobs.ts`
|
||||
- `toStartBody` 增加 `resume: p.resume ?? false`。
|
||||
|
||||
### 3.3 前端：`startBackgroundJob` 改为复用线程续跑
|
||||
|
||||
文件：`frontend-web/src/roundtable-planning/pages/RoundtablePlanningPage.tsx`（`startBackgroundJob` 约 1119 行）
|
||||
|
||||
当前端已有现成会话（`step2.threadIds` 且 `step2.coordinatorName` 非空，说明研讨已 init）时，改传：
|
||||
|
||||
```ts
|
||||
const canContinue = !!(step2.threadIds && step2.coordinatorName);
|
||||
await startJob({
|
||||
draftId: draftId ?? undefined,
|
||||
// 续跑：用「继续」式 seed，而不是 buildInitialIntent()（后者会让总控从头理解任务）
|
||||
intent: buildInitialIntent(), // 仍传，作为无线程回退时的首轮 prompt
|
||||
seedMessage: canContinue
|
||||
? "用户已将本次研讨转入后台，请基于现有进展继续上一轮的协调与派活；若已可收口则直接综合输出结论。"
|
||||
: undefined,
|
||||
agents,
|
||||
model: selectedModel || undefined,
|
||||
orchestrationMode,
|
||||
chain: chainMeta,
|
||||
orchestrationPlan: orchestrationMode === "dag" ? orchestrationPlan : null,
|
||||
// 关键三件套：复用前端真实线程 + 协调名 + resume 标志
|
||||
threadIds: canContinue ? step2.threadIds ?? undefined : undefined,
|
||||
coordinatorName: canContinue ? step2.coordinatorName ?? undefined : undefined,
|
||||
resume: canContinue,
|
||||
});
|
||||
```
|
||||
|
||||
> 没有现成会话（极早期、线程未建）时 `canContinue=false`，自动回退到「新建线程 + 原 intent」的旧行为，安全。
|
||||
|
||||
### 3.4 前端：挂起前先停本地编排（避免交出去时线程还在飞）
|
||||
|
||||
文件：同上，`handleBackgroundSuspend`（约 1161 行）
|
||||
|
||||
当前顺序是 `startBackgroundJob()` → `resetToStep1()`（reset 里才 `stopOrchestration`）。复用线程时必须**先停再交**：
|
||||
|
||||
```ts
|
||||
const handleBackgroundSuspend = useCallback(async () => {
|
||||
if (step2.selectedAgents.length === 0) { /* toast 提示后 return */ }
|
||||
// 1. 先彻底停掉前端在飞的那一轮，尽量让线程在交给后台前变空闲
|
||||
step2.stopOrchestration();
|
||||
step2.setIsRoundTableRunning(false);
|
||||
try {
|
||||
// 2. 再起后台作业（复用线程续跑）
|
||||
await startBackgroundJob();
|
||||
} catch (err) { /* toast 失败后 return */ }
|
||||
// 3. 最后重置回第一步（后台作业继续在历史记录里跑）
|
||||
resetToStep1();
|
||||
triggerToast("🚀 已转入后台，将基于当前进度继续研讨…");
|
||||
void drafts.refresh();
|
||||
}, [...]);
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 4. 关键风险：线程占用（409）与三种处理
|
||||
|
||||
复用前端线程的唯一风险：**在某一轮在飞时挂起**，`stopOrchestration()` 只断了客户端 SSE，服务端那次 run 可能还在跑一小会儿；后台作业首轮 leader 向同一总控线程 post → 撞 409 / rollback。
|
||||
|
||||
| 方案 | 做法 | 取舍 |
|
||||
|---|---|---|
|
||||
| **A（推荐起步）** | 仅前端 stop + 后端 `is_resume` 复用，接受小概率 409 | 实现最简，真正同进度续跑；轮间挂起 100% 安全，仅「某轮在飞时挂起」有小概率 409 —— 走编排循环**现有 rollback/重试容错**兜底 |
|
||||
| **B（最稳）** | 同 A，但后端在复用线程续跑前，**先 cancel 这些线程上可能残留的 run**（参考 takeover 里 `cancelJob` 的 `task.cancel()` 思路，落到 thread-run 级别）再 post | 几乎无 409；后端要多写「按 thread_id 取消活跃 run」的逻辑 |
|
||||
| **C（不推荐）** | 不复用线程，新建空线程，但把「目前为止的研讨」塞进 seed_message 让总控续写 | 零 409，但**席位子线程上下文全丢**、不是真正同进度、leader prompt 臃肿 |
|
||||
|
||||
> 建议：先上 **A**，真机验证 409 命中率；若频繁再加 **B** 的后端线程级 cancel。
|
||||
|
||||
---
|
||||
|
||||
## 5. 适用范围（哪些挂起入口改续跑）
|
||||
|
||||
三个触发源（见 `autoSuspendRef`）：
|
||||
|
||||
1. **手动「后台挂起」按钮**（`handleBackgroundSuspend`）——**改续跑**。能在 start 前干净 stop，最适合复用线程。
|
||||
2. **路由卸载自动挂起**（切到其它页面，`fire`）——可改续跑（能走完整异步链）；卸载瞬间 stop 不一定彻底，按 A 方案的容错兜底。
|
||||
3. **刷新/关标签页 keepalive**（`beforeunload` 的 `beacon` → `startJobKeepalive`）——**建议保持新建线程**。关页时无法保证前端 run 已停，复用线程 409 风险最高；这里走「新建线程 + 原 intent」的安全回退即可（用户可在历史记录里点「查看」再用我们已实现的 takeover 接管回前端续跑）。
|
||||
|
||||
> 落地顺序建议：先只改 #1，验证稳定后再决定是否推广到 #2。
|
||||
|
||||
---
|
||||
|
||||
## 6. 与已实现「查看=接管回前端」的关系
|
||||
|
||||
二者是一对对称操作，复用同一套「线程同属一个后端库、可跨前后端续跑」的事实：
|
||||
|
||||
- **后台挂起（本方案）**：前端 → 后端。前端 stop，把前端线程交给后台 `is_resume` 续跑。
|
||||
- **查看接管（已实现，见 memory）**：后端 → 前端。`cancelJob` 停后台，`hydrateFromDraft` 恢复作业真实线程，`resumeLiveTakeover()` 由前端在那些线程上续跑。
|
||||
|
||||
实现本方案时可直接参照 takeover 的线程复用与 `is_resume` 经验。
|
||||
|
||||
---
|
||||
|
||||
## 7. 测试
|
||||
|
||||
后端（`make test`）：
|
||||
- 新增/扩展 `tests/test_roundtable_jobs.py`：`start_job` 带 `resume=true + threadIds` ⇒ `StartParams.is_resume=True` 且**不调** `init_threads`、直接用传入线程；`resume=false` 或无 `threadIds` ⇒ 走 `init_threads` 新建（回归旧行为）。
|
||||
- 回归：`tests/test_roundtable_inprocess_request.py`、`tests/test_roundtable_step3_pipeline.py`（确认续跑不破坏 Step3 双产物）。
|
||||
|
||||
前端（`pnpm typecheck`）+ 真机：
|
||||
- 研讨进行到几轮后点「后台挂起」→ 历史记录里点「查看」→ 确认后台是**接着已有对话**往下派活，而非从第 1 轮重来。
|
||||
- 轮间挂起 / 某轮在飞时挂起两种时机各验一次，观察是否出现 409（A 方案下应被 rollback 容错吸收）。
|
||||
|
||||
---
|
||||
|
||||
## 8. 落地清单（TL;DR）
|
||||
|
||||
- [ ] 后端 `StartJobRequest.resume` 字段 + `start_job` 透传 `is_resume`
|
||||
- [ ] 前端 `StartJobPayload.resume` + `toStartBody`
|
||||
- [ ] 前端 `startBackgroundJob`：有会话时传 `threadIds/coordinatorName/seedMessage/resume`
|
||||
- [ ] 前端 `handleBackgroundSuspend`：调整为 **先 stop 再 start 再 reset**
|
||||
- [ ] 后端测试（resume 复用 vs 新建分支）
|
||||
- [ ] 真机验证「接着进度跑」+ 409 命中率
|
||||
- [ ]（可选）方案 B：后端复用线程前按 thread_id 取消残留 run
|
||||
File diff suppressed because it is too large.
Load diff
@@ -1,610 +0,0 @@
|
||||
# 圆桌 DAG 编排 + 并行执行显示 —— 分步实现需求文档
|
||||
|
||||
> 目标读者：实现本功能的前端 / 后端工程师。
|
||||
> 本文是**可逐步落地**的需求 + 设计文档，每个 Phase 都能独立交付、独立验收。
|
||||
|
||||
---
|
||||
|
||||
## 0. 背景与目标
|
||||
|
||||
### 0.1 现状
|
||||
|
||||
Step 2 圆桌目前有两种编排模式（`orchestrationMode`，见 `hooks/useStep2Orchestration.ts`）：
|
||||
|
||||
- **`free`（自由模式）**：总控（leader）自由决定派活给谁，`runOrchestration` 循环驱动。
|
||||
- **`chain`（线性链条）**：用户把席位排成一条**线性**链 `A → B → C`，`runChainOrchestration` 按顺序逐个派活，最后总控收口。链路用 `DispatchChainModal.tsx` 的**横向流程图**展示（总控 → 席位1 → 席位2 → … → 共识 → 报告）。
|
||||
|
||||
链条数据现在只是一个**有序数组** `selectedAgents`（线性），节点状态类型见 `lib/job-types.ts`（`ChainNodeState = pending | active | done`）。
|
||||
|
||||
### 0.2 目标（本次新增）
|
||||
|
||||
1. **DAG 编排**：用户可视化指定**智能体之间的直接依赖关系**——有的串行（`A → B → C`），有的并行（`A B C` 同时跑）。编排结果是一张**流程图（DAG）**。
|
||||
2. **并行执行**：DAG 里同一层（互不依赖）的智能体**并行**跑，墙钟时间从「串行求和」降到「并行取最慢」。
|
||||
3. **并行禁止写文件**：并行的智能体**禁用文件写入 / 沙箱命令**，避免它们同时打开沙箱互相打架。产物以消息正文交付；落文件留给串行 / 收口阶段。
|
||||
4. **并行高性能显示**：并行时收起左侧、顶部头像条切换、主区同时显示 2 个智能体的消息流，其余**只接流不渲染**，避免页面卡死。
|
||||
|
||||
### 0.3 设计总原则
|
||||
|
||||
- **复用** 现有 special run / leader run / rollback 抢占 / 轻量交付状态接口（`GET /threads/{id}/last-ai-message`），不另起炉灶。
|
||||
- **线性链是 DAG 的特例**：`chain` 模式可平滑迁移成"每层 1 个节点"的 DAG，老数据不破坏。
|
||||
- **分层（Stage）优先于自由连线**：编辑交互首选「分层」模型（每个 stage 放若干并行席位，stage 间串行），实现成本低、覆盖 95% 的"串 / 并"诉求；自由连线画布作为后续可选增强（见 §2.3）。
|
||||
|
||||
### 0.4 已确认的设计决策（✅ 本期按这些实现）
|
||||
|
||||
> 以下三项已与需求方确认锁定，实现时直接按此，不再二选一。
|
||||
|
||||
1. **✅ 编排模型 = 分层 Stage**：本期只做分层（每个 stage 放若干并行席位、stage 间串行）。**自由连线画布本期不做**，仅在数据模型层预留（§1.2）以免将来返工。
|
||||
2. **✅ 并行跳过 leader、前端直派**：DAG 由用户显式指定，执行时前端**直接**对 stage 内席位发起 special run，**不**每个 stage 再跑一轮总控派活；只在最后 `finalSynthesis` 跑一轮 leader 综合（即 §3.2 的方案 B）。
|
||||
3. **✅ 主区固定 2 列，窄屏降 1 列**：并行时主区最多并排显示 2 个被选中智能体的消息流；窄屏（断点见 §5.3）降为 1 列 + 头像条横向滚动。不做 3+ 列。
|
||||
|
||||
---
|
||||
|
||||
## 1. 数据模型
|
||||
|
||||
### 1.1 DAG 的「分层」表示（推荐，主模型）
|
||||
|
||||
把 DAG 表示成**有序的 stage 列表**，每个 stage 内的席位并行、stage 之间串行：
|
||||
|
||||
```ts
|
||||
// lib/dag-types.ts （新建）
|
||||
|
||||
/** 一个并行批次：stage 内所有席位并行执行。 */
|
||||
export interface OrchestrationStage {
|
||||
/** stage 稳定 id（拖拽重排 / diff 用）。 */
|
||||
id: string;
|
||||
/** 本 stage 并行执行的席位（agent_id 列表）。1 个 = 退化为串行节点。 */
|
||||
agentIds: string[];
|
||||
}
|
||||
|
||||
/** 用户编排出的执行计划。stages 顺序即串行顺序。 */
|
||||
export interface OrchestrationPlan {
|
||||
/** 编排模式标识，持久化进 step2。 */
|
||||
mode: "dag";
|
||||
/** 串行执行的批次；每批内并行。 */
|
||||
stages: OrchestrationStage[];
|
||||
/** 收口：是否在所有 stage 完成后由总控综合（默认 true）。 */
|
||||
finalSynthesis: boolean;
|
||||
}
|
||||
```
|
||||
|
||||
- **串行** `A → B → C` = `stages: [{agentIds:[A]}, {agentIds:[B]}, {agentIds:[C]}]`。
|
||||
- **并行** `A B C` = `stages: [{agentIds:[A,B,C]}]`。
|
||||
- **混合** `A →(B C)→ D` = `stages: [{[A]}, {[B,C]}, {[D]}]`。
|
||||
|
||||
> 为什么用分层而不是「邻接表 / 边」：执行调度本就是「拓扑分批」，分层就是分批的结果，前端不必再做拓扑排序，也天然避免环；编辑器交互更直观（拖卡片进 stage）。真正的自由连线 DAG 见 §2.3（可选）。
|
||||
|
||||
### 1.2 通用 DAG 表示（§2.3 自由连线时才需要）
|
||||
|
||||
若将来做自由连线画布，再引入边表示，并在执行前做一次**拓扑排序 → 分层**，落回 §1.1 的 stages 结构给调度器用：
|
||||
|
||||
```ts
|
||||
export interface DagEdge { from: string; to: string } // from 完成后才能跑 to
|
||||
export interface DagGraph { nodes: string[]; edges: DagEdge[] }
|
||||
// 执行前：toposortToStages(graph) -> OrchestrationStage[]
|
||||
```
|
||||
|
||||
### 1.3 持久化
|
||||
|
||||
- 复用 `roundtable_drafts.step2` JSON（见后端 `app.gateway.routers` roundtable-drafts + `hooks/useStep2Orchestration.ts` 里对 `step2` 的读写）。
|
||||
- 在 `step2` 增加字段：`orchestrationPlan: OrchestrationPlan | null`。
|
||||
- `orchestrationMode` 扩展为 `"free" | "chain" | "dag"`；`chain` 旧数据读取时即时转换为单节点-stage 的 `dag`（兼容层，不改老 draft）。
|
||||
|
||||
### 1.4 校验（纯函数，可单测）
|
||||
|
||||
`lib/dag-types.ts` 内提供：
|
||||
|
||||
```ts
|
||||
validatePlan(plan: OrchestrationPlan, selectedAgentIds: string[]): {
|
||||
ok: boolean;
|
||||
errors: string[]; // 面向用户的中文提示
|
||||
}
|
||||
```
|
||||
|
||||
校验项：
|
||||
- 每个 `agentId` 必须在本次选定席位 `selectedAgents` 内（防止编排了不存在的席位 → 复用后端 `_partition_dispatch` 同思路的"必须在 roster 内"）。
|
||||
- 不允许空 stage；不允许同一席位出现在多个 stage（一个席位只跑一次）。
|
||||
- 至少 1 个 stage。
|
||||
- （可选）单 stage 并行度上限提示（如 > 6 并行给 warning：远程模型并发 + 沙箱压力）。
|
||||
|
||||
---
|
||||
|
||||
## 2. 编排编辑器（流程图构建）交互
|
||||
|
||||
### 2.1 入口与整体布局
|
||||
|
||||
- 在 Step 2 顶部模式切换里新增 **「依赖编排」**（`dag`）模式（与现有 自由 / 链条 并列）。
|
||||
- 选择 `dag` 后弹出 / 进入**编排编辑器**（复用 Dialog 大弹窗，参考 `DispatchChainModal` 的尺寸 `w-[94vw] max-w-[1240px]`）。
|
||||
|
||||
### 2.2 推荐交互：分层（Stage）编辑器（Phase 2 实现）
|
||||
|
||||
```
|
||||
┌ 依赖编排 ───────────────────────────────────────── [校验✓] [保存] ┐
|
||||
│ 可选席位（点/拖加入某个 stage） │
|
||||
│ [情报收集] [环境评估] [方案设计] [风险审查] [执行规划] [前端实现] │
|
||||
│ ─────────────────────────────────────────────────────────────── │
|
||||
│ Stage 1（并行） Stage 2（并行） Stage 3 │
|
||||
│ ┌──────────┐ ─────▶ ┌──────────┐ ┌────────┐ ─▶ ┌──────────┐ │
|
||||
│ │ 情报收集 │ │ 方案设计 │ │风险审查│ │ 执行规划 │ │
|
||||
│ └──────────┘ └──────────┘ └────────┘ └──────────┘ │
|
||||
│ [＋ 拖席位到此] [＋] [＋] │
|
||||
│ [＋ 新建 Stage（串在最后）] │
|
||||
│ ─────────────────────────────────────────────────────────────── │
|
||||
│ 说明：同一 Stage 内的席位会**并行**执行；Stage 之间**串行** │
|
||||
│ （后一个 Stage 能看到前面所有 Stage 的交付摘要）。 │
|
||||
└───────────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
交互要点：
|
||||
- **加入 stage**：从顶部可选席位**拖拽**到某个 stage，或点席位 →「加入 Stage N」。
|
||||
- **移动 / 重排**：席位可在 stage 间拖动；stage 可整体左右拖动重排（改串行顺序）。
|
||||
- **删除**：席位 / 空 stage 可删。
|
||||
- **并行度提示**：stage 内 ≥2 席位即标「并行」徽标；过多给 warning。
|
||||
- **实时校验**：`validatePlan` 实时跑，错误在顶部红条提示，「保存」在校验通过前禁用。
|
||||
- **预览流程图**：编辑区下方实时渲染只读流程图缩略（复用 §7 的 DAG 流程图组件）。
|
||||
|
||||
> 拖拽可用现成轻量方案（HTML5 DnD 或 `@dnd-kit`，看项目是否已有依赖；没有就先用「点选 + 加入/移出按钮」的无拖拽版，Phase 2 先上无拖拽、Phase 2.5 再加拖拽）。
|
||||
|
||||
### 2.3 可选增强：自由连线画布（后续，非本期必须）
|
||||
|
||||
- 节点自由摆放 + 拉线建边 → `DagGraph`（§1.2）。
|
||||
- 实现成本高（画布 / 连线 / 命中检测 / 自动布局），且分层模型已覆盖绝大多数诉求。**本期不做**，仅在数据模型上预留 §1.2，避免将来返工。
|
||||
|
||||
### 2.4 验收（Phase 2）
|
||||
|
||||
- 能拖 / 点把席位分配到多个 stage，保存进 `step2.orchestrationPlan`。
|
||||
- 刷新 / 重进 draft 能恢复编排。
|
||||
- 非法编排（空 stage / 席位重复 / 席位不在 roster）有明确中文提示且无法保存。
|
||||
|
||||
---
|
||||
|
||||
## 3. 执行调度（前端编排循环 + 后端复用）
|
||||
|
||||
### 3.1 新增 `runDagOrchestration`（`useStep2Orchestration.ts`）
|
||||
|
||||
与现有 `runOrchestration` / `runChainOrchestration` 并列，逻辑：
|
||||
|
||||
```
|
||||
for (stageIndex, stage of plan.stages):
|
||||
if isStale() break // 复用现有 epoch 抢占（人工干预/暂停）
|
||||
// 1) 总控按本 stage 的席位并行派活（一轮 leader，dispatched = stage.agentIds）
|
||||
// —— 或跳过 leader，前端直接对 stage 内每个席位发起 special run（见 3.2）
|
||||
// 2) 并行跑 stage 内所有席位
|
||||
await Promise.all(stage.agentIds.map(seat =>
|
||||
streamMultiAgent({ agentType:"special", agentName:seat,
|
||||
newMessage: buildStageTask(seat, plan, stageIndex),
|
||||
parallelNoFile: stage.agentIds.length > 1, // §4
|
||||
... })))
|
||||
// 3) 等本 stage 全部 ✅（Promise.all 天然 barrier）再进入下一 stage
|
||||
// 4) finalSynthesis：最后一轮 leader（synthesisMode=true）综合所有 stage 交付
|
||||
```
|
||||
|
||||
要点：
|
||||
- **批内并行**：`Promise.all`（不是现在的 `for await` 串行）。
|
||||
- **批间串行**：`Promise.all` 即 barrier，自动等齐。
|
||||
- **依赖可见性**：进入 stage N 的席位时，把**前面所有 stage 已交付席位的摘要**拼进它的 `newMessage`（复用 §交付状态：网关 `_collect_delivery_status` / `GET last-ai-message?max_chars=N`，前端也可在 stage 边界拉一次摘要拼进 task）。这样 `(B C)` 能看到 `A` 的产出。
|
||||
- **抢占 / 干预 / 暂停**：每个 stage 边界检查 `isStale()`（epoch），与现有自由 / 链条模式一致。
|
||||
- **rollback**：special run 已统一 `multitask_strategy=rollback`，并行多席位各自不同 thread，互不冲突。
|
||||
|
||||
### 3.2 是否每 stage 都走 leader？ — ✅ 已定：前端直派（方案 B）
|
||||
|
||||
**本期采用方案 B**：DAG 由用户**显式**指定，不需要 leader 再"决定派给谁"。前端**直接**对 stage 内每个席位发起 special run，省掉每 stage 的 leader LLM 往返（更快、更可控）；只在所有 stage 完成后、`finalSynthesis` 为真时跑**一轮** leader（`synthesisMode=true`）综合所有交付。
|
||||
|
||||
> （备选方案 A：每 stage 跑一轮 leader 派活——更贴现有架构但每 stage 多一次 leader LLM、更慢。**本期不采用**，留档备查。）
|
||||
|
||||
### 3.3 后端改动
|
||||
|
||||
- **基本不需要新接口**：special run（`POST /api/multi-agent/run/stream` agent_type=special）、leader run、轻量交付读（`GET /threads/{id}/last-ai-message`）、轻量追加（`POST /threads/{id}/messages/append`）都已就绪。
|
||||
- 仅需 §4 的「并行禁写文件」run policy 开关。
|
||||
|
||||
### 3.4 验收（Phase 3）
|
||||
|
||||
- `(B C)` 两席位**同时**出现在执行（后端日志两条 special run 时间重叠；前端两个流并行增长）。
|
||||
- stage 间严格串行（stage 2 不早于 stage 1 全部完成）。
|
||||
- stage N 席位的输入里能看到 stage <N 的交付摘要。
|
||||
|
||||
---
|
||||
|
||||
## 4. 并行时禁止写文件（防沙箱打架）
|
||||
|
||||
### 4.1 后端 run policy
|
||||
|
||||
`app/gateway/roundtable_run_policy.py` 现有角色：`leader` / `seat` / `report`。新增一个角色或开关：
|
||||
|
||||
- 方案：给 `seat` 增加一个「并行变体」`seat_parallel`，其 `excluded_tools` 在 `seat` 基础上**再禁** `write_file` / `bash` / `str_replace` / `present_files`（与 `report` 角色禁写的思路一致）。
|
||||
- `_special_run`（`multi_agent.py`）按请求里的一个新 flag（如 `no_file: true`）选择 `seat_parallel` policy。
|
||||
|
||||
### 4.2 前端传参
|
||||
|
||||
- `streamMultiAgent` 请求体新增 `parallel_no_file?: boolean`（或复用 `no_file`）。
|
||||
- `runDagOrchestration` 在 `stage.agentIds.length > 1`（真并行）时置 `true`；单节点 stage（实际串行）不限制，可正常写文件。
|
||||
|
||||
### 4.3 产物归属
|
||||
|
||||
- 并行席位的产物 = **消息正文**（被 `_collect_delivery_status` / 综合块读取）。
|
||||
- 需要落文件 / 画图的（如 Step3 报告、`roundtable-report`）放在**串行的收口阶段**单独跑，不在并行批次里。
|
||||
|
||||
### 4.4 验收（Phase 4）
|
||||
|
||||
- 并行 stage 的席位 run 日志显示 `Excluded ... ['write_file','bash','str_replace','present_files', ...]`。
|
||||
- 并行期间不再出现多个席位抢同一沙箱 / 写文件冲突。
|
||||
|
||||
---
|
||||
|
||||
## 5. 并行执行的前端显示
|
||||
|
||||
### 5.1 布局总览（并行进行时）
|
||||
|
||||
```
|
||||
┌ 多智能体圆桌会商中心 ─────────────────────────────────────────────┐
|
||||
│ [头像条：进行中的并行智能体] [模型选择 ▾] [暂停会商] │
|
||||
│ (情报)● (方案)● (风险)● (执行) … │
|
||||
│ ↑选中 ↑选中 loading hidden │
|
||||
│ ───────────────────────────────────────────────────────────────── │
|
||||
│ ┌─ 方案设计（选中）──────────┐ ┌─ 风险审查（选中）──────────┐ │
|
||||
│ │ 一、整体架构…（流式渲染） │ │ 1. 合规风险 IDFA/OAID… │ │
|
||||
│ │ … │ │ … │ │
|
||||
│ └────────────────────────────┘ └────────────────────────────┘ │
|
||||
└───────────────────────────────────────────────────────────────────┘
|
||||
（左侧"参与角色"栏在并行时整体收起，腾出横向空间放两列消息）
|
||||
```
|
||||
|
||||
### 5.2 关键交互（与需求逐条对齐）
|
||||
|
||||
1. **收起左侧**：并行执行时把左侧「参与角色 / 人工参与」栏整体折叠（一个 `isParallelRunning` 状态控制；非并行恢复）。
|
||||
2. **头像条**：位置在**消息列表上方、模型选择器左侧**。展示**当前并行批次的所有进行中智能体头像**：
|
||||
- **进行中** → 头像加 loading 动画（复用 `DispatchChainModal` 的 `animate-ping` 光环 + `Loader2Icon animate-spin` 角标）。
|
||||
- **已完成** → 打勾（复用现有 `CheckIcon` 角标）。
|
||||
- **选中** → 高亮描边（当前正在主区显示的那 2 个）。
|
||||
3. **主区同时显示 2 个**：被选中的 2 个智能体的消息列表**并排两列**渲染。
|
||||
4. **多于 2 个先隐藏**：未选中的智能体**不渲染 DOM**（见 §6 性能）。
|
||||
5. **点头像切换**：点击未选中的头像 → 把它设为选中之一（替换当前 2 个里的一个，或按"最近点击保留 2 个"的 LRU 策略）；被换下的智能体**停止渲染但继续接流**。
|
||||
|
||||
### 5.3 选中策略 — ✅ 固定 2 列，窄屏降 1 列
|
||||
|
||||
- 维护 `selectedParallelSeats: string[]`，**最多 2**（固定 2 列）。**不做 3+ 列**。
|
||||
- **响应式**：宽屏 2 列并排；窄屏（建议断点 `< 1024px` / `lg`）降为**1 列**，头像条横向滚动切换；此时 `selectedParallelSeats` 仍可存 2 个，但主区只渲染当前 1 个（点头像切换哪个在前）。
|
||||
- 默认选中本批次**最先开始**的（宽屏 2 个 / 窄屏 1 个）。
|
||||
- 点头像：若已选中则取消；未选中则加入（超过上限时按 LRU 踢掉最早选中的）。被换下的席位**停止渲染但继续接流**（§6）。
|
||||
|
||||
### 5.4 验收（Phase 5）
|
||||
|
||||
- 并行时左侧收起、出现头像条、主区两列。
|
||||
- 进行中头像有 loading；完成打勾；选中有高亮。
|
||||
- 点头像能切换主区显示的智能体，且切换是**即时**的（数据已在内存，见 §6）。
|
||||
|
||||
---
|
||||
|
||||
## 6. 流式高性能策略（避免页面卡死）
|
||||
|
||||
> 核心矛盾：N 个并行席位同时高频 `onTextDelta`，若都 setState + Markdown 渲染会卡死。
|
||||
|
||||
### 6.1 全接收、惰性渲染
|
||||
|
||||
- **所有**并行 run 的流都正常订阅（不丢数据）：每个席位的增量文本**累积进 store / ref**（`Map<agentId, { text, status, charCount }>`）。
|
||||
- **只有被选中（主区显示）的席位**才把累积文本挂到 DOM 渲染（Markdown）。
|
||||
- **未选中**席位：只更新内存 + 头像条上的轻量进度（字数 / loading），**不挂载消息 DOM**。
|
||||
- 切换选中时：把该席位已累积的全文一次性渲染 + 继续接增量。
|
||||
|
||||
### 6.2 渲染节流
|
||||
|
||||
- 选中席位的流式渲染**节流/合帧**：用 `requestAnimationFrame` 或 ~50–100ms 批量 flush 累积的 token，而不是每个 token 一次 setState。
|
||||
- Markdown 渲染对"流式中"可降级（纯文本 / 轻量渲染），**完成后**再做完整 Markdown（现有 `MessageBubble` 的 `streaming` 标志可复用：streaming 时简渲、done 后全渲）。
|
||||
|
||||
### 6.3 头像条进度
|
||||
|
||||
- 头像条只显示**廉价**信息：loading 动画 + 可选字数（`charCount`，从内存 Map 取，节流更新，例如每 500ms 刷一次）。不触发消息体渲染。
|
||||
|
||||
### 6.4 数据流（建议结构）
|
||||
|
||||
```
|
||||
streamMultiAgent(seat).onTextDelta(text) →
|
||||
parallelStore.append(seatId, text) // 纯内存累积，O(1)
|
||||
if seatId ∈ selectedParallelSeats: // 仅选中的进 React state（节流）
|
||||
scheduleFlush(seatId) // rAF / 节流 setState
|
||||
else:
|
||||
bumpCharCount(seatId) // 仅更新头像条进度（节流）
|
||||
```
|
||||
|
||||
### 6.5 验收（Phase 6）
|
||||
|
||||
- 6+ 席位并行时页面不卡（输入 / 滚动 / 切换流畅）。
|
||||
- 切到一个之前隐藏的席位，能立刻看到它**到目前为止的全部**输出（证明流没丢、只是没渲染）。
|
||||
|
||||
---
|
||||
|
||||
## 7. 流程图执行态可视化（DAG 版）
|
||||
|
||||
- 把现有 `DispatchChainModal.tsx`（横向**线性**流程图）扩展 / 复刻成 **DAG 流程图**：
|
||||
- 按 stage 分**列**，stage 内席位**纵向并列**，stage 间用连接线。
|
||||
- 节点状态复用 `ChainNodeState`（pending / active / done）+ 现有动画（`animate-ping` / `Loader2Icon` / `CheckIcon`）。
|
||||
- 末尾接 `共识 → 结果绘制` 里程碑（沿用现有）。
|
||||
- 用途：编辑器里的只读预览（§2.2 底部）+ 执行时的总览（可选小窗 / 顶部缩略）。
|
||||
|
||||
```
|
||||
总控 ─▶ ┌ Stage1 ┐ ─▶ ┌ Stage2 ┐ ─▶ ┌ Stage3 ┐ ─▶ 🤝共识 ─▶ 📊报告
|
||||
│ 情报● │ │ 方案✓ │ │ 执行○ │
|
||||
└────────┘ │ 风险● │ └────────┘
|
||||
└────────┘
|
||||
```
|
||||
|
||||
### 7.1 验收（Phase 7）
|
||||
|
||||
- 编辑器底部 / 执行时能看到 DAG 流程图，节点状态随执行实时更新（active/done）。
|
||||
|
||||
### 7.2 实施结论（2026-06-07，✅ 已完成）
|
||||
|
||||
- **新组件** `components/DagFlowChart.tsx`：纯展示组件，复刻 `DispatchChainModal` 的节点 / 连线 / 里程碑视觉（`animate-ping` 光环 + `Loader2Icon` 转圈 + `CheckIcon` 打勾 + `rt-chain-connector` 连线动画），但按 **stage 分列**：总控 ─▶ ┌Stage1┐ ─▶ ┌Stage2┐ ─▶ 🤝共识 ─▶ 📊结果绘制，stage 内席位**纵向并列**。`compact` 紧凑模式给缩略图用。所有状态由 props 传入（`stages` / `resolveSeat` / `seatStates` / `coordinatorState` / `consensusState` / `reportState`）；不传 `seatStates` → 全 pending（只读预览）。`aggregateStageState` 把 stage 内席位聚合成列状态(任一 active→active、全 done→done)驱动 stage 间连线。
|
||||
- **编辑器只读预览**（§2.2 底部）：`BusinessChainEditorPage` 底部加「编排预览」section，把 `stageGroups`(去空 stage) 映射成 `DagFlowChart` 的 stages + `resolvePreviewSeat`(席位名/配色)，`compact`、全 pending、随编辑实时变。
|
||||
- **执行态总览**（Step 2 顶部缩略）：`useStep2Orchestration` 新增 `dagNodeStates: Record<agentId, ChainNodeState>`，`runDagOrchestration` 跨 stage 维护(进入 stage→该批 active、席位终态→done、`resetStep2State` 清空)；经 `RoundtablePlanningPage` 把 `orchestrationMode/orchestrationPlan/dagNodeStates` 传到 `Step2Panel`，dag 模式下会话区顶部 `compact` 渲染 DagFlowChart。总控/共识里程碑状态由 `Step2Panel` 从节点状态 + `isOrchestrating`/`hasConsensus` 派生（中转轮时无席位 active + isOrchestrating → 总控 active）。
|
||||
- 改动文件：`components/DagFlowChart.tsx`(新)、`pages/BusinessChainEditorPage.tsx`、`hooks/useStep2Orchestration.ts`、`components/Step2Panel.tsx`、`pages/RoundtablePlanningPage.tsx`。typecheck 零新增(基线 65)。
|
||||
- **下次实地验收**：①编辑器底部随分层编辑实时画出流程图;②Step2 dag 模式顶部出现缩略流程图,跑起来后当前 stage 列点亮(active 转圈)、完成打勾、中转时总控亮、最后共识亮。
|
||||
|
||||
---
|
||||
|
||||
## 8. 分步实现计划（每步可独立交付 + 验收）
|
||||
|
||||
| Phase | 内容 | 主要文件 | 依赖 | 可独立验收 |
|
||||
|---|---|---|---|---|
|
||||
| **1** | DAG 数据模型 + 校验 + 持久化字段 | `lib/dag-types.ts`（新）、draft step2 读写 | — | 纯函数单测：`validatePlan` / `chain→dag` 兼容转换 |
|
||||
| **2** | 分层编排编辑器（先无拖拽：点选加入/移出） | `components/DagPlanEditor.tsx`（新）、Step2 模式切换 | P1 | 能编排并保存/恢复，非法编排有提示 |
|
||||
| **2.5** | 编辑器拖拽 | 同上 + `@dnd-kit`（若引入） | P2 | 拖拽分配/重排 |
|
||||
| **3** | `runDagOrchestration`：批内并行 + 批间串行 + 依赖摘要注入 | `hooks/useStep2Orchestration.ts`、`api/multi-agent.ts` | P1 | 后端日志并行 special run 时间重叠；stage 严格串行 |
|
||||
| **4** | 并行禁写文件 run policy | `roundtable_run_policy.py`、`multi_agent.py`、`api/multi-agent.ts` | P3 | 并行 run 的 excluded_tools 含 write_file/bash/... |
|
||||
| **5** | 并行显示骨架：收起左侧 + 头像条 + 两列 + 选中切换 | `components/Step2Panel.tsx`、`components/ParallelSeatStrip.tsx`（新） | P3 | 见 §5.4 |
|
||||
| **6** | 流式高性能：全接收 / 惰性渲染 / 节流 | `hooks/useStep2Orchestration.ts`（parallelStore）、`Step2Panel` | P5 | 见 §6.5（6+ 并行不卡） |
|
||||
| **7** | DAG 流程图执行态可视化 | `components/DagFlowChart.tsx`（复刻自 `DispatchChainModal`） | P2、P3 | 见 §7.1 |
|
||||
| **8** | 边界 / 异常 | 各处 | 全部 | 见 §9 |
|
||||
|
||||
> 落地顺序建议：**P1 → P3 → P4 → P5 → P6**（先把"能并行跑 + 能看 + 不卡"打通），**P2 编辑器**可与 P3 并行开发（先用一份 hardcode/简单 UI 的 plan 驱动 P3 联调），**P7 / P2.5 / 2.3** 作为增强后置。
|
||||
|
||||
---
|
||||
|
||||
## 9. 边界与异常（Phase 8）
|
||||
|
||||
- **某并行节点失败 / 超时**：单个席位失败不拖垮整 stage；标该节点 error，stage 其余 ✅ 仍推进；收口时把"X 失败"如实告知总控（复用交付状态 ⬜ + 失败标记）。
|
||||
- **依赖未满足**：调度器保证 stage 串行，理论不会发生；防御性校验 stage N 启动前 stage <N 必须全部终态（done/failed）。
|
||||
- **人工干预 / 暂停 / 恢复**：在 stage 边界检查 `isStale()`（epoch）；干预指令插入后，从"下一个未开始的 stage"继续（已完成 stage 不重跑）。
|
||||
- **并行禁写文件 vs 席位本想写文件**：在席位 SOUL / 派活提示里说明"本轮为并行研讨，产出请直接写在回复正文，不要写文件"，降低模型困惑。
|
||||
- **窄屏**：主区两列降为一列 + 头像条横向滚动。
|
||||
- **单 stage 全并行（一次性 N 个）**：给并行度 warning（远程模型并发限流 / 成本）。
|
||||
|
||||
---
|
||||
|
||||
## 10. 与现有代码的衔接点速查
|
||||
|
||||
- 编排循环：`hooks/useStep2Orchestration.ts`（`runOrchestration` / `runChainOrchestration` 旁新增 `runDagOrchestration`；复用 `isStale()` epoch、`abortAllInFlight`、`appendStreamingDialogue` / `appendToDialogue`）。
|
||||
- 流式 API：`api/multi-agent.ts`（`streamMultiAgent`，新增 `parallel_no_file` 入参；`onTextDelta(text, messageId)` 已是按席位回调，天然支持并行分流）。
|
||||
- 显示：`components/Step2Panel.tsx`（消息列表 `step2RoundtableDialogues.map`；新增头像条 `ParallelSeatStrip` + 两列容器 + 折叠左侧）。
|
||||
- 流程图：`components/DispatchChainModal.tsx` / `lib/job-types.ts` / `lib/business-chain.ts`（复刻成 DAG 版）。
|
||||
- 常量 / 元数据：`lib/roundtable-constants.ts`（`BUILTIN_AGENT_META` 头像/中文名、`buildAgentNameToRole`）。
|
||||
- 后端 run policy：`app/gateway/roundtable_run_policy.py`（新增 `seat_parallel`）、`app/gateway/routers/multi_agent.py`（`_special_run` 读 `no_file` flag）。
|
||||
- 交付摘要：`GET /api/threads/{id}/last-ai-message?max_chars=N`（依赖注入 stage N 输入用）。
|
||||
|
||||
---
|
||||
|
||||
## 11. 一句话总结
|
||||
|
||||
把「线性链条」升级为「分层 DAG」：**编辑器**用分层 stage（拖卡片）表达串/并；**调度**用 `Promise.all` 批内并行、批间 barrier 串行，依赖摘要在 stage 边界注入；**并行禁写文件**避免沙箱打架；**显示**收起左侧、头像条切换、主区两列、未选中只接流不渲染（rAF 节流）保证不卡。全部建立在现有 special/leader run + 轻量交付接口之上，后端几乎零新接口。
|
||||
|
||||
---
|
||||
|
||||
## 12. 已知问题与排查记录（2026-06-06）
|
||||
|
||||
### 12.1 ❗并行席位「步骤条」不显示 —— 根因：`seat_parallel` 工具面禁得过多
|
||||
|
||||
**现象**：并行两列里的席位气泡看不到步骤条（`MessageStepsCard`），而串行席位有。
|
||||
|
||||
**排查结论（已定位，待修）**：
|
||||
|
||||
1. **步骤条只渲染「工具调用」步骤**。`MessageStepsCard` 的 `isToolStep` 只认 `tool_calling:<toolName>` 的步骤进折叠卡；纯阶段步骤（`connecting:` / `streaming:` 这类 reasoning）只渲染成**底部一行小灯泡**，不出大卡（见 `MessageStepsCard.tsx` L80-100）。
|
||||
2. **并行席位几乎调不动任何会产生步骤的工具**。后端 `app/gateway/roundtable_run_policy.py`：
|
||||
- `_SEAT_EXCLUDED = [web_search, ask_clarification, present_files, view_image, agent_orchestration]` —— **注意 `web_search` 对所有席位（串行也）就禁了**（注释「圆桌研讨里联网搜索一律禁用」）。
|
||||
- `_SEAT_PARALLEL_EXCLUDED = _SEAT_EXCLUDED + [write_file, bash, str_replace]`。
|
||||
- 于是并行席位可用工具几乎只剩 `read_file / ls / task / write_todos / memory / hindsight_*`，研究/分析类席位基本一个都不会调 → **没有任何 `tool_calling` 步骤** → 步骤条（大卡）不出现。
|
||||
3. **对比串行为何「有步骤条」**：串行席位（`seat` policy）**能 `write_file`**，会产生 `tool_calling:write_file` 步骤 → 大卡出现。并行被禁了 write_file + 又没有 search，所以一个工具步都没有。
|
||||
|
||||
**用户诉求**：并行时**只禁真正会冲突的写操作**，搜索信息 / 调用技能等**必要操作不要禁**。
|
||||
|
||||
**关键事实（影响方案选择）**：圆桌每个席位是**独立 thread = 独立 sandbox**（后端 per-thread 隔离 `.deer-flow/users/{uid}/threads/{tid}/`）。所以并行席位写文件**在后端并不会互相冲突**；当初 §4「并行禁写文件」的「打架」其实主要是**前端单沙箱预览**（`activeArtifactThreadId` 只能同时显示一个）+ 产物归集到正文的简化考量，并非后端真冲突。→ 这意味着「并行禁写」的必要性比原设想弱，可放宽。
|
||||
|
||||
**✅ 决策：选 A（2026-06-06 已实施）**：把 `seat_parallel` 放宽到**只禁 `write_file / bash / str_replace`**（防前端单沙箱预览串台 + 产物走正文）+ 仍禁 `ask_clarification / present_files / view_image / agent_orchestration`（打断/越权类），但**放开 `web_search` 与技能**（搜索不碰 sandbox；技能靠 `read_file`）。改在 `app/gateway/roundtable_run_policy.py` 的 `_SEAT_PARALLEL_EXCLUDED`（**不再继承** seat 的 web_search 禁用，显式列举），测试 `tests/test_roundtable_run_policy.py` 同步更新（`test_seat_parallel_allows_web_search_and_skills`）。
|
||||
→ 并行席位现在能联网搜索/用技能 → 产生 `tool_calling` 步骤 → 步骤条显示。
|
||||
|
||||
> 备选 B（连写文件也放开，沙箱按 thread 隔离）/ C（重新评估「圆桌串行也一律禁 web_search」）**本期不做**，留档。串行席位目前仍禁搜索（`seat` policy 不变）。
|
||||
|
||||
### 12.2 其它已修（本轮反馈）
|
||||
|
||||
- 最左侧**全局工作区侧栏**并行时收起：改用与「打开沙箱」同一路径 `useSidebarSafe()?.setOpen(false)`（之前只收了圆桌「参与角色」栏 `step2SidebarExpanded`）。并行结束**不自动展开**（与沙箱收起不回弹一致）。
|
||||
- 并行两列改回**完整 Markdown**（撤销 §6.2 流式纯文本降级 `plainWhileStreaming`），与串行同款 `Step2Message`。
|
||||
- 头像条**不再显示字数**（去掉 charCount 徽标 + 字段）。
|
||||
- **致命 bug 已修**：用户干预 / 并行跑完后**不再重跑整张 DAG**。`dagPlanExecutedRef` 闸门 + DAG 各阶段跑完后释放锁 `await runOrchestration(handoff)` 交回总控继续派活/收口；后续干预/恢复一律走 leader 自由派活。
|
||||
- **并行两列宽表溢出已修（2026-06-07）**：模型输出的 Markdown **表格**（如「法律免责条款」表）没有宽度约束 / 横向滚动容器，在并行两列的窄列里会**超出左列、盖到相邻列**。修法：`MessageBubble.sharedMarkdownComponents.table`（+ `markdown-components.tsx`）把 `<table>` 套进 `<div class="max-w-full overflow-x-auto">`——表格保持自然宽度、超出时**列内横向滚动**而非撑破列宽。`min-w-0` 链(grid `minmax(0,1fr)` → 列 `min-w-0` → 气泡 `flex-1 min-w-0`)本就齐全,只差 table 这一层。
|
||||
- **关于「并行席位无步骤条」**：步骤条(`MessageStepsCard` 大卡)只在席位产生 `tool_calling` 步骤时出现(§12.1)。纯分析型席位若本轮**直接写分析正文、没调用任何工具**(web_search 等),就只显示一行「正在输出」思考指示、不出大卡——这是设计使然,非解析失败。要确认是否真调了工具:浏览器 devtools 设 `window.__MULTI_AGENT_DEBUG__=true` 看 `frame# event=messages` 里有无 tool_calls,或查后端 special run 日志。
|
||||
|
||||
### 12.3 ✅需求变更：Stage 间「经总控中转 + 广播」（2026-06-06 提出，2026-06-07 已实施）
|
||||
|
||||
**当前实现（方案 B，§3.2）**：DAG 各 stage 由前端**直接**派活，stage 间不跑 leader，靠前端把 `buildDependencySummary` 拼进下一 stage 的 task（`priorSummary`）。各 stage 跑完后才 `await runOrchestration` 交回总控收口。
|
||||
|
||||
**新诉求**：业务链条调用时，**每一轮 stage 之间都要经总控中转 + 广播**，而不是前端直接串下去。具体：
|
||||
|
||||
1. 第 1 轮并行调用 stage1 的 A、B、C；
|
||||
2. A/B/C **全部完成后**，把这一批的**输出广播出来**（广播给其它席位 + 总控 thread）；
|
||||
3. **总控（leader）接收**这批交付 → 再**往下继续派活**（派 stage2 的 D、E）；
|
||||
4. 派活时把（承上的）消息**广播给下一轮的 D、E**；
|
||||
5. 以此类推，直到所有 stage 跑完再收口。
|
||||
|
||||
> 即把 §3.2 的「方案 B 前端直派」改成 **「方案 A' = 计划约束下的总控中转」**：stage 边界插入一轮 leader，由它综合上一批交付并驱动下一批（下一批席位仍由编排计划**钉死**为 D、E，不是 leader 自由乱派——leader 负责「承上启下 + 广播上下文」，计划负责「派给谁」）。
|
||||
|
||||
**实施要点（下次会话）**：
|
||||
|
||||
- **每个 stage 完成 → 广播该批输出**：后端 `_special_run` 已对每个 seat 做 detached 后广播（「子智能体X完成…交付内容为:…」→ 其它席位 + 总控）。但它是 **detached/best-effort + 截断**；要让总控**可靠**地在下一轮前看到，需在 stage barrier 处**等广播落地**（或前端在 stage 边界用 `GET /threads/{id}/last-ai-message` 主动收齐各席位交付再喂给 leader）。
|
||||
- **插入 leader 中转轮**：`runDagOrchestration` 在进入下一 stage 前，跑一轮 leader（`synthesisMode=true` 让后端注入上一批**完整**交付），让总控产出一段「承上启下」说明；把这段说明（而非机械 `priorSummary`）作为下一 stage 席位 task 的上下文 / 或广播给下一批席位 thread（复用 `POST /threads/{id}/messages/append`）。
|
||||
- **下一批仍由计划钉死**：leader 中转轮**不**自由决定派给谁（计划已定 D、E）；可参考 `buildChainLeaderTask` 的「约束 leader 只派给指定席位」思路，或干脆 leader 只做综合、前端照计划派 D、E 并把 leader 综合 + 广播注入其 task。
|
||||
- **与「并行禁写」「§12.1 放开搜索」并存**：中转/广播不影响工具面。
|
||||
- **与致命 bug 修复并存**：最后一个 stage 跑完仍走「交回总控（`runOrchestration`）收口/继续派活」；用户干预仍走 leader、不重跑 DAG。
|
||||
|
||||
**影响文件（预估）**：`hooks/useStep2Orchestration.ts`（`runDagOrchestration` 的 stage 循环里加 leader 中转 + 广播收齐）、可能复用后端 `_leader_run`（synthesisMode 注入全文）/ `/messages/append`（广播）/ `/last-ai-message`（收齐交付）。后端基本无需新接口。
|
||||
|
||||
**✅ 实施结论（2026-06-07）**：后端**零改动**，全部在前端 `runDagOrchestration` 内完成，复用既有「传摘要按需传全文」优化链路：
|
||||
|
||||
1. **中转轮位置**：每个**非最后**的 stage barrier 到达后（`Promise.all` 收齐本批交付），插入**一轮 leader**「中转轮」，展示为「总控协调（承上启下）」气泡。最后一个 stage 后**不**插中转，综合/收口仍交给末尾的 `runOrchestration` 接管（与致命 bug 修复一致）。
|
||||
2. **可靠收齐交付（传摘要）**：中转轮跑 `streamMultiAgent({agentType:"leader", synthesisMode:false})`。后端 `_leader_run` 本就在每轮 leader 前用 `_collect_delivery_status`（走**轻量 `GET /last-ai-message?max_chars=600` 接口**）**客观核对**各席位真实交付并把**摘要**注入总控输入（`_build_seat_status_block`）——比 detached 广播可靠（直接实读各席位 thread，不依赖 best-effort 广播是否落地）。故中转轮传 `synthesisMode:false`：只综合本批、**不收口**，只注入摘要；**完整交付正文**留到末尾收口的 synthesis 轮（`_build_full_deliverables_block`，全 ✅ 才注入）按需读取——即「**传摘要按需传全文**」。
|
||||
3. **承上启下驱动下一批**：中转轮产出的综合说明存进 `leaderBridge`，作为**下一 stage** 席位 task 的上下文（`buildStageTask({priorSummary: leaderBridge, priorIsBridge:true})`，措辞为「承上启下综合说明」而非机械「交付摘要」），替代原来的 `buildDependencySummary(deliveries)`。中转失败/无产出时回退机械依赖摘要，不阻断。
|
||||
4. **下一批由计划钉死**：中转轮提示（`buildDagTransitTask`）明确告知总控「下一阶段席位已由编排计划钉死为 X、Y，你**不要**自行 `agent_orchestration` 派活、**不要**收口」，前端仍照计划直派下一批；即便弱模型仍派了活，前端忽略其 `dispatched`、只取 `content` 当 bridge。
|
||||
5. **广播**：本批输出广播给其它席位 + 总控 = 后端 `_special_run` 既有 detached 后广播（截断摘要）；承上消息给下一批 = 注入其 task `new_message`（席位 run 输入必见，比 `/messages/append` 更可靠，省一次写盘）。
|
||||
6. **澄清/抢占/失败**：中转轮触发澄清 → 暂停 DAG（回复后走 leader 自由派活、不重跑 DAG，沿用 `dagPlanExecutedRef` 闸门）；被 `isStale()`/abort 抢占 → 静默收尾；异常 → 该气泡标错误、清空 bridge 继续。
|
||||
|
||||
**纯函数**：`lib/dag-types.ts` 新增 `buildDagTransitTask`（中转轮提示）+ `buildStageTask` 加 `priorIsBridge` 形参；`lib/dag-types.test.ts` 加 4 个单测（共 27 个全过）。
|
||||
|
||||
> ⚠️ **§12.3 的「单独中转轮 + 前端直派」已被 §12.4 取代**（2026-06-07 同日）：见下。`buildDagTransitTask` / `buildStageTask` / `buildDependencySummary` 仍保留在 lib（带单测），但 `runDagOrchestration` 不再调用它们。
|
||||
|
||||
### 12.4 ✅ 需求再变更：DAG 改成「总控驱动每 stage 派活」（2026-06-07，§3.2 方案 A）
|
||||
|
||||
**新诉求（用户原话）**：把并行(DAG)逻辑改得和串行一样——串行时总控给多个智能体派活、它们完成后汇报给总控、总控再继续往下派活；并行也改成「由总控判断要执行哪些步骤，这些步骤完成后再汇报给总控，总控继续往下派活」。
|
||||
|
||||
**已确认的两个关键决策**：
|
||||
1. **每批「派给谁」**：✅ **计划钉死成员，但走总控派活**（不是总控自由乱派）。即 stage 成员仍由编排计划定死，但改成总控每轮**真正** `agent_orchestration` 派活给这些席位（约束只派给它们、且每个都派到）。= §3.2「方案 A」+ 批内并行。
|
||||
2. **适用范围**：✅ **只改 dag 模式**（recommend 仍串行逐个、chain 不动）。
|
||||
|
||||
**实现（后端零改动，全在 `runDagOrchestration`）**：每个 stage 两步——
|
||||
- **(a) 总控派活轮**：跑一轮 `leader`，提示 `buildDagStageLeaderTask`（首 stage 带任务背景；后续 stage 靠后端注入的此前交付摘要承上；列出本 stage 钉死席位的 `agent_name`，要求逐个派到、不派清单外、不收口）。`synthesisMode:false` → 后端 `_build_seat_status_block` 走轻量 `last-ai-message` 注入各席位**摘要**（传摘要按需传全文）。
|
||||
- **(b) 批内并行**：本 stage 席位并行跑总控派给各自的 task——`pickDagSeatTask(leaderRes.dispatched, content, agentId, name)` 命中其 `agent_name`；总控漏派/写偏则回退总控正文/模板，**不借用别席位 task**（多席位并行避免张冠李戴）。完成后交付经后端 `_special_run` detached 后广播回总控；下一 stage 的派活轮即可看到。
|
||||
- barrier 后进入下一 stage，总控再派下一批。全部 stage 跑完 → 交回 `runOrchestration` 收口（synthesisMode 按需注入全文）。
|
||||
|
||||
**相比 §12.3 的变化**：§12.3 是「前端直派 + stage 间单独插一轮只综合的中转轮」；§12.4 把**每个 stage 本身**就改成「总控派活轮 + 批内并行」——总控既承上（注入的摘要）又派活，不再需要单独的中转轮。与 chain 模式同构，区别是一次钉死一批、批内并行。
|
||||
|
||||
**纯函数**：`lib/dag-types.ts` 新增 `buildDagStageLeaderTask`（每 stage 派活提示）+ `pickDagSeatTask`（取某席位 task，多席位不张冠李戴）；`lib/dag-types.test.ts` 加 5 个单测（共 32 个全过）。`runDagOrchestration` 不再用 `buildStageTask/buildDependencySummary/buildDagTransitTask`（保留在 lib）。
|
||||
|
||||
**Phase 7 流程图兼容**：总控派活轮无席位 active + `isOrchestrating` → 流程图「总控」节点显示 active（承上派活中），随后该 stage 列点亮，无缝衔接。
|
||||
|
||||
**下次实地验收**：①每个 stage 前出现「总控协调（派活中/承上派活中）」气泡,内容是总控对本 stage 席位的派活说明;②后端日志该轮 leader 的 `_partition_dispatch` 派给的正是本 stage 钉死席位;③这些席位并行跑(special 日志时间重叠);④完成后下一 stage 总控派活轮能看到上一批交付摘要(`[leader-roster]` 日志)。
|
||||
|
||||
### 12.5 ✅ 流程图弹窗化 + DAG「快速/沙箱」执行子模式（2026-06-07）
|
||||
|
||||
三个改动一起做:
|
||||
|
||||
**(1) 流程图改成弹窗**:Step2 顶部不再内联 `DagFlowChart`(挤占两列消息高度,截图 `样式优化-1.png`)。改成一条细栏「分层编排执行中 · 共 N 个阶段 ＋ [查看编排流程图]」按钮,点开 `Dialog` 看**全尺寸**流程图(非 compact)。编辑器(`BusinessChainEditorPage`)底部预览保持内联不变。改 `Step2Panel.tsx`(加 `dagFlowOpen` state + Dialog)。
|
||||
|
||||
**(2) 推荐弹窗新增 DAG 执行子模式(快速/沙箱)**:`BusinessChainPicker` 在选中链条**有分层并行**(走 dag)时,「进入研讨」旁出现「执行方式:快速·双列禁写 / 沙箱·单列可写」二选一(只 dag 时出现,§已确认)。
|
||||
- **快速(`fast`,默认)**:同 stage 席位**并行**、主区**双列**、`parallelNoFile=true`(seat_parallel **禁写文件**)。= 之前的行为。
|
||||
- **沙箱(`sandbox`)**:同 stage 席位**仍并行**,但主区**单列**(`selectedParallelSeats` 最多 1)、`parallelNoFile=false`(seat policy **允许写文件**);点头像切换查看,沙箱跟随当前席位。
|
||||
- 数据流:`SeatSelectionResult.dagExecMode` → `RoundtablePlanningPage` state(持久化进 `DraftStep2(Run)Snapshot.dagExecMode`,mirror orchestrationPlan)→ `useStep2Orchestration` option → `runDagOrchestration`。
|
||||
|
||||
**(3) 沙箱跟随当前显示席位(防 bug 重点,用户特别强调)**:`useStep2Orchestration`
|
||||
- `sandboxFollowThreadIdRef` = 当前显示席位的 thread;`openSeatVirtualArtifact` 的 auto 守卫:sandbox 模式下,**后台并行席位**写文件时若不是当前显示席位,**不自动打开**(否则沙箱内容和主区列表错位)。
|
||||
- 切换显示席位的 effect(keyed on `selectedParallelSeats` + `dagExecMode`):该席位**有产物**(stub 里有 write_file/str_replace)→ `openSeatLatestArtifact` 强制打开它最新的文件;**无产物**→ 关闭沙箱面板(`setActiveArtifactThreadId(null)` + `deselectArtifact` + `setArtifactsOpen(false)`)。
|
||||
- 当前显示席位 live 写文件 → `handleArtifactToolPhase` auto-open,守卫放行(follow===该席位)。后台席位写 → 守卫拦掉、只缓存 stub;切到它时 effect 再打开。
|
||||
- `toggleParallelSeat`:sandbox 单列 = 点头像切换(始终保留 1 个,不取消到空);fast 双列 = 原 LRU 2 个。
|
||||
|
||||
**改动文件**:`lib/business-chain.ts`(SeatSelectionResult.dagExecMode)、`components/BusinessChainPicker.tsx`(子模式 UI)、`api/drafts.ts`(快照 dagExecMode)、`pages/RoundtablePlanningPage.tsx`(state/snapshot/hydrate/confirm/reset 全链)、`hooks/useStep2Orchestration.ts`(option + parallelNoFile/单列 + 沙箱跟随 + toggle cap)、`components/Step2Panel.tsx`(流程图弹窗 + 单列 grid)。typecheck 零新增(基线 65),dag-types 32 单测全过。
|
||||
|
||||
**下次实地验收(沙箱跟随是重点)**:①推荐弹窗选有并行的链条 → 出现「快速/沙箱」二选一;②快速=双列禁写(无沙箱);③沙箱=单列、席位能写文件、点头像切换列表时沙箱内容**同步**切换:有产物显示、无产物关闭;④后台席位写文件不会把沙箱抢到非当前列表;⑤流程图点按钮弹窗、全尺寸、节点随执行点亮。
|
||||
|
||||
### 12.6 ✅ 修:并行席位「跟着派活」（2026-06-07）
|
||||
|
||||
**现象**(截图 `bug-1.png`):总控派活后,并行席位的气泡里显示的是**总控的整段派活叙述**(「我来调度各智能体…第1阶段任务清单:席位1…席位2…」),还去调 `skill_list` 找 `agent_orchestration`——席位在**模仿派活**。
|
||||
|
||||
**根因**:总控这轮**没真正调用 `agent_orchestration` 工具**(弱模型 deepseek 用文字罗列了派活计划)→ 后端解析的 `dispatched=[]` → 前端 `pickDagSeatTask` 回退把**总控整段叙述**喂给**每个**席位 → 席位照抄叙述、模仿派活。(席位本就被 run policy 禁了 `agent_orchestration`,**派不了活**,只是空转一轮叙述 + 调 `skill_list`。)
|
||||
|
||||
**修 v1(护栏,效果不足)**:`pickDagSeatTask` 加角色边界护栏 + 回退措辞改「挑出属于你的那份」;`buildDagStageLeaderTask` 加「必须真正调用工具」。→ **没拦住**:总控对 deepseek 仍只写文字不调工具(`dispatched` 还是空),席位拿到的**任务正文本身**就是那段「我来调度…席位1/2/3」,护栏被淹没;3/4 席位仍模仿派活,有的甚至绕道 `bash`+`python requests.get('/api/agents')` 调内部接口派活(bug-1~4)。
|
||||
|
||||
**✅ 修 v2(根治:回退绝不注入派活叙述)**:`pickDagSeatTask` 去掉 `leaderContent` 入参——
|
||||
- **命中**总控结构化派活(真调了工具)→ `请你作为「X」，完成总控派给你的本阶段任务：<dispatched task>` + 护栏;
|
||||
- **未命中 / `dispatched` 空**(总控只写文字)→ **绝不**把总控派活叙述喂给席位,改给**干净本职任务**:「请你作为「X」,从你的专业职责出发,独立完成本阶段属于你的分析与交付。任务背景与此前各阶段交付已在你的会话上下文里(后端广播注入),据此承接」+ 护栏。护栏强化到「不要用 bash/python/脚本访问内部接口(/api/agents、agent_orchestration)派活、不要复述『我来调度』总控口吻」。
|
||||
- `buildDagStageLeaderTask` 的「必须真正调用工具」保留(让命中率更高、尽量走 exact-match)。
|
||||
- `dag-types.test.ts` 改 2 单测(共 33 过)。typecheck 65。
|
||||
|
||||
> ⚠️ **残留次要污染源(若仍复发再处理)**:DAG 每个 stage 的 leader 轮,后端 `_leader_run` 会把 leader 的输入(含「请你作为总控…派活」)**预广播**进所有席位线程(`总控智能体收到用户消息:…`)。席位的**当轮任务**已干净(修 v2),这条只是历史上下文,理论上被「最新的干净任务 + 护栏」压制。若实测仍模仿派活,下一步:后端给 DAG leader 轮加 `suppress_pre_broadcast` / 或广播一条**净化**过的背景(去掉派活措辞)。
|
||||
|
||||
### 12.7 ✅ 修 4 项(2026-06-07):席位仍派活(v3 后端)+ 沙箱关不掉 + 并行不自动下滑 + 流程图只到 Step2
|
||||
|
||||
实测 §12.6 仍有席位派活(视觉设计师 `skill_manage` 找 agent-orchestration、列 dispatch all 4)——**坐实**了「残留次要污染源」:后端 leader 预广播把派活 prompt 推进席位线程,弱模型从历史里捡到派活意图。一起修了 4 项:
|
||||
|
||||
1. **席位仍派活(根治·后端)**:`_leader_run` 加 `suppress_pre_broadcast` 参数;DAG「每 stage 派活轮」前端传 `suppressPreBroadcast: true`(`api/multi-agent.ts` → `suppress_pre_broadcast`,run/stream 读取)。为真时:**不**把总控派活 prompt 预广播给席位、**也不**把「总控给 X 派活」detached 广播给兄弟席位。recommend/chain 不传该 flag,行为不变。配套:`pickDagSeatTask` 加 `seedMessage` 入参——预广播抑制后席位线程没有用户原始消息了,把**任务背景**直接写进席位 task(此前各阶段交付仍由 special 完成后广播注入)。→ 席位线程里**任何地方都没有派活语句**,根治模仿。
|
||||
2. **沙箱关闭按钮关不掉**:沙箱跟随 effect 之前每次 render(deps 身份变)都重跑 → 把用户刚关的沙箱强制重开。加 `lastReconciledSeatRef` 守卫:**只在显示席位真正变化时**才开/关沙箱;用户手动关闭(席位没变)时 effect 早退,不再回弹。
|
||||
3. **并行时自动下滑失效**:平铺列表 ↔ 网格(两列/单列)切换时 `scrollHeight` 突变、`scrollTop` 被浏览器 clamp → `useAutoScroll` 误判「用户上滚」而停跟随。`Step2Panel` 加 effect:`isParallelRunning` / `selectedParallelSeats` 变化时 `scrollStep2ToBottom()` 重新拉到底 + 恢复跟随。
|
||||
4. **流程图只展示 Step2**:`DagFlowChart` 加 `showReportMilestone`(默认 true);Step2 弹窗 + 编辑器预览传 `false` —— 只到「共识」(Step2 终点),去掉「结果绘制」(那是 Step3)。
|
||||
|
||||
**改动**:后端 `multi_agent.py`(`_leader_run` suppress 两处广播 + run/stream 读 flag);前端 `api/multi-agent.ts`、`lib/dag-types.ts`(pickDagSeatTask + seedMessage)、`hooks/useStep2Orchestration.ts`(传 flag + seed + 沙箱跟随守卫)、`components/Step2Panel.tsx`(自动下滑 effect + showReportMilestone)、`components/DagFlowChart.tsx`、`pages/BusinessChainEditorPage.tsx`。`dag-types.test.ts` 共 **34 单测过**;前端 tsc 65 基线零新增;后端 `py_compile` OK。
|
||||
|
||||
### 12.8 ✅ 席位产出 md 文件 + 禁写 .py（2026-06-07）
|
||||
|
||||
**现象**:某次 7 席位**全都没产出 md 文件**;且一直有的问题——席位**有时写 .py 脚本**(用户要求:不写 .py,其它格式可以)。
|
||||
|
||||
**根因(没 md)**:两层都堵了——(a)**快速模式** `seat_parallel` policy 禁了 `write_file`(原「产物走正文」设计);(b)DAG 席位**提示词**写的是「直接输出你的交付正文」,即使沙箱模式(允许写)也不会去写文件。
|
||||
|
||||
**修**:
|
||||
1. **后端 policy 放开写文件、统一禁 bash**(`roundtable_run_policy.py`):
|
||||
- `seat_parallel`:去掉 `write_file` / `str_replace` 禁用(**可写 md**),保留禁 `bash` + 打断/越权类;仍放开 `web_search`。
|
||||
- `seat`:新增禁 `bash`(原来允许)——圆桌席位是「文档产出者」,不该跑 shell/python 脚本(弱模型爱写 .py 再跑生成产物,乱且抓不到预览);改为**直接 write_file 写成品文档**。
|
||||
- `report`:`bash` 已随 `seat` 禁,改为 `[*_SEAT_EXCLUDED, "str_replace"]`(集合不变)。
|
||||
- 测试 `test_roundtable_run_policy.py`:`test_seats_can_write_files_but_not_run_bash`(seat / seat_parallel 都可写文件、都禁 bash),6 用例过。
|
||||
2. **前端提示词**(`pickDagSeatTask`)加「产物要求」:**把完整交付写成一个 .md 文件**(write_file 到 outputs)+ 正文给摘要;**禁写 .py 或任何脚本文件**,.md/.html/.txt 等其它格式都可以,唯独不要 .py。
|
||||
3. **快速模式(双列)不自动开沙箱**(`openSeatVirtualArtifact` 新增守卫):放开写文件后,双列里多席位并发写文件会让单沙箱预览反复串台——fast 模式 DAG 并行批次**完全不自动开沙箱**(文件仍写盘,Step3/文件卡可查看);沙箱模式(单列)仍按「沙箱跟随当前席位」预览。
|
||||
|
||||
**关于 .py 的强度**:`write_file` 是全局工具(所有 agent 共用),无法在网关层按扩展名硬拦,故 .py 禁止走**提示词**(强约束)+ **禁 bash**(写了也跑不了,脚本生成产物的动机被掐断)。若实测仍偶发写 .py 且需**硬拦**,下一步:给 lead_agent 的 `write_file` 工具加一个「按 run context 禁某些扩展名」的开关,网关 `_run_payload` 对圆桌席位置 `[".py"]`(只影响圆桌 run,不动其它 agent)。
|
||||
|
||||
**改动**:`roundtable_run_policy.py`、`test_roundtable_run_policy.py`、`lib/dag-types.ts`、`lib/dag-types.test.ts`(35 过)、`hooks/useStep2Orchestration.ts`。前端 tsc 65、后端 py_compile OK + 6 policy 单测过。⚠️ **改了后端,需重启 Gateway**。
|
||||
|
||||
### 12.9 ✅ 席位「按需取前序全文」只读工具 read_peer_delivery（2026-06-07）
|
||||
|
||||
**诉求**:A,B,C → D,E 时,D,E 默认拿 A,B,C 的**摘要**,但识别到需要**全文**时也能取到。之前只有总控(leader)能「摘要默认+按需全文」(跨 thread 读),下游席位**不能**(席位间 thread 隔离、只有 1500 字截断广播)。
|
||||
|
||||
**实现**(真·按需,全文只在调用工具时才进 LLM 上下文):
|
||||
1. **新 harness 只读工具** `read_peer_delivery(seat_name)`(`packages/harness/deerflow/tools/builtins/roundtable_peers_tool.py`):从 `runtime.context["roundtable_peer_deliveries"]`(网关注入的 `[{name,content}]`)按席位名取全文;空名→列出可取席位;非圆桌场景→「无」。只读 context dict,**不跨 thread、不碰 sandbox、无网络**,纯 `deerflow.*`(不违反 harness→app 防火墙)。注册进 `tools/tools.py` 的 builtin(始终在,只读无副作用)。
|
||||
2. **网关注入 context**:`_run_payload` 加 `extra_context`;`_special_run` 收 `peer_deliveries` → 注入 `context.roundtable_peer_deliveries`;run/stream 读 `peer_deliveries`(规整 `[{name,content}]`)。**不进 prompt → 不调工具就不耗 token**。
|
||||
3. **run policy**:`read_peer_delivery` 从 `leader`(总控自己跨 thread 读,不需要)+ `report`(Step3 不读前序)排除;**席位(seat/seat_parallel)保留**。
|
||||
4. **前端**:`runDagOrchestration` 跨 stage 累积**完整交付** `priorDeliveries`,stage>0 席位的 special run 带 `peerDeliveries`(`api/multi-agent.ts` → `peer_deliveries`);`pickDagSeatTask` 加 `hasPeers` 形参,有前序时在 task 里提示「需要某前序席位全文时调用 read_peer_delivery('<名>'),平时用摘要」。
|
||||
|
||||
**最终语义**:D、E 默认看 A、B、C 的**摘要**(1500 字广播);识别到不够 → 调 `read_peer_delivery('情报收集')` 取**全文**。总控仍是摘要默认+综合轮全文。**= 用户要的「摘要默认、按需全文」对席位也成立了。**
|
||||
|
||||
**改动**:后端 `roundtable_peers_tool.py`(新)、`tools/tools.py`、`roundtable_run_policy.py`、`multi_agent.py`;前端 `api/multi-agent.ts`、`lib/dag-types.ts`、`lib/dag-types.test.ts`(36 过)、`hooks/useStep2Orchestration.ts`。前端 tsc 65;后端 py_compile OK + 6 policy 单测过 + 工具导入/行为 smoke 通过。⚠️ **改了后端 + 新增 harness 工具,需重启 Gateway**。
|
||||
|
||||
### 12.10 ✅ 修:沙箱模式第一个席位不自动开沙箱（2026-06-07，前端）
|
||||
|
||||
**现象**:沙箱(单列)模式并行时,**第一个**席位的消息列表不自动打开沙箱(要手动开);切换到其它席位时正常自动打开。
|
||||
|
||||
**根因**:沙箱跟随 effect(§12.5)在选中席位**还没写产物**时走「关闭沙箱」分支,里面同时调了 `deselectArtifact()` **和** `setArtifactsOpen(false)`。后者是 context 的 **wrapped setOpen**(`context.tsx:73`):`关闭时若 autoOpen 为真 → 把 autoOpen 置 false`。于是第一个席位刚选中(还没写文件)→ autoOpen 被关 → 该席位**随后 live 写文件**时,`openSeatVirtualArtifact` 的 `!flags.autoOpen` 守卫拦掉自动打开。切换到**已写过**文件的席位能开,是因为那条路走 `openSeatLatestArtifact`(force,auto=false,绕过 autoOpen 守卫)。
|
||||
|
||||
**修**(一行):关闭分支只保留 `deselectArtifact()`(它走 raw `setOpen(false)`,**不**动 autoOpen),删掉多余的 `setArtifactsOpen(false)`。这样面板关掉、但 autoOpen 仍为真 → 第一个席位 live 写文件时正常自动开沙箱。(若用户**手动**关过沙箱,autoOpen 本就为 false,不会自动重开 —— 尊重用户。)改 `hooks/useStep2Orchestration.ts` 沙箱跟随 effect。tsc 65。
|
||||
|
||||
### 12.11 ✅ 修:快速模式又生成文件了 + Fragment 警告（2026-06-07）；记录无关 404
|
||||
|
||||
1. **快速模式不应生成文件(回退 §12.8 的过度放开)**:§12.8 为「让席位产 md」把 `seat_parallel`(快速模式)也放开了 write_file —— 但**快速模式(双列)的设计就是产物走正文、不写文件**,写文件归**沙箱模式**(单列 + 沙箱预览)。修:
|
||||
- `roundtable_run_policy.py`:`seat_parallel` 重新**禁 write_file / str_replace**(保留禁 bash);`seat`(沙箱模式)仍可写。
|
||||
- `useStep2Orchestration.ts`:`parallelNoFile = dagExecMode !== "sandbox"`(快速模式一律禁写,含单节点 stage;沙箱模式放开)。
|
||||
- `pickDagSeatTask` 加 `allowFiles` 形参:沙箱(true)=写 .md 文件 + 禁 .py;快速(false)=产物走正文、不写文件。hook 按 `dagExecMode === "sandbox"` 传。
|
||||
- 测试:`test_sandbox_seat_writes_files_fast_parallel_does_not`(6 过);`dag-types.test.ts` 拆成沙箱/快速两个产物要求用例(37 过)。
|
||||
- **语义**:快速=双列、产物走正文、不写文件;沙箱=单列、写 .md、沙箱跟随预览。
|
||||
2. **`Invalid prop data-lov-id supplied to React.Fragment`**:`DagFlowChart` 的 stage 循环用了具名 `<Fragment key>`,dev 代码标注插件(lovable-tagger)给它注入 `data-lov-id`,而 Fragment 只接受 key/children → React 警告。修:换成 `<div className="contents" key=...>`(display:contents 不生成盒子,Connector/StageColumn 仍是外层 flex 直接子项,布局不变,且能接收 data-*)。移除 `Fragment` import。
|
||||
3. **(无关,非本期引入)`/api/getTokenByEmail?email=…%40local.deerflow` 404 ×60**:来自 `src/core/studio/api.ts` 的 `fetchStudioToken`(studio/expert-auth 引导),用户邮箱被映射到 legacy `@local.deerflow` 域、后端 `getTokenByEmail` 不认 → 404,React Query 跨多个 `useStudioToken` 调用点重试 → 重复多次。**与圆桌 DAG 改动无关**(圆桌头像是本地 SVG,本期未新增任何前端 GET)。需要的话单独排查 studio token 链路。
|
||||
|
||||
---
|
||||
|
||||
## 13. 实施进度（截至 2026-06-06，下次会话续）
|
||||
|
||||
| Phase | 内容 | 状态 |
|
||||
|---|---|---|
|
||||
| 1 | DAG 数据模型 + 校验 + 持久化字段（`lib/dag-types.ts`，24 单测） | ✅ 完成 |
|
||||
| 2 | 分层编辑器（**做进业务链条页**）：链条加 `stages` 字段（前后端 + alembic `20260607_03` + 11 后端单测）；`BusinessChainEditorPage` 竖列分组 Stage（DnD + 按钮）；纯逻辑 `lib/chain-stages.ts`（19 单测） | ✅ 完成 |
|
||||
| 3 | `runDagOrchestration`：批内 `Promise.all` 并行 + 批间 barrier 串行 + 依赖摘要注入（方案 B 前端直派） | ✅ 完成 |
|
||||
| 4 | 并行 run policy `seat_parallel`（`roundtable_run_policy.py` + `multi_agent.py` 读 `parallel_no_file`，6 后端单测）；**§12.1 方案 A 已放开 web_search/技能** | ✅ 完成 |
|
||||
| 5 | 并行显示：收起左侧 + 头像条 + 两列 + 选中切换 | ✅ 完成（本轮修了侧栏/Markdown/字数） |
|
||||
| 6 | 流式高性能：ref 全量累积 + 节流 flush（仅选中渲染） | ✅ 完成 |
|
||||
| 7 | DAG 流程图执行态可视化（`DispatchChainModal` 复刻成按 stage 分列；后改成**弹窗**见 §12.5） | ✅ 完成（2026-06-07） |
|
||||
|
||||
**端到端闭环已通**：链条页编排分层 → 存库 → 选链条进 Step2 → `runDagOrchestration` 并行直派 → 各阶段完交回总控派活/收口。
|
||||
|
||||
**✅ §12.3 已实施（2026-06-07）**：见 §12.3「实施结论」。`runDagOrchestration` 在每个非末尾 stage barrier 后插入 leader 中转轮（`synthesisMode:false` 走轻量 `last-ai-message` 注入摘要、收口轮才按需全文），承上启下驱动下一批；后端零改动；`dag-types.ts` 加 `buildDagTransitTask` + `buildStageTask.priorIsBridge`，27 单测全过。
|
||||
|
||||
**下次会话 TODO（优先级从高到低）**：
|
||||
1. **运行时实地验收 §12.3**：①每个 stage 之间出现「总控协调（承上启下）」气泡，内容是对上一批的综合；②后端日志中转轮 `[leader-roster] ... 注入席位交付状态` 显示上一批 ✅、未跑的 ⬜；③下一 stage 席位 task 里带的是总控承上说明（非机械摘要）；④中转轮不应真的派活（即使派了前端也忽略）。
|
||||
2. 运行时实地验收（旧）：①§12.1 后并行席位能搜索、步骤条显示；②并行 special 日志时间重叠 + `Excluded` 不含 web_search 含 write_file；③用户干预后是「总控派活」而非重跑 DAG；④最左全局侧栏 + 参与角色栏并行收起且不回弹。
|
||||
3. **✅ Phase 7 DAG 流程图可视化已完成（2026-06-07）**：见 §7.2。`DagFlowChart` 组件 + 编辑器只读预览 + Step2 执行态缩略(`dagNodeStates` 实时驱动)。
|
||||
4. （可选）串行也确定性注入上一轮交付摘要。
|
||||
|
||||
**关键文件速查**：
|
||||
- 前端：`hooks/useStep2Orchestration.ts`（`runDagOrchestration` / `dagPlanExecutedRef` / 并行 store+flush）、`components/Step2Panel.tsx`（两列+头像条）、`components/ParallelSeatStrip.tsx`、`lib/dag-types.ts`、`lib/chain-stages.ts`、`pages/BusinessChainEditorPage.tsx`、`pages/RoundtablePlanningPage.tsx`（侧栏收起 effect + handleConfirmAgents dag 分支）、`components/BusinessChainPicker.tsx`（链条→dag）。
|
||||
- 后端：`app/gateway/roundtable_run_policy.py`（`seat_parallel`，**待放宽**）、`app/gateway/routers/multi_agent.py`（`_special_run` 读 `parallel_no_file`）、`persistence/roundtable_chains/`（`stages` 列）、`migrations/versions/20260607_03_chain_stages.py`。
|
||||
- 测试：`frontend-web` 用 esbuild 转译 `.test.ts` 后 `node --test`（项目无 vitest）；后端 `PYTHONPATH=. uv run pytest tests/test_roundtable_chain_stages.py tests/test_roundtable_run_policy.py`。
|
||||
|
||||
@@ -1,375 +0,0 @@
|
||||
# 浅蓝色 UI 规范
|
||||
|
||||
> 本文档由设计稿截图（`1.1 全局样式-色彩`、`1.2 全局样式-色彩附录`）识别整理而成。
|
||||
> 全部色值已与"色彩附录"编号表逐一核对，保证准确。
|
||||
|
||||
---
|
||||
|
||||
## 1.1 色彩体系总览
|
||||
|
||||
调色体系分为 **主色** 与 **辅色** 两大类。每个色相按用途分为若干阶梯：
|
||||
|
||||
| 阶梯 | 数量 | 用途 |
|
||||
| --- | --- | --- |
|
||||
| 浅色 | 3 | 用于大面积底色 |
|
||||
| 中亮色 | 6 | 用于标签、图表、提示等 |
|
||||
| 深色 | 4（+基准色，共 5 个深阶） | 用于小面积按钮、图标等 |
|
||||
| 线性渐变色 | 1 | 用于特殊装饰性小面积底色 |
|
||||
| 径向渐变色 | 1 | 用于特殊装饰性小面积底色 |
|
||||
|
||||
> 中性色（灰阶）略有不同：浅色用于背景 / 白色文字及分割线，中亮色用于备注文字及图标，深色用于正文，且**无渐变色**。
|
||||
|
||||
---
|
||||
|
||||
## 主色
|
||||
|
||||
### 主色 - 红（基准 `#CF1322`）
|
||||
|
||||
| 用途 | 色值 |
|
||||
| --- | --- |
|
||||
| 浅色 | `#FEFAFA` · `#FEF6F6` · `#FEEDED` |
|
||||
| 中亮色 | `#FDDADC` · `#FCB5B9` · `#FA9196` · `#F86C73` · `#F74750` · `#F5222D` |
|
||||
| 深色 | `#CF1322` · `#BC0D1E` · `#A8071A` · `#820014` · `#5C0011` |
|
||||
| 线性渐变色 | `#F5222D` → `#A8071A` |
|
||||
| 径向渐变色 | `#F86C73` → `#F5222D` |
|
||||
|
||||
### 主色 - 蓝（基准 `#0062D9`）
|
||||
|
||||
| 用途 | 色值 |
|
||||
| --- | --- |
|
||||
| 浅色 | `#F9FCFF` · `#F4F9FF` · `#EAF4FF` |
|
||||
| 中亮色 | `#D4E9FF` · `#AAD2FF` · `#80BCFF` · `#55A6FE` · `#2B8FFE` · `#0079FE` |
|
||||
| 深色 | `#0062D9` · `#0056C6` · `#004AB3` · `#00368C` · `#002466` |
|
||||
| 线性渐变色 | `#0079FE` → `#004AB3` |
|
||||
| 径向渐变色 | `#55A6FE` → `#0079FE` |
|
||||
|
||||
---
|
||||
|
||||
## 辅色
|
||||
|
||||
### 辅色 - 橙（基准 `#D97B00`）
|
||||
|
||||
| 用途 | 色值 |
|
||||
| --- | --- |
|
||||
| 浅色 | `#FFFCF9` · `#FFFAF4` · `#FFF6EA` |
|
||||
| 中亮色 | `#FFEED4` · `#FFDDAA` · `#FFCC80` · `#FFBB55` · `#FFAA2B` · `#FF9900` |
|
||||
| 深色 | `#D97B00` · `#C66D00` · `#B35F00` · `#8C4600` · `#663000` |
|
||||
| 线性渐变色 | `#FF9900` → `#B35F00` |
|
||||
| 径向渐变色 | `#FFBB55` → `#FF9900` |
|
||||
|
||||
### 辅色 - 绿（基准 `#1EB07F`）
|
||||
|
||||
| 用途 | 色值 |
|
||||
| --- | --- |
|
||||
| 浅色 | `#FAFEFD` · `#F6FDFB` · `#EEFCF7` |
|
||||
| 中亮色 | `#DCF8EE` · `#BAF1DD` · `#97EBCD` · `#74E4BC` · `#52DDAB` · `#2FD69A` |
|
||||
| 深色 | `#1EB07F` · `#189D72` · `#118A65` · `#07634B` · `#043D30` |
|
||||
| 线性渐变色 | `#2FD69A` → `#118A65` |
|
||||
| 径向渐变色 | `#74E4BC` → `#2FD69A` |
|
||||
|
||||
### 辅色 - 青蓝（基准 `#2F9ACC`）
|
||||
|
||||
| 用途 | 色值 |
|
||||
| --- | --- |
|
||||
| 浅色 | `#FBFDFE` · `#F7FCFE` · `#EFFAFE` |
|
||||
| 中亮色 | `#E0F5FD` · `#C1EBFB` · `#A2E1F9` · `#83D7F7` · `#64CDF5` · `#45C3F3` |
|
||||
| 深色 | `#2F9ACC` · `#2788B9` · `#1E76A6` · `#115580` · `#0B3959` |
|
||||
| 线性渐变色 | `#45C3F3` → `#1E76A6` |
|
||||
| 径向渐变色 | `#83D7F7` → `#45C3F3` |
|
||||
|
||||
### 中性色 - 灰（基准 `#222222`）
|
||||
|
||||
| 用途 | 色值 |
|
||||
| --- | --- |
|
||||
| 浅色（背景、白色文字及分割线） | `#FFFFFF` · `#FAFAFA` · `#F5F5F5` |
|
||||
| 中亮色（备注文字及图标） | `#E8E8E8` · `#DBDBDB` · `#C1C1C1` · `#A6A6A6` · `#7F7F7F` · `#575757` |
|
||||
| 深色（正文） | `#3D3D3D` · `#222222` |
|
||||
|
||||
---
|
||||
|
||||
## 1.2 色彩附录（编号对照表）
|
||||
|
||||
编号规则：`ZS` = 主色（主色调），`FS` = 辅色 / 中性色。`-01~-03` 为浅色，`-04~-09` 为中亮色，`-10~-13` 为深色，`-14` 为线性渐变色，`-15` 为径向渐变色；不带后缀的为该色相基准深色。
|
||||
|
||||
### 主色 ZS
|
||||
|
||||
| 编号 | 红 (ZS-001) | 编号 | 蓝 (ZS-002) |
|
||||
| --- | --- | --- | --- |
|
||||
| ZS-001 | `#CF1322` | ZS-002 | `#0062D9` |
|
||||
| ZS-001-01 | `#FEFAFA` | ZS-002-01 | `#F9FCFF` |
|
||||
| ZS-001-02 | `#FEF6F6` | ZS-002-02 | `#F4F9FF` |
|
||||
| ZS-001-03 | `#FEEDED` | ZS-002-03 | `#EAF4FF` |
|
||||
| ZS-001-04 | `#FDDADC` | ZS-002-04 | `#D4E9FF` |
|
||||
| ZS-001-05 | `#FCB5B9` | ZS-002-05 | `#AAD2FF` |
|
||||
| ZS-001-06 | `#FA9196` | ZS-002-06 | `#80BCFF` |
|
||||
| ZS-001-07 | `#F86C73` | ZS-002-07 | `#55A6FE` |
|
||||
| ZS-001-08 | `#F74750` | ZS-002-08 | `#2B8FFE` |
|
||||
| ZS-001-09 | `#F5222D` | ZS-002-09 | `#0079FE` |
|
||||
| ZS-001-10 | `#BC0D1E` | ZS-002-10 | `#0056C6` |
|
||||
| ZS-001-11 | `#A8071A` | ZS-002-11 | `#004AB3` |
|
||||
| ZS-001-12 | `#820014` | ZS-002-12 | `#00368C` |
|
||||
| ZS-001-13 | `#5C0011` | ZS-002-13 | `#002466` |
|
||||
| ZS-001-14 | `#F5222D` → `#A8071A` | ZS-002-14 | `#0079FE` → `#004AB3` |
|
||||
| ZS-001-15 | `#F86C73` → `#F5222D` | ZS-002-15 | `#55A6FE` → `#0079FE` |
|
||||
|
||||
### 辅色 / 中性色 FS
|
||||
|
||||
| 编号 | 橙 (FS-001) | 绿 (FS-002) | 青蓝 (FS-003) | 中性灰 (FS-004) |
|
||||
| --- | --- | --- | --- | --- |
|
||||
| FS-00n | `#D97B00` | `#1EB07F` | `#2F9ACC` | `#222222` |
|
||||
| -01 | `#FFFCF9` | `#FAFEFD` | `#FBFDFE` | `#FFFFFF` |
|
||||
| -02 | `#FFFAF4` | `#F6FDFB` | `#F7FCFE` | `#FAFAFA` |
|
||||
| -03 | `#FFF6EA` | `#EEFCF7` | `#EFFAFE` | `#F5F5F5` |
|
||||
| -04 | `#FFEED4` | `#DCF8EE` | `#E0F5FD` | `#E8E8E8` |
|
||||
| -05 | `#FFDDAA` | `#BAF1DD` | `#C1EBFB` | `#DBDBDB` |
|
||||
| -06 | `#FFCC80` | `#97EBCD` | `#A2E1F9` | `#C1C1C1` |
|
||||
| -07 | `#FFBB55` | `#74E4BC` | `#83D7F7` | `#A6A6A6` |
|
||||
| -08 | `#FFAA2B` | `#52DDAB` | `#64CDF5` | `#7F7F7F` |
|
||||
| -09 | `#FF9900` | `#2FD69A` | `#45C3F3` | `#575757` |
|
||||
| -10 | `#C66D00` | `#189D72` | `#2788B9` | `#3D3D3D` |
|
||||
| -11 | `#B35F00` | `#118A65` | `#1E76A6` | — |
|
||||
| -12 | `#8C4600` | `#07634B` | `#115580` | — |
|
||||
| -13 | `#663000` | `#043D30` | `#0B3959` | — |
|
||||
| -14 | `#FF9900` → `#B35F00` | `#2FD69A` → `#118A65` | `#45C3F3` → `#1E76A6` | — |
|
||||
| -15 | `#FFBB55` → `#FF9900` | `#74E4BC` → `#2FD69A` | `#83D7F7` → `#45C3F3` | — |
|
||||
|
||||
> 注：中性灰（FS-004）无渐变阶，最深至 `#222222`（基准）/`#3D3D3D`（FS-004-10），用于正文文字。
|
||||
|
||||
---
|
||||
|
||||
# 组件规范
|
||||
|
||||
> 以下为各组件的视觉规范与实现方式。颜色统一引用上文色值，建议在项目中先抽取为 CSS 变量 / Tailwind token，组件只引用语义化变量。
|
||||
|
||||
## 语义化 Token 建议
|
||||
|
||||
把色彩体系映射成语义变量，组件层只用语义名，便于换肤与维护：
|
||||
|
||||
```css
|
||||
:root {
|
||||
/* 品牌 / 主色（蓝） */
|
||||
--color-primary: #0062D9; /* 主按钮、实底选中：ZS-002 深色基准 */
|
||||
--color-primary-hover: #0079FE; /* 悬浮态：ZS-002-09 */
|
||||
--color-primary-active: #0056C6; /* 按下态：ZS-002-10 */
|
||||
--color-primary-bg: #EAF4FF; /* 浅底选中底色：ZS-002-03 */
|
||||
--color-primary-border: #0079FE; /* 描边选中边框 */
|
||||
|
||||
/* 中性 */
|
||||
--color-text: #3D3D3D; /* 正文：FS-004-10 */
|
||||
--color-text-secondary: #A6A6A6; /* 次要 / 占位符：FS-004-07 */
|
||||
--color-border: #DBDBDB; /* 默认边框：FS-004-05 */
|
||||
--color-bg-disabled: #F5F5F5; /* 失效底色：FS-004-03 */
|
||||
--color-fill-disabled: #A6A6A6; /* 主按钮失效底色：FS-004-07 */
|
||||
--color-surface: #FFFFFF; /* 卡片 / 控件底 */
|
||||
}
|
||||
```
|
||||
|
||||
> 项目使用 Tailwind v4 + shadcn，可在 `@theme` 中声明这些变量后用 `bg-primary` / `text-primary` 等使用；下文示例同时给出等价的 Tailwind 写法。
|
||||
|
||||
---
|
||||
|
||||
## 2.1 基础 - 按钮（Button）
|
||||
|
||||
### 按钮尺寸
|
||||
|
||||
| 规格 | 高度 | 圆角 | 用途 |
|
||||
| --- | --- | --- | --- |
|
||||
| 小 | **28px** | 4px | 空间小的板块内部，如表格中、卡片标题等 |
|
||||
| 正常 | **32px** | 4px | 常用按钮（默认尺寸） |
|
||||
|
||||
水平内边距约 12–16px，字号 14px，图标与文字间距 4–8px。
|
||||
|
||||
### 常用按钮类型
|
||||
|
||||
| 类型 | 视觉 | 用途 |
|
||||
| --- | --- | --- |
|
||||
| 主按钮 | 实底 `--color-primary`，白字 | 一个操作区域**只能有一个**主按钮 |
|
||||
| 主按钮·失效 | 实底灰 `--color-fill-disabled`，白字 | 主按钮禁用态 |
|
||||
| 次要按钮 | 白底 + `--color-primary` 边框 + 蓝字 | 重要度低于主按钮，高于默认按钮 |
|
||||
| 默认按钮 | 白底 + `--color-border` 边框 + `--color-text` | 默认动作（取消等） |
|
||||
| 默认·失效 | `--color-bg-disabled` 底 + 浅边框 + `--color-text-secondary` | 次要/默认按钮禁用态 |
|
||||
| 文本按钮 | 无边框无底，蓝字（可加下划线） | 表格操作、文件下载等 |
|
||||
| 图标按钮 | 仅图标，可选 描边/实底/浅底 × 方形/圆形 | 增强辨识度或节省空间 |
|
||||
|
||||
实现（Tailwind 示意）：
|
||||
|
||||
```tsx
|
||||
// 主按钮
|
||||
<button className="h-8 px-4 rounded text-sm text-white bg-[#0062D9]
|
||||
hover:bg-[#0079FE] active:bg-[#0056C6]
|
||||
disabled:bg-[#A6A6A6] disabled:cursor-not-allowed">确定</button>
|
||||
|
||||
// 次要按钮（次按钮）
|
||||
<button className="h-8 px-4 rounded text-sm text-[#0062D9] bg-white
|
||||
border border-[#0062D9] hover:bg-[#EAF4FF]">点击</button>
|
||||
|
||||
// 默认按钮
|
||||
<button className="h-8 px-4 rounded text-sm text-[#3D3D3D] bg-white
|
||||
border border-[#DBDBDB] hover:border-[#0079FE] hover:text-[#0079FE]
|
||||
disabled:bg-[#F5F5F5] disabled:text-[#A6A6A6] disabled:border-[#DBDBDB]">取消</button>
|
||||
|
||||
// 文本按钮
|
||||
<button className="text-sm text-[#0062D9] hover:underline">文本按钮样式</button>
|
||||
|
||||
// 图标按钮（实底圆形）
|
||||
<button className="size-7 grid place-items-center rounded-full text-white bg-[#0062D9]
|
||||
hover:bg-[#0079FE]"><TrashIcon className="size-4" /></button>
|
||||
```
|
||||
|
||||
图标按钮三种底 × 两种形状：
|
||||
- **描边**：白底 + `--color-border` 边框，图标用 `--color-text`
|
||||
- **实底**：`--color-primary` 底 + 白图标
|
||||
- **浅底**：`--color-primary-bg`（`#EAF4FF`）底 + 蓝图标
|
||||
- 形状：方形 `rounded`（4px）/ 圆形 `rounded-full`
|
||||
|
||||
### 按钮组合
|
||||
|
||||
| 组合 | 说明 |
|
||||
| --- | --- |
|
||||
| 文字 + 下拉按钮 | 按钮右侧带 `⌄` 箭头；有 默认 / 主按钮 两种变体 |
|
||||
| 图标 + 文字按钮 | `+ 正常`，前置图标；有 描边 / 实底 两种变体 |
|
||||
| 文本按钮 + 图标 | 蓝色文本链按钮：`更多 »` `展开 ⌄` `下一页 »|` `收起 ⌃` `|⌜ 上一页` |
|
||||
| 主按钮 + 默认按钮 | `确定`(主) `取消`(默认) `返回`(默认) 并排，间距 8–12px |
|
||||
|
||||
### 其他状态
|
||||
|
||||
- **悬浮状态（hover）**：主按钮悬浮时底色变亮（`--color-primary-hover` `#0079FE`），用于 Banner / 登录 / 视觉冲击强的场景。
|
||||
- **加载中（loading）**：按钮内前置旋转图标 `⟳ 加载中`，期间 `disabled` 并降低交互；用于异步操作等待反馈。
|
||||
|
||||
```tsx
|
||||
<button className="h-8 px-4 rounded text-sm text-white bg-[#0062D9] inline-flex items-center gap-1.5" disabled>
|
||||
<Spinner className="size-4 animate-spin" /> 加载中
|
||||
</button>
|
||||
```
|
||||
|
||||
---
|
||||
|
||||
## 5.9 输入 - 单选框（选择按钮 / Segmented Radio）
|
||||
|
||||
一组互斥选项，外观为按钮组。两种排布 × 三种选中样式。
|
||||
|
||||
### 排布方式
|
||||
|
||||
| 排布 | 说明 |
|
||||
| --- | --- |
|
||||
| 拼接式 | 选项首尾相连，仅最外侧两角圆角，相邻共用 1px 边框 |
|
||||
| 分离式 | 选项之间留间距（约 8px），每个独立 4px 圆角，各自带完整边框 |
|
||||
|
||||
### 选中（已选）样式 × 三种
|
||||
|
||||
| 样式 | 已选 | 未选 |
|
||||
| --- | --- | --- |
|
||||
| 描边式 | 白底 + 蓝边框 `--color-primary` + 蓝字 | 白底 + `--color-border` 边框 + `--color-text` |
|
||||
| 浅底式 | `--color-primary-bg`(`#EAF4FF`) 底 + 蓝字 + 蓝边框 | 同上 |
|
||||
| 实底式 | `--color-primary` 实底 + 白字 | 同上 |
|
||||
|
||||
- **带统计**：选项文字后追加计数，如 `已选(230)` / `未选(230)`，计数用次要色 `--color-text-secondary`。
|
||||
|
||||
实现（拼接式 + 描边选中）：
|
||||
|
||||
```tsx
|
||||
<div className="inline-flex rounded overflow-hidden border border-[#DBDBDB] divide-x divide-[#DBDBDB]">
|
||||
{options.map((o) => (
|
||||
<button key={o.value}
|
||||
className={cn(
|
||||
"h-8 px-4 text-sm bg-white",
|
||||
value === o.value
|
||||
? "text-[#0062D9] ring-1 ring-inset ring-[#0079FE] z-10" // 描边选中
|
||||
: "text-[#3D3D3D] hover:text-[#0079FE]"
|
||||
)}>
|
||||
{o.label}{o.count != null && <span className="text-[#A6A6A6]">({o.count})</span>}
|
||||
</button>
|
||||
))}
|
||||
</div>
|
||||
```
|
||||
|
||||
> 浅底式把选中态换成 `bg-[#EAF4FF] text-[#0062D9]`；实底式换成 `bg-[#0062D9] text-white`。
|
||||
> 分离式去掉 `divide-x` / `overflow-hidden`，外层用 `gap-2`，每个按钮单独 `rounded border`。
|
||||
|
||||
### 按钮式页签（Tab as Button）
|
||||
|
||||
图标 + 文字的页签型按钮：
|
||||
- **激活** 两种：描边（白底 + 蓝边框 + 蓝字）/ 实底（`--color-primary` 底 + 白字）
|
||||
- **正常**：白底 + `--color-border` 边框 + `--color-text`，前置图标
|
||||
- 结构与拼接式选择按钮一致，区别在每项前置一个功能图标。
|
||||
|
||||
---
|
||||
|
||||
## 5.11 输入 - 选择器（Select / Dropdown）
|
||||
|
||||
单项输入选择，下拉单选。
|
||||
|
||||
### 状态
|
||||
|
||||
| 状态 | 触发框 |
|
||||
| --- | --- |
|
||||
| 占位 | `请选择` 占位文字用 `--color-text-secondary`，灰边框 `--color-border`，右侧 `⌄` |
|
||||
| 展开/聚焦 | 边框变 `--color-primary`，箭头翻转为 `⌃` |
|
||||
| 已选 | 显示选中项文字（正文色），保持聚焦时蓝边框 |
|
||||
|
||||
### 下拉面板
|
||||
|
||||
- 白底、圆角 4px、投影（如 `shadow-lg`），最大高度后出现滚动条
|
||||
- 选项默认 `--color-text`；**hover / 选中** 项：`--color-primary-bg`(`#EAF4FF`) 底 + `--color-primary` 文字
|
||||
- **分组**：分组标题用小号灰字 `--color-text-secondary`，组之间用 1px 分割线 `--color-border`
|
||||
|
||||
实现示意：
|
||||
|
||||
```tsx
|
||||
// 触发框
|
||||
<button className="h-8 w-full px-3 inline-flex items-center justify-between rounded
|
||||
border border-[#DBDBDB] bg-white text-sm
|
||||
data-[state=open]:border-[#0062D9]">
|
||||
<span className={selected ? "text-[#3D3D3D]" : "text-[#A6A6A6]"}>{selected ?? "请选择"}</span>
|
||||
<ChevronDown className="size-4 text-[#A6A6A6] transition-transform data-[state=open]:rotate-180" />
|
||||
</button>
|
||||
|
||||
// 下拉项
|
||||
<li className={cn(
|
||||
"h-9 px-3 flex items-center text-sm cursor-pointer rounded-sm",
|
||||
active ? "bg-[#EAF4FF] text-[#0062D9]" : "text-[#3D3D3D] hover:bg-[#EAF4FF]"
|
||||
)}>{label}</li>
|
||||
|
||||
// 分组标题
|
||||
<div className="px-3 py-1.5 text-xs text-[#A6A6A6]">分组标题1</div>
|
||||
```
|
||||
|
||||
> 推荐直接基于 Radix `Select`（项目已有 shadcn 封装）改这些类名落地。
|
||||
|
||||
---
|
||||
|
||||
## 5.14 输入 - 开关（Switch）
|
||||
|
||||
两种形态：
|
||||
|
||||
### 1. 滑块开关（多用于表单、表格中）
|
||||
|
||||
- **开**：胶囊底 `--color-primary`，白色圆形滑块在右，左侧白字「开」
|
||||
- **关**：胶囊底灰 `--color-text-secondary`/`#C1C1C1`，滑块在左，右侧灰字「关」
|
||||
- 切换时滑块带左右过渡动画
|
||||
|
||||
```tsx
|
||||
<button role="switch" aria-checked={on}
|
||||
className={cn(
|
||||
"relative h-6 w-12 rounded-full transition-colors text-[11px] text-white",
|
||||
on ? "bg-[#0062D9]" : "bg-[#C1C1C1]"
|
||||
)}>
|
||||
<span className={cn("absolute top-0.5 size-5 rounded-full bg-white transition-all",
|
||||
on ? "left-[26px]" : "left-0.5")} />
|
||||
<span className={cn("absolute top-1/2 -translate-y-1/2", on ? "left-2" : "right-2 text-[#7F7F7F]")}>
|
||||
{on ? "开" : "关"}
|
||||
</span>
|
||||
</button>
|
||||
```
|
||||
|
||||
### 2. 单选框作为开关（分段式）
|
||||
|
||||
- `开 | 关` 两段式：激活段 `--color-primary` 实底白字，未激活段白底灰字 + `--color-border` 边框
|
||||
- 即 5.9 拼接式「实底选择按钮」的两选项特例
|
||||
|
||||
```tsx
|
||||
<div className="inline-flex rounded overflow-hidden border border-[#DBDBDB]">
|
||||
<button className={cn("h-7 px-4 text-sm", on ? "bg-[#0062D9] text-white" : "bg-white text-[#3D3D3D]")}>开</button>
|
||||
<button className={cn("h-7 px-4 text-sm border-l border-[#DBDBDB]", !on ? "bg-[#0062D9] text-white" : "bg-white text-[#3D3D3D]")}>关</button>
|
||||
</div>
|
||||
```
|
||||
@@ -1,161 +0,0 @@
|
||||
# 需求文档 · 问答中心左侧栏改造（0616）
|
||||
|
||||
> 整理自 2026-06-16 需求草稿。本批需求统一围绕「问答中心」菜单下 5 个页面入口的**左侧侧边栏**进行改造，目标是让每个业务页面的左栏从「通用导航」变为「贴合该业务场景的快捷入口 + 历史记录」。
|
||||
>
|
||||
> 菜单结构参考：`frontend-web/src/core/page-layout/sidebar-menu.ts` 中 `qa-center`（问答中心）分组，含 5 个子项，正好对应下文 5 条需求。
|
||||
|
||||
## 0. 背景与共性约定
|
||||
|
||||
| 需求 | 菜单项 | id | 现有路由 | 对应页面/组件 |
|
||||
|------|--------|----|----------|--------------|
|
||||
| 1 | 通用问答 | `general-qa` | `/page/workspace/chats/new` | Workspace Chat + 左栏（`recent-chat-list.tsx` 等） |
|
||||
| 2 | 多智能体会商 | `multi-agent-roundtable` | `/page/strategy/qa/multi-agent-roundtable` | `roundtable-planning/pages/RoundtablePlanningPage.tsx` |
|
||||
| 3 | zzJC支持 | `qa-zzjc` | `/page/strategy/agents/6bf/chats/new` | `pages/AgentChatPage.tsx` |
|
||||
| 4 | zzXD应用 | `qa-zzxd` | `/page/strategy/agents/10d69686718349688e58a8790cb61024/chats/new` | `pages/AgentChatPage.tsx` |
|
||||
| 5 | 课题研究 | `qa-research` | `/page/canvas/ai-writing` | `open-canvas/pages/AIWritingPage.tsx` |
|
||||
|
||||
**共性约定（适用于全部 5 条，除非单条另有说明）：**
|
||||
|
||||
- 「更多」一律为可点击入口；点击后跳转到对应的完整列表/管理页（具体目标见各条「待确认」）。
|
||||
- 列表为空时的兜底：参考各条「现状」中的级联/空态规则；无内容时该区块整体隐藏或显示占位文案（待 UI 设计确认，默认隐藏，与现有 `recent-chat-list.tsx` 一致）。
|
||||
- 左栏宽度、折叠交互沿用现有侧边栏体系（`SidebarGroup` 系列组件）。
|
||||
- 所有「智能体」相关数据复用 `core/agents`（`Agent` 类型已含 `is_pinned` / `is_favorite` / `featured_order`，scope 支持 `mine` / `square` / `builtin`）。
|
||||
|
||||
---
|
||||
|
||||
## 1. 通用问答 — 左侧栏改造
|
||||
|
||||
**页面**：`/page/workspace/chats/new`
|
||||
|
||||
### 1.1 现状
|
||||
当前左栏主要展示「最近会话」列表（`recent-chat-list.tsx`，已过滤 `scheduler` / `bootstrap` / `roundtable` 类型线程，置顶会话优先）。
|
||||
|
||||
### 1.2 目标布局（自上而下）
|
||||
左栏从上到下依次为三块：**新建回答按钮 → 常用智能体 → 聊天记录**。
|
||||
|
||||
#### ① 顶部：新建回答
|
||||
- 左栏最顶部放置「**新建回答**」按钮（主操作）。
|
||||
- 点击 = 开启一次全新的通用问答会话（等价于跳转 `/page/workspace/chats/new`）。
|
||||
|
||||
#### ② 常用智能体（最多 3 个 + 更多）
|
||||
- 展示「常用智能体」卡片/列表项，**最多 3 个**。
|
||||
- 数据来源采用**级联兜底**逻辑（取到即用，不足再降级）：
|
||||
1. 若用户有**置顶智能体**（`is_pinned = true`）→ 展示置顶的；
|
||||
2. 否则展示**用户自己的智能体**（`scope = "mine"`）；
|
||||
3. 仍没有 → 展示**默认广场智能体**（`scope = "square"` / `builtin`，即广场默认推荐）。
|
||||
- 三者按上述优先级取，最终最多取 3 个。
|
||||
- 下方显示「**更多**」入口。
|
||||
- 点击某个常用智能体 → 进入该智能体的对话（待确认：是否为该智能体新建会话）。
|
||||
|
||||
#### ③ 聊天记录（最多 10 条 + 更多）
|
||||
- 展示用户的问答聊天记录，**最多 10 条**（沿用现有 `useThreads` + 过滤规则）。
|
||||
- 下方显示「**更多**」入口。
|
||||
- 点击「更多」→ **跳转到问答记录页面**（完整聊天记录列表页）。
|
||||
|
||||
### 1.3
|
||||
- 「常用智能体」点击后「打开该智能体主页」
|
||||
- 「常用智能体」的「更多」跳转目标页 智能体管理 `/page/strategy/agents`。
|
||||
- 「聊天记录」的「问答记录页面」具体路由http://localhost:5173/#/page/workspace/chats。
|
||||
- 级联兜底是（置顶组够 3 个就完全不展示我的/广场）如果不够「补足到 3 个
|
||||
|
||||
---
|
||||
|
||||
## 2. 多智能体会商 — 左侧栏重做
|
||||
|
||||
**页面**：`/page/strategy/qa/multi-agent-roundtable`（`RoundtablePlanningPage`）
|
||||
|
||||
### 2.1 现状
|
||||
圆桌会商页已有业务链选择、分析草稿等能力（`roundtable-planning/api/chains.ts`、`hooks/useRoundtableDrafts.ts`）。需求要求**左栏整体重做**。
|
||||
|
||||
### 2.2 目标布局（自上而下）
|
||||
|
||||
#### ① 顶部：业务链（公共，前 3 个）
|
||||
- 顶部展示**业务链**，仅取**公共**链条（`listChains("public")`，即所有已发布 `is_public = true` 的链条）。
|
||||
- 仅展示**前 3 个**。
|
||||
- 点击某条业务链 → 以该业务链启动一次会商（沿用现有"会商自动串联圆桌"流程，见记忆 `business-chain-mapping`）。
|
||||
|
||||
#### ② 下方：分析记录
|
||||
- 业务链下方展示「**分析记录**」列表。
|
||||
- 数据来源：圆桌分析草稿/历史（`useRoundtableDrafts` / `listDrafts`）。
|
||||
- 点击某条分析记录 → 打开对应的历史会商分析。
|
||||
|
||||
### 2.3 待确认
|
||||
- 「业务链」是否需要「更多」入口跳转到业务链条配置页（`/page/strategy/business-chains`）？需求未提及，默认仅展示前 3 个、无更多。 需要更多入口
|
||||
- 「分析记录」是否需要分页/「更多」？需求未提及，默认展示全部或合理上限 分析记录默认全部吧。
|
||||
- 公共业务链「前 3 个」的排序口径（创建时间倒序？已有 `listChains` 默认 newest first） 按照现在默认的显示来吧，然后要支持用户单选，如果用户选择后，点击进入多智能体研讨按钮时，直接按照选择的业务链进去执行，如果用户取消勾选，或没有勾选就还是打开弹窗。
|
||||
|
||||
---
|
||||
|
||||
## 3. zzJC支持 — 左侧栏改造（按类型聚合智能体）
|
||||
|
||||
**页面**：`/page/strategy/agents/6bf/chats/new`（当前为单一智能体 `6bf` 的对话页）
|
||||
|
||||
### 3.1 现状
|
||||
当前 zzJC支持直接进入某个固定智能体（id=`6bf`）的对话页（`AgentChatPage`）。
|
||||
|
||||
### 3.2 目标布局
|
||||
- **左栏按「类型」分组**，展示该类型下的智能体列表（"类型"待确认，推测为智能体 `tags` / 分类）。
|
||||
- 智能体分组**下方**展示**聊天记录**。
|
||||
- 左栏**宽度与通用问答（需求 1）保持一致**（"左侧是问答宽"）。
|
||||
|
||||
### 3.3 Admin 设置入口
|
||||
- 在 **admin 账号权限下**，页面**右侧顶部显示「设置」图标按钮**（齿轮图标）。
|
||||
- 点击设置按钮 → 打开设置面板，可对该模块进行配置。
|
||||
- 非 admin 账号不显示设置按钮（沿用 `sidebar-menu.ts` 中 `adminOnly` 同款角色判断）。
|
||||
|
||||
### 3.4 待确认
|
||||
- 「类型」的确切含义：是智能体的 `tags`，还是 zzJC 业务自定义的一组分类？
|
||||
- 现 zzJC 入口是单个智能体 `6bf`；改为「按类型展示多个智能体」后，**进入该菜单时默认落地页是什么**（智能体列表？还是仍默认进 `6bf` 对话）？是否需要新建一个 zzJC 聚合页。
|
||||
- Admin「设置」具体可配置项：配置「该模块纳入哪些类型/哪些智能体」？还是其他（如默认智能体、欢迎语）？需明确设置项清单。
|
||||
- 「聊天记录」过滤范围：仅 zzJC 相关会话，还是全部会话？
|
||||
|
||||
---
|
||||
|
||||
## 4. zzXD应用 — 左侧栏改造
|
||||
|
||||
**页面**：`/page/strategy/agents/10d69686718349688e58a8790cb61024/chats/new`
|
||||
|
||||
### 4.1 需求
|
||||
**与需求 3（zzJC支持）完全一致**：
|
||||
- 左栏按类型展示该类型下的智能体 + 下方聊天记录；
|
||||
- 左栏宽度同通用问答；
|
||||
- admin 账号下右侧顶部显示设置图标按钮，可设置。
|
||||
|
||||
### 4.2 待确认
|
||||
- 同需求 3 全部待确认项。
|
||||
- zzXD 与 zzJC 是否共用同一套「按类型 + 设置」组件（仅数据源/默认智能体不同），还是各自独立配置？建议**抽象为同一可复用组件**，通过参数区分。
|
||||
|
||||
---
|
||||
|
||||
## 5. 课题研究 — 左侧栏改造（研究模板 + 笔记本）
|
||||
|
||||
**页面**：`/page/canvas/ai-writing`（`AIWritingPage`）
|
||||
|
||||
### 5.1 目标布局（自上而下）
|
||||
|
||||
#### ① 上半部分：研究模板
|
||||
- 左栏上半部分展示「**研究模板**」。
|
||||
- 研究模板 = **原来的「文章类型」**（`open-canvas/hooks/useArticleTypes.ts` 的 `articleTypes`，即 AI 写作表单里的「文章类型」选择项）。
|
||||
- 点击某个研究模板 → 以该模板/文章类型发起课题研究（沿用现有 `articleType` 设置逻辑）。
|
||||
|
||||
#### ② 下半部分：笔记本
|
||||
- 左栏下半部分展示「**笔记本**」。
|
||||
- 笔记本内容 = **「简洁模式」左侧菜单下方展示的笔记本内容**（`useStudioNotebooks`，即 `workspace-nav-studio.tsx` 中渲染的 studio notebooks 列表）。
|
||||
- 点击某个笔记本 → 打开该笔记本（沿用现有 `/page/workspace/studio/notebooks/{id}` 行为，或作为课题研究的素材来源 `materialSource = 'notebook'`，待确认）。
|
||||
|
||||
> 备注：草稿原文写作「简介模式」，结合代码应为「**简洁模式**」（`appearance-settings-page.tsx` 中的全局界面模式：全量 / 简洁；简洁模式下笔记本在左栏菜单展示）。
|
||||
|
||||
### 5.2 待确认
|
||||
- 「研究模板」点击行为：仅预填表单 `articleType`，还是直接开始生成？
|
||||
- 「笔记本」点击行为：跳转笔记本详情页，还是将笔记本设为当前课题研究的素材来源（`materialSource='notebook'`）？
|
||||
- 上/下两部分是否各自可折叠、是否需要「更多」入口与数量上限？需求未提及。
|
||||
|
||||
---
|
||||
|
||||
## 6. 整体待确认与建议
|
||||
|
||||
1. **复用性**：需求 3 / 4（zzJC / zzXD）布局完全相同，建议实现为同一套「按类型分组 + 聊天记录 + admin 设置」可复用组件，通过配置区分模块。
|
||||
2. **左栏宽度统一**：需求 1 / 3 / 4 都要求"问答宽"，建议统一左栏宽度常量，避免各页面各写一套。
|
||||
3. **「更多」目标页**：需求 1 涉及「智能体更多」「聊天记录更多→问答记录页」两个跳转目标需先确定路由是否已存在。
|
||||
4. **Admin 设置项清单**：需求 3 / 4 的设置面板具体能配置什么，需要产品给出字段清单后才能落地。
|
||||
5. **入口语义变更**：需求 3 / 4 把原「单一智能体对话」入口改为「按类型聚合」，菜单 `path` 与默认落地页可能需要调整，需同步确认。
|
||||
@@ -1,148 +0,0 @@
|
||||
# 定时任务 — 模板填充「源数据抽取」手工测试清单
|
||||
|
||||
> 覆盖功能：定时任务「模板数据转换器」内置 agent —— 把用户传入的源数据转换成 HTML
|
||||
> 模板的 `DATA` 结构并填充生成可预览页面。**重点测「源数据抽取」**：源可能是
|
||||
> ① 纯 JSON、② 含 JSON 的文本、③ 纯文本，三种形态都要正确处理。
|
||||
>
|
||||
> 测试环境：后端 `http://localhost:8001`（`make dev`），前端 `http://localhost:5174`（`pnpm dev`）。
|
||||
> 相关代码：`deerflow/runtime/scheduler/template_fill.py`（`extract_embedded_json`）、
|
||||
> `service.py`（`_run_template_fill_attempt` / `_build_conversion_prompt`）、
|
||||
> `pages/ScheduledTasksPage.tsx`。
|
||||
|
||||
---
|
||||
|
||||
## 前置 — 进入表单
|
||||
|
||||
| # | 步骤 | 预期结果 |
|
||||
|---|------|----------|
|
||||
| 0.1 | 任务管理 → 新建定时任务 | 弹出「新建定时任务」对话框 |
|
||||
| 0.2 | 「执行智能体」下拉选择 **模板数据转换器** | 表单出现「HTML 模板」下拉；「必须生成 HTML/Markdown」两个开关隐藏 |
|
||||
| 0.3 | 「HTML 模板」下拉选择 **no.1** | 下方出现「上传 .json」按钮 + 源数据文本框；说明含「未提供源 JSON 时，将直接展示该模板的默认数据」|
|
||||
|
||||
---
|
||||
|
||||
## T1 — 源形态①：纯 JSON
|
||||
|
||||
| # | 步骤 / 输入 | 预期结果 |
|
||||
|---|------|----------|
|
||||
| 1.1 | 在源数据框粘贴纯 JSON（示例见下「样例 A」） | 文本框正常接收，无报错 |
|
||||
| 1.2 | 保存任务 → 「立即执行」 | 执行记录新增一条，状态最终 succeeded |
|
||||
| 1.3 | 打开执行记录 → HTML 预览 | 渲染出 no.1 大屏，**总览指标/各列表数据来自粘贴的 JSON**（而非模板默认值）|
|
||||
| 1.4 | 执行记录正文 | 文案为「已根据源数据填充模板「no.1」生成页面」|
|
||||
|
||||
## T2 — 源形态②：含 JSON 的文本（抽取重点）
|
||||
|
||||
| # | 步骤 / 输入 | 预期结果 |
|
||||
|---|------|----------|
|
||||
| 2.1 | 粘贴「样例 B」（前后是中文说明，中间夹一段 JSON） | 文本框正常接收 |
|
||||
| 2.2 | 保存 → 立即执行 → 查看 HTML | 页面数据以**夹带的 JSON 为准**填充；正文里的补充说明（如「重点突出 X」）被一并参考 |
|
||||
| 2.3 | 验证抽取正确性：JSON 字符串值里含 `}`（如 `"备注": "进度 80%}尚未完成"`） | 不会被误判为 JSON 块提前结束，整段 JSON 被完整抽取 |
|
||||
|
||||
## T3 — 源形态③：纯文本（无 JSON）
|
||||
|
||||
| # | 步骤 / 输入 | 预期结果 |
|
||||
|---|------|----------|
|
||||
| 3.1 | 粘贴「样例 C」（一段纯中文描述，无任何 JSON） | 文本框正常接收 |
|
||||
| 3.2 | 保存 → 立即执行 → 查看 HTML | 智能体**理解文本语义**后据实填充模板，产出完整可预览页面（数据可能较稀疏但结构完整）|
|
||||
| 3.3 | 不应报「未返回 JSON」类硬失败（除非模型彻底失败重试耗尽）| 状态 succeeded |
|
||||
|
||||
## T4 — 空源 → 模板默认数据
|
||||
|
||||
| # | 步骤 | 预期结果 |
|
||||
|---|------|----------|
|
||||
| 4.1 | 源数据框留空，保存 → 立即执行 | **不调用模型**，直接用 no.1 模板自带的默认数据生成 |
|
||||
| 4.2 | 执行记录正文 | 「未提供源数据，已用模板「no.1」的默认数据生成页面」|
|
||||
| 4.3 | HTML 预览 | 展示模板内置的默认大屏 |
|
||||
|
||||
## T5 — 上传 .json 文件
|
||||
|
||||
| # | 步骤 | 预期结果 |
|
||||
|---|------|----------|
|
||||
| 5.1 | 点击「上传 .json 文件」，选一个合法 `.json` | 文件内容被读入源数据文本框 |
|
||||
| 5.2 | 上传一个非法 JSON 内容的文件 | toast 提示「文件不是合法 JSON，已填入文本框请检查」，内容仍填入（可手改）|
|
||||
| 5.3 | 连续上传同一个文件两次 | 第二次仍能触发读取（input 已被重置 value）|
|
||||
|
||||
## T6 — no.2（只展示模板）
|
||||
|
||||
| # | 步骤 | 预期结果 |
|
||||
|---|------|----------|
|
||||
| 6.1 | 「HTML 模板」下拉选 **no.2** | **隐藏**上传按钮与源数据文本框；显示「该模板暂为直接展示已有数据，无需填写」|
|
||||
| 6.2 | 即使之前填过源数据，保存 → 立即执行 | 忽略源数据，**原样输出** no.2 现有页面 |
|
||||
| 6.3 | 执行记录正文 | 「已展示模板「no.2」的现有数据」|
|
||||
|
||||
## T7 — 编辑回显
|
||||
|
||||
| # | 步骤 | 预期结果 |
|
||||
|---|------|----------|
|
||||
| 7.1 | 对已保存的模板任务点「编辑」 | 「执行智能体」「HTML 模板」「源数据」均正确回显 |
|
||||
| 7.2 | 把模板从 no.1 改成 no.2 再保存 | 切到 no.2 时源数据区即时隐藏；保存后执行走只展示分支 |
|
||||
|
||||
---
|
||||
|
||||
## 样例数据（可直接复制粘贴到源数据框）
|
||||
|
||||
### 样例 A — 纯 JSON
|
||||
|
||||
```json
|
||||
{
|
||||
"overview": {
|
||||
"核心指标": [
|
||||
{ "名称": "综合区域", "数值": 7, "单位": "区", "说明": "测试数据" },
|
||||
{ "名称": "组织与单位", "数值": 188, "单位": "个", "说明": "测试数据" }
|
||||
]
|
||||
},
|
||||
"regions": [
|
||||
{ "name": "测试区域甲", "scope": "本岛西岸" }
|
||||
],
|
||||
"equipment": [
|
||||
{ "name_zh": "测试装备X", "category": "无人系统", "status": "列装" }
|
||||
]
|
||||
}
|
||||
```
|
||||
|
||||
> 预期：总览第一项「综合区域」显示 **7**，「组织与单位」显示 **188**，区域/装备出现「测试区域甲」「测试装备X」。
|
||||
|
||||
### 样例 B — 含 JSON 的文本
|
||||
|
||||
```
|
||||
这是本月台海态势数据，请据此填充大屏，重点突出离岛防卫：
|
||||
|
||||
{"overview": {"核心指标": [{"名称": "综合区域", "数值": 9, "单位": "区", "说明": "含离岛"}]}, "regions": [{"name": "离岛防区", "scope": "外岛", "备注": "进度 80%}尚未完成"}]}
|
||||
|
||||
以上数据来自 6 月例会，务必准确，不要臆造未提供的字段。
|
||||
```
|
||||
|
||||
> 预期：① 数据以中间 JSON 为准（综合区域=9、出现「离岛防区」）；② 字符串里的 `}`（`进度 80%}`）不破坏抽取；③ 前后的「重点突出离岛防卫」「务必准确」作为补充说明被模型参考。
|
||||
|
||||
### 样例 C — 纯文本
|
||||
|
||||
```
|
||||
请整理一份台湾滨海防卫概览：覆盖本岛沿岸 9 个综合区域，组织与单位约 280 个，
|
||||
装备条目近百项（含无人系统、防空、海军平台等类别），并列出若干关键节点与近期演习。
|
||||
数据以离线知识为准，没有的字段留空即可。
|
||||
```
|
||||
|
||||
> 预期：智能体理解语义后据实填充，产出结构完整的页面（具体数值取决于模型，可较稀疏），不报硬失败。
|
||||
|
||||
---
|
||||
|
||||
## 后端单元测试（自动化补充）
|
||||
|
||||
```bash
|
||||
# 在 offline-backend-20260512/backend/ 下运行
|
||||
PYTHONPATH=. uv run --no-sync pytest tests/test_template_fill.py -v # 引擎：抽取/校验/填充/注册表
|
||||
PYTHONPATH=. uv run --no-sync pytest tests/test_template_fill_attempt.py -v # 调度器执行分支：三形态/默认/只展示/重试
|
||||
PYTHONPATH=. uv run --no-sync pytest tests/test_template_builder_seed.py -v # 内置 agent 种子
|
||||
```
|
||||
|
||||
抽取相关关键用例：
|
||||
- `test_extract_embedded_json_pure_object` / `_pure_array` —— 纯 JSON。
|
||||
- `test_extract_embedded_json_text_with_json` / `_fenced` —— 含 JSON 的文本 / ```json 围栏。
|
||||
- `test_extract_embedded_json_plain_text_returns_none` —— 纯文本返回 None。
|
||||
- `test_extract_embedded_json_ignores_braces_in_strings` —— 引号内 `}` 不误判。
|
||||
- `test_text_with_embedded_json_surfaced_in_prompt` —— 转换 prompt 单独标注「识别到的结构化数据」。
|
||||
- `test_plain_text_source_converts` —— 纯文本仍照常转换。
|
||||
- `test_empty_source_uses_template_default` —— 空源用模板默认数据。
|
||||
- `test_display_only_template_renders_as_is` —— no.2 原样输出。
|
||||
|
||||
三个文件应全部通过（0 failed）。
|
||||
@@ -13,11 +13,9 @@
|
||||
} catch (e) {}
|
||||
})();
|
||||
</script>
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<link rel="icon" type="image/svg+xml" href="/favicon.svg" />
|
||||
<title>CM 助手</title>
|
||||
<script src="/runtime-config.js"></script>
|
||||
</head>
|
||||
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
|
||||
<title>CM 助手</title>
|
||||
</head>
|
||||
<body>
|
||||
<div id="root"></div>
|
||||
<script type="module" src="/src/main.tsx"></script>
|
||||
|
||||
Generated
+10140
-10540
File diff suppressed because it is too large.
Load diff
@@ -7,12 +7,9 @@
|
||||
"dev": "vite",
|
||||
"build": "node --max-old-space-size=8192 ./node_modules/vite/bin/vite.js build",
|
||||
"build:sit": "node --max-old-space-size=8192 ./node_modules/vite/bin/vite.js build --mode sit",
|
||||
"build:test": "node --max-old-space-size=8192 ./node_modules/vite/bin/vite.js build --mode test",
|
||||
"build:network": "node --max-old-space-size=8192 ./node_modules/vite/bin/vite.js build --mode network",
|
||||
"preview": "vite preview",
|
||||
"typecheck": "tsc --noEmit",
|
||||
"check:chrome109-css": "node scripts/check-chrome109-css.mjs",
|
||||
"check:legacy-chrome-css": "node scripts/check-chrome109-css.mjs"
|
||||
"typecheck": "tsc --noEmit"
|
||||
},
|
||||
"dependencies": {
|
||||
"@codemirror/lang-css": "^6.3.1",
|
||||
@@ -25,7 +22,6 @@
|
||||
"@dnd-kit/core": "^6.3.1",
|
||||
"@dnd-kit/sortable": "^10.0.0",
|
||||
"@dnd-kit/utilities": "^3.2.2",
|
||||
"@fontsource/noto-sans-sc": "^5.2.9",
|
||||
"@langchain/core": "^1.1.15",
|
||||
"@langchain/langgraph-sdk": "^1.5.3",
|
||||
"@radix-ui/react-avatar": "^1.1.11",
|
||||
@@ -59,7 +55,6 @@
|
||||
"@uiw/react-codemirror": "^4.25.4",
|
||||
"@xyflow/react": "^12.10.0",
|
||||
"ai": "^6.0.33",
|
||||
"axios": "1.16.1",
|
||||
"best-effort-json-parser": "^1.2.1",
|
||||
"canvas-confetti": "^1.9.4",
|
||||
"class-variance-authority": "^0.7.1",
|
||||
@@ -68,7 +63,6 @@
|
||||
"codemirror": "^6.0.2",
|
||||
"date-fns": "^4.1.0",
|
||||
"dayjs": "^1.11.21",
|
||||
"dnd-kit": "^0.0.2",
|
||||
"docx": "^9.5.1",
|
||||
"dotenv": "^17.2.3",
|
||||
"embla-carousel-react": "^8.6.0",
|
||||
@@ -77,7 +71,6 @@
|
||||
"hast": "^1.0.0",
|
||||
"katex": "^0.16.28",
|
||||
"lucide-react": "^0.562.0",
|
||||
"mermaid": "^11.15.0",
|
||||
"motion": "^12.26.2",
|
||||
"nanoid": "^5.1.6",
|
||||
"next-themes": "^0.4.6",
|
||||
@@ -112,7 +105,6 @@
|
||||
"@types/react": "^19.0.0",
|
||||
"@types/react-dom": "^19.0.0",
|
||||
"@vitejs/plugin-react": "^5.0.4",
|
||||
"lightningcss": "^1.30.2",
|
||||
"postcss": "^8.5.3",
|
||||
"tailwindcss": "^4.0.15",
|
||||
"tw-animate-css": "^1.4.0",
|
||||
|
||||
@@ -1,11 +1,5 @@
|
||||
import tailwindcss from "@tailwindcss/postcss";
|
||||
import legacyColorFallback from "./scripts/postcss-legacy-color-fallback.mjs";
|
||||
import lightningCssLegacyChrome from "./scripts/postcss-lightningcss-chrome109.mjs";
|
||||
|
||||
export default {
|
||||
plugins: [
|
||||
tailwindcss(),
|
||||
legacyColorFallback(),
|
||||
lightningCssLegacyChrome(),
|
||||
],
|
||||
};
|
||||
export default {
|
||||
plugins: {
|
||||
"@tailwindcss/postcss": {},
|
||||
},
|
||||
};
|
||||
Binary file not shown.
|
Before Width: | Height: | Size: 39 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 53 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 33 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 33 KiB |
@@ -1,28 +0,0 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 48 48" role="img" aria-labelledby="title">
|
||||
<title id="title">CM 助手</title>
|
||||
<defs>
|
||||
<linearGradient id="logoTopLeft" x1="7" y1="7" x2="25" y2="20" gradientUnits="userSpaceOnUse">
|
||||
<stop stop-color="#5DB5FF"/>
|
||||
<stop offset="1" stop-color="#328AF2"/>
|
||||
</linearGradient>
|
||||
<linearGradient id="logoTopRight" x1="24" y1="3" x2="43" y2="19" gradientUnits="userSpaceOnUse">
|
||||
<stop stop-color="#62B2FF"/>
|
||||
<stop offset="1" stop-color="#4B98F4"/>
|
||||
</linearGradient>
|
||||
<linearGradient id="logoLeft" x1="5" y1="14" x2="24" y2="45" gradientUnits="userSpaceOnUse">
|
||||
<stop stop-color="#2C7CF0"/>
|
||||
<stop offset="1" stop-color="#1F5EDA"/>
|
||||
</linearGradient>
|
||||
<linearGradient id="logoBottom" x1="9" y1="31" x2="43" y2="40" gradientUnits="userSpaceOnUse">
|
||||
<stop stop-color="#276DE3"/>
|
||||
<stop offset="1" stop-color="#4E9DF5"/>
|
||||
</linearGradient>
|
||||
</defs>
|
||||
|
||||
<!-- The mark is drawn as four independent folded faces. Together they form
|
||||
the open C-shaped cube used in the product reference. -->
|
||||
<path d="M5 13.5 24 3v10L14 18.5 5 13.5Z" fill="url(#logoTopLeft)"/>
|
||||
<path d="M24 3 43 13.5 34 18.5 24 13V3Z" fill="url(#logoTopRight)"/>
|
||||
<path d="M5 13.5 14 18.5v11L24 35v10L5 34.5v-21Z" fill="url(#logoLeft)"/>
|
||||
<path d="M14 29.5 24 35l10-5.5 9 5L24 45 5 34.5l9-5Z" fill="url(#logoBottom)"/>
|
||||
</svg>
|
||||
|
Before Width: | Height: | Size: 1.4 KiB |
@@ -1,29 +0,0 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 64 64" role="img" aria-label="智能助手中心">
|
||||
<defs>
|
||||
<!-- 品牌玫红渐变填充的圆角磁贴 -->
|
||||
<linearGradient id="tile" x1="6" y1="4" x2="58" y2="60" gradientUnits="userSpaceOnUse">
|
||||
<stop stop-color="#FB7185"/>
|
||||
<stop offset="1" stop-color="#E11D48"/>
|
||||
</linearGradient>
|
||||
<!-- 气泡内 AI 闪光星：玫红渐变，与白色气泡形成对比 -->
|
||||
<linearGradient id="spark" x1="24" y1="18" x2="40" y2="36" gradientUnits="userSpaceOnUse">
|
||||
<stop stop-color="#FB7185"/>
|
||||
<stop offset="1" stop-color="#E11D48"/>
|
||||
</linearGradient>
|
||||
</defs>
|
||||
|
||||
<!-- 磁贴背景 -->
|
||||
<rect width="64" height="64" rx="15" fill="url(#tile)"/>
|
||||
|
||||
<!-- 辅助小闪光，点出「多种智能工具」 -->
|
||||
<path fill="#FFFFFF" fill-opacity="0.85"
|
||||
d="M47 43 Q47.6 46.4 50.5 47 Q47.6 47.6 47 51 Q46.4 47.6 43.5 47 Q46.4 46.4 47 43 Z"/>
|
||||
|
||||
<!-- 助手对话气泡（白色）+ 尾巴 -->
|
||||
<rect x="12" y="13" width="40" height="28" rx="10" fill="#FFFFFF"/>
|
||||
<path fill="#FFFFFF" d="M21 40 H31 L19 49 Z"/>
|
||||
|
||||
<!-- 气泡内的 AI 闪光星 -->
|
||||
<path fill="url(#spark)"
|
||||
d="M32 17.5 Q33.2 25.8 41.5 27 Q33.2 28.2 32 36.5 Q30.8 28.2 22.5 27 Q30.8 25.8 32 17.5 Z"/>
|
||||
</svg>
|
||||
|
Before Width: | Height: | Size: 1.2 KiB |
@@ -1,100 +0,0 @@
|
||||
// 前端运行时配置文件
|
||||
//
|
||||
// 说明：
|
||||
// 1. 本文件会在打包时原样复制到 dist/runtime-config.js。
|
||||
// 2. 应用启动时会优先读取 window.__CM_RUNTIME_CONFIG__ 里的配置。
|
||||
// 3. 部署后如果需要切换接口地址，直接修改 dist/runtime-config.js 即可，不需要重新 pnpm build。
|
||||
// 4. 这里的默认值来自 .env.network；如某个环境不用某项，可以留空字符串。
|
||||
// 5. 布尔开关也用字符串形式填写："true" / "false"。
|
||||
window.__CM_RUNTIME_CONFIG__ = {
|
||||
// DeerFlow Gateway 地址：前端自有后端接口，例如 /api/agents、/api/auth、/api/threads 等。
|
||||
// VITE_BACKEND_BASE_URL: "http://10.19.200.5:2026",
|
||||
|
||||
// LangGraph 流式对话接口地址：普通问答、智能体对话等流式能力使用。
|
||||
// VITE_LANGGRAPH_BASE_URL: "http://10.19.200.5:2026/api",
|
||||
|
||||
// Canvas/AI 写作独立 LangGraph 地址：为空时前端会回退使用 VITE_LANGGRAPH_BASE_URL。
|
||||
VITE_CANVAS_LANGGRAPH_BASE_URL: "",
|
||||
|
||||
// 轻应用「新建笔记」系列接口地址。
|
||||
VITE_NOTE_API_BASE_URL: "http://47.88.25.99:8000",
|
||||
|
||||
// CMZS 后端接口：任务管理、任务分析、三情分析、行动分析等业务接口。
|
||||
VITE_API_BASE_URL: "https://ch1.b.uat.4.cn/zncm",
|
||||
|
||||
// 佳来模型对话接口。
|
||||
VITE_AIPI_CHAT_URL: "https://ch1.b.uat.4.cn/gpt",
|
||||
|
||||
// 多轮检索/生成接口。
|
||||
VITE_AIPI_GENERATE_CHAT_URL: "https://ch1.b.uat.4.cn/jialai_b1_chat",
|
||||
|
||||
// 任务问答接口。
|
||||
VITE_AIPI_TASK_CHAT_URL: "https://ch1.b.uat.4.cn/qq_b1_chat",
|
||||
|
||||
// B 系统知识库前端地址：用于打开知识详情、原文详情等。
|
||||
VITE_AIPI_KNOWLEDGE_URL: "https://chlib.b.uat.4.cn/knowlibrary",
|
||||
|
||||
// B1 COH 系统地址。
|
||||
VITE_AIPI_B1COH_URL: "https://ch1.b.uat.4.cn/web",
|
||||
|
||||
// B2 行为体/网络资源系统地址。
|
||||
VITE_AIPI_B2ACTOR_URL: "https://ty.b.uat.4.cn",
|
||||
|
||||
// B2 网络 XWT iframe 路径。
|
||||
VITE_AIPI_B2ACTOR_XWT_PATH: "/fromB/toXWT?isFrame=1",
|
||||
|
||||
// 知识库 iframe 地址换取接口。
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_API_URL: "https://ch1.b.uat.4.cn/knowes/api/iframeKnowledgeUrl",
|
||||
|
||||
// 知识库 iframe 皮肤参数。
|
||||
VITE_AIPI_KNOWLEDGE_IFRAME_SKIN: "red",
|
||||
|
||||
// MopeChat 接口地址。
|
||||
VITE_AIPI_MOPECHAT_URL: "https://ch1.b.uat.4.cn/likun_b1_mope",
|
||||
|
||||
// 工作室/空间后端接口：空间列表、空间详情、素材上传等使用。
|
||||
VITE_AIPI_BACKENP_URL: "http://47.88.25.99:8000",
|
||||
|
||||
// 工作室/空间后端别名：兼容旧配置命名，通常与 VITE_AIPI_BACKENP_URL 保持一致。
|
||||
VITE_NOTEBOOK_API_BASE_URL: "http://47.88.25.99:8000",
|
||||
|
||||
// 工作室/空间前端地址：用于 auth/callback、新建对话等跳转。
|
||||
VITE_AIPI_FRONTEND_URL: "http://47.88.25.99:3000",
|
||||
|
||||
// Consumer 接口：深链任务详情 cop-task-detail、研判情报报告 selectSqReport 等使用。
|
||||
VITE_AIPI_CONSUMER_URL: "http://10.19.200.5:8002/consumer",
|
||||
|
||||
// A 系统跳转地址：人物、组织、事件等列表项打开链接使用。
|
||||
VITE_AIPI_PRESENT_URL: "http://present.a.uat.4.cn/present/domain",
|
||||
|
||||
// B 系统 knowes 根地址。
|
||||
VITE_AIPI_KNOWES_URL: "http://ch1.b.uat.4.cn/knowes",
|
||||
|
||||
// MaxKey 鉴权地址：onlineTicket 换 token 使用。
|
||||
VITE_AUTH_BASE_URL: "https://auth.y.sit.4.cn",
|
||||
|
||||
// PG 中任务 iframe 地址。
|
||||
VITE_AIPI_PGLINKTO_URL: "https://pg.b.uat.4.cn",
|
||||
|
||||
// XK 中任务 iframe 地址。
|
||||
VITE_AIPI_XKLINKTO_URL: "https://kz.b.uat.4.cn",
|
||||
|
||||
// 是否使用真实任务看板数据：true=真实接口；false=本地 mock。
|
||||
VITE_USE_REAL_DATA: "true",
|
||||
|
||||
// 登录深链任务详情是否使用本地 mock：false=真实接口；true=本地 mock。
|
||||
VITE_MOCK_COP_TASK_DETAIL: "false",
|
||||
|
||||
// 大屏绘制 selectSqReport 是否使用本地 mock：不在运行时强制覆盖，
|
||||
// 由各环境的 .env 文件决定（开发环境为 true，生产环境为 false）。
|
||||
|
||||
// 静态站点模式开关；一般部署业务系统时保持 false。
|
||||
VITE_STATIC_WEBSITE_ONLY: "false",
|
||||
|
||||
// 紧凑导航开关：true 时固定使用简洁布局，工作台左侧栏收起为菜单图标，
|
||||
// 隐藏系统 Logo、换肤按钮以及“全量/简洁模式”切换。修改后刷新页面即可生效。
|
||||
VITE_FORCE_COLLAPSED_SIDEBAR: "false",
|
||||
|
||||
// GitHub OAuth Token；不使用 GitHub 相关能力时留空。
|
||||
VITE_GITHUB_OAUTH_TOKEN: "",
|
||||
};
|
||||
@@ -1,109 +0,0 @@
|
||||
import fs from "node:fs";
|
||||
import path from "node:path";
|
||||
import postcss from "postcss";
|
||||
|
||||
const unsupportedColor = /(?:color-mix|oklch|oklab|lab|lch)\(/i;
|
||||
const assetsDir = path.resolve("dist/assets");
|
||||
|
||||
if (!fs.existsSync(assetsDir)) {
|
||||
throw new Error("dist/assets 不存在，请先运行 pnpm build");
|
||||
}
|
||||
|
||||
const cssFiles = fs
|
||||
.readdirSync(assetsDir)
|
||||
.filter((file) => file.endsWith(".css"));
|
||||
|
||||
if (cssFiles.length === 0) {
|
||||
throw new Error("dist/assets 中没有 CSS 构建产物");
|
||||
}
|
||||
|
||||
const failures = [];
|
||||
let modernDeclarations = 0;
|
||||
|
||||
const requiredLegacyFallbacks = [
|
||||
{
|
||||
file: "src/components/ui/textarea.tsx",
|
||||
pattern: /rows=\{rows\s*\?\?\s*3\}/,
|
||||
description: "Textarea needs a native rows fallback",
|
||||
},
|
||||
{
|
||||
file: "src/components/ai-elements/prompt-input.tsx",
|
||||
pattern: /<InputGroup className="h-auto min-h-16">/,
|
||||
description: "Prompt textarea must not rely on :has() for its parent height",
|
||||
},
|
||||
];
|
||||
|
||||
for (const file of cssFiles) {
|
||||
const root = postcss.parse(
|
||||
fs.readFileSync(path.join(assetsDir, file), "utf8"),
|
||||
{ from: file },
|
||||
);
|
||||
|
||||
root.walkDecls((declaration) => {
|
||||
if (!unsupportedColor.test(declaration.value)) return;
|
||||
modernDeclarations += 1;
|
||||
|
||||
let ancestor = declaration.parent;
|
||||
while (ancestor) {
|
||||
if (
|
||||
ancestor.type === "atrule" &&
|
||||
ancestor.name === "supports" &&
|
||||
unsupportedColor.test(ancestor.params)
|
||||
) {
|
||||
return;
|
||||
}
|
||||
ancestor = ancestor.parent;
|
||||
}
|
||||
|
||||
let previous = declaration.prev();
|
||||
while (previous?.type === "comment") previous = previous.prev();
|
||||
const hasSafeFallback =
|
||||
!declaration.prop.startsWith("--") &&
|
||||
previous?.type === "decl" &&
|
||||
previous.prop === declaration.prop &&
|
||||
!unsupportedColor.test(previous.value);
|
||||
|
||||
if (!hasSafeFallback) {
|
||||
failures.push(
|
||||
`${file}:${declaration.source?.start?.line ?? "?"} ${declaration.prop}: ${declaration.value}`,
|
||||
);
|
||||
}
|
||||
});
|
||||
}
|
||||
|
||||
function walk(directory) {
|
||||
return fs.readdirSync(directory, { withFileTypes: true }).flatMap((entry) => {
|
||||
const fullPath = path.join(directory, entry.name);
|
||||
return entry.isDirectory() ? walk(fullPath) : [fullPath];
|
||||
});
|
||||
}
|
||||
|
||||
for (const file of walk(path.resolve("src"))) {
|
||||
if (!/\.[jt]sx?$/.test(file)) continue;
|
||||
const source = fs.readFileSync(file, "utf8");
|
||||
if (unsupportedColor.test(source)) {
|
||||
failures.push(
|
||||
`${path.relative(process.cwd(), file)} contains a runtime-only modern color function`,
|
||||
);
|
||||
}
|
||||
}
|
||||
|
||||
for (const fallback of requiredLegacyFallbacks) {
|
||||
const source = fs.readFileSync(path.resolve(fallback.file), "utf8");
|
||||
if (!fallback.pattern.test(source)) {
|
||||
failures.push(`${fallback.file}: ${fallback.description}`);
|
||||
}
|
||||
}
|
||||
|
||||
if (failures.length > 0) {
|
||||
console.error("Legacy Chrome CSS compatibility check failed:\n");
|
||||
console.error(failures.slice(0, 50).join("\n"));
|
||||
if (failures.length > 50) {
|
||||
console.error(`... and ${failures.length - 50} more`);
|
||||
}
|
||||
process.exit(1);
|
||||
}
|
||||
|
||||
console.log(
|
||||
`Legacy Chrome (100+) CSS compatibility check passed (${modernDeclarations} guarded modern declarations).`,
|
||||
);
|
||||
@@ -1,70 +0,0 @@
|
||||
/**
|
||||
* Tailwind 4 gracefully falls back from color-mix() to the base semantic
|
||||
* color, but that turns utilities such as bg-primary/5 into an opaque fill on
|
||||
* Chrome 109. Replace that fallback with an sRGB-channel based alpha color.
|
||||
* The modern color-mix() declaration remains in its @supports block.
|
||||
*/
|
||||
|
||||
const LEGACY_RGB_TOKENS = new Set([
|
||||
"accent",
|
||||
"background",
|
||||
"border",
|
||||
"brand-200",
|
||||
"card",
|
||||
"destructive",
|
||||
"foreground",
|
||||
"input",
|
||||
"muted",
|
||||
"muted-foreground",
|
||||
"primary",
|
||||
"primary-foreground",
|
||||
"ring",
|
||||
"secondary",
|
||||
"sidebar-accent",
|
||||
"sidebar-border",
|
||||
"sidebar-foreground",
|
||||
"workspace-brand",
|
||||
]);
|
||||
|
||||
const TRANSPARENT_MIX_RE =
|
||||
/color-mix\(\s*in\s+[\w-]+\s*,\s*var\(--([\w-]+)\)\s+([\d.]+)%\s*,\s*transparent\s*\)/i;
|
||||
|
||||
function legacyAlphaColor(value) {
|
||||
const match = value.match(TRANSPARENT_MIX_RE);
|
||||
if (!match || !LEGACY_RGB_TOKENS.has(match[1])) return null;
|
||||
|
||||
const [, token, percentage] = match;
|
||||
return `rgb(var(--legacy-${token}-rgb) / ${percentage}%)`;
|
||||
}
|
||||
|
||||
export default function legacyColorFallback() {
|
||||
return {
|
||||
postcssPlugin: "postcss-legacy-color-fallback",
|
||||
OnceExit(root) {
|
||||
root.walkAtRules("supports", (atRule) => {
|
||||
if (!atRule.params.includes("color-mix")) return;
|
||||
|
||||
atRule.walkDecls((modernDeclaration) => {
|
||||
const fallbackValue = legacyAlphaColor(modernDeclaration.value);
|
||||
if (!fallbackValue) return;
|
||||
|
||||
// Tailwind emits the old-browser declaration immediately before
|
||||
// the nested @supports block, in the same utility/variant rule.
|
||||
let fallbackDeclaration = atRule.prev();
|
||||
while (fallbackDeclaration?.type === "comment") {
|
||||
fallbackDeclaration = fallbackDeclaration.prev();
|
||||
}
|
||||
|
||||
if (
|
||||
fallbackDeclaration?.type === "decl" &&
|
||||
fallbackDeclaration.prop === modernDeclaration.prop
|
||||
) {
|
||||
fallbackDeclaration.value = fallbackValue;
|
||||
}
|
||||
});
|
||||
});
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
legacyColorFallback.postcss = true;
|
||||
@@ -1,33 +0,0 @@
|
||||
import postcss from "postcss";
|
||||
import { transform } from "lightningcss";
|
||||
|
||||
const LEGACY_CHROME = 100 << 16;
|
||||
|
||||
/**
|
||||
* Keep Vite's PostCSS pipeline (required by Tailwind 4), then compile the
|
||||
* resulting CSS to Chrome 100-compatible syntax for both dev and production.
|
||||
*/
|
||||
export default function lightningCssLegacyChrome() {
|
||||
return {
|
||||
postcssPlugin: "postcss-lightningcss-legacy-chrome",
|
||||
OnceExit(root, { result }) {
|
||||
const filename = root.source?.input.file ?? result.opts.from ?? "style.css";
|
||||
const output = transform({
|
||||
filename,
|
||||
code: Buffer.from(root.toString()),
|
||||
targets: { chrome: LEGACY_CHROME },
|
||||
minify: false,
|
||||
});
|
||||
const compiled = postcss.parse(output.code.toString(), { from: filename });
|
||||
|
||||
root.removeAll();
|
||||
root.append(compiled.nodes);
|
||||
|
||||
for (const warning of output.warnings) {
|
||||
result.warn(warning.message, { plugin: "lightningcss" });
|
||||
}
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
lightningCssLegacyChrome.postcss = true;
|
||||
@@ -1,102 +1,47 @@
|
||||
import { useEffect, useMemo } from "react";
|
||||
import { useMemo } from "react";
|
||||
import { Navigate, Route, Routes } from "react-router-dom";
|
||||
import { ThemeProvider as NextThemesProvider } from "next-themes";
|
||||
|
||||
import { DesensitizeWordsLoader } from "@/components/desensitize-words-loader";
|
||||
import { FontScaleApplier } from "@/components/font-scale";
|
||||
import { QueryClientProvider } from "@/components/query-client-provider";
|
||||
import { AuthSyncWatcher } from "@/core/auth/sync";
|
||||
import { I18nProvider } from "@/core/i18n/context";
|
||||
import { fetchYLogConfig } from "@/core/log/y-log-config";
|
||||
import { detectLocale } from "@/core/i18n/locale";
|
||||
import { resolveForcedThemeForPath } from "@/core/page-layout/chats-iframe-route";
|
||||
import { ChatProvider } from "@/contexts/ChatContext";
|
||||
import { ThemeProvider } from "@/contexts/ThemeContext";
|
||||
import {
|
||||
resolveEmbedRoundtableTheme,
|
||||
resolveEmbedTheme,
|
||||
} from "@/core/embed/params";
|
||||
import { usePublicEmbedTheme } from "@/core/embed/background";
|
||||
import { readWujieProps } from "@/core/embed/wujie";
|
||||
import {
|
||||
resolveForcedEmbedTheme,
|
||||
syncEmbedSessionFromSearch,
|
||||
} from "@/core/embed/embed-session";
|
||||
|
||||
import { HashRouter, useLocation } from "react-router-dom";
|
||||
|
||||
import LandingPage from "./pages/LandingPage";
|
||||
import LoginPage from "./pages/LoginPage";
|
||||
import PageRoutes from "./pages/PageRoutes";
|
||||
import PublicScheduledTaskPage, {
|
||||
PublicScheduledTaskFullscreenPage,
|
||||
} from "./pages/PublicScheduledTaskPage";
|
||||
import PublicScheduledTaskPage from "./pages/PublicScheduledTaskPage";
|
||||
import PublicSharePage from "./pages/PublicSharePage";
|
||||
import PublicLeaderboardPage from "./pages/PublicLeaderboardPage";
|
||||
import PublicRoundtableSharePage from "./pages/PublicRoundtableSharePage";
|
||||
import EmbedChatPage from "./pages/EmbedChatPage";
|
||||
import EmbedRoundtablePage from "./pages/EmbedRoundtablePage";
|
||||
import DashboardStandalonePage from "@/roundtable-planning/pages/DashboardStandalonePage";
|
||||
import { ChatRuntime } from "./pages/WorkspaceRoutes";
|
||||
import { NotFoundPage } from "./pages/NotFoundPage";
|
||||
|
||||
const PUBLIC_EMBED_PREFIXES = ["/share/", "/public/", "/roundtable-share/"];
|
||||
|
||||
function AppProviders({ children }: { children: React.ReactNode }) {
|
||||
const locale = useMemo(() => detectLocale(), []);
|
||||
const location = useLocation();
|
||||
const publicEmbedTheme = usePublicEmbedTheme();
|
||||
// iframe 嵌入会话（按标签页 sessionStorage）集中在此同步：任何带 `&embed=1` 的入口（深链登录页
|
||||
// 或直接深链路由）一进来就启用、`embed=0` 关闭，后续干净路由靠会话维持去 chrome + 强制主题。
|
||||
useEffect(() => {
|
||||
syncEmbedSessionFromSearch(location.search);
|
||||
}, [location.search]);
|
||||
// 操作日志(Y-Log)：启动早期预热上报配置，使 writeYLog 同步读缓存即可工作
|
||||
// （刷新后 module 缓存清空，token 登录用户也能继续上报）。失败静默。
|
||||
useEffect(() => {
|
||||
void fetchYLogConfig();
|
||||
}, []);
|
||||
const forcedTheme = useMemo(() => {
|
||||
// iframe 嵌入模式（外部系统 `&embed=1`）最高优先：按本标签页的嵌入主题强制明/暗。
|
||||
// 走 next-themes forcedTheme（不写 localStorage）→ 同浏览器另开的正常系统页不受影响。
|
||||
const embedTheme = resolveForcedEmbedTheme(location.search);
|
||||
if (embedTheme) return embedTheme;
|
||||
const routeTheme = resolveForcedThemeForPath(location.pathname);
|
||||
if (routeTheme) return routeTheme;
|
||||
if (location.pathname === "/embed/roundtable") {
|
||||
const wp = readWujieProps();
|
||||
const raw =
|
||||
(typeof wp.theme === "string" && wp.theme.trim()) ||
|
||||
new URLSearchParams(location.search).get("theme")?.trim() ||
|
||||
"dark";
|
||||
return resolveEmbedRoundtableTheme(raw);
|
||||
}
|
||||
if (location.pathname.startsWith("/embed/chats")) {
|
||||
const raw = new URLSearchParams(location.search).get("theme")?.trim();
|
||||
return raw ? resolveEmbedTheme(raw) : undefined;
|
||||
}
|
||||
if (PUBLIC_EMBED_PREFIXES.some((p) => location.pathname.startsWith(p))) {
|
||||
return publicEmbedTheme;
|
||||
}
|
||||
return undefined;
|
||||
}, [location.pathname, location.search, publicEmbedTheme]);
|
||||
const forcedTheme = location.pathname === "/landing" ? "light" : undefined;
|
||||
|
||||
return (
|
||||
<NextThemesProvider
|
||||
attribute="class"
|
||||
storageKey="strategy-theme"
|
||||
defaultTheme="light"
|
||||
themes={["light", "dark", "lightblue", "lightgreen"]}
|
||||
enableSystem={false}
|
||||
disableTransitionOnChange
|
||||
forcedTheme={forcedTheme}
|
||||
>
|
||||
{/* QueryClientProvider must wrap ThemeProvider because the custom
|
||||
ThemeProvider reads the admin-configured default via useQuery. */}
|
||||
<QueryClientProvider>
|
||||
<I18nProvider initialLocale={locale}>
|
||||
<ThemeProvider>
|
||||
<ChatProvider>
|
||||
<AuthSyncWatcher />
|
||||
<DesensitizeWordsLoader />
|
||||
<FontScaleApplier />
|
||||
{children}
|
||||
</ChatProvider>
|
||||
@@ -107,47 +52,23 @@ function AppProviders({ children }: { children: React.ReactNode }) {
|
||||
);
|
||||
}
|
||||
|
||||
/**
|
||||
* 根路由 `/` → `/login` 重定向，**保留 query string**。
|
||||
* 外部系统的任务深链落在 `#/?taskId=...&goPath=...`，裸 `<Navigate to="/login">` 会丢掉
|
||||
* query，导致 LoginPage 读不到参数；这里把 search 一并带过去。
|
||||
*/
|
||||
function RootRedirect() {
|
||||
const { search } = useLocation();
|
||||
return <Navigate to={`/login${search}`} replace />;
|
||||
}
|
||||
|
||||
export function App() {
|
||||
return (
|
||||
<HashRouter>
|
||||
<AppProviders>
|
||||
<Routes>
|
||||
<Route path="/" element={<RootRedirect />} />
|
||||
<Route path="/" element={<Navigate to="/login" replace />} />
|
||||
<Route path="/login" element={<LoginPage />} />
|
||||
<Route path="/login/:username" element={<LoginPage />} />
|
||||
<Route path="/landing" element={<LandingPage />} />
|
||||
{/* Public read-only conversation viewer — no authentication. */}
|
||||
<Route path="/share/:shareCode" element={<PublicSharePage />} />
|
||||
<Route path="/public/leaderboard" element={<PublicLeaderboardPage />} />
|
||||
<Route
|
||||
path="/roundtable-share/:shareCode"
|
||||
element={<PublicRoundtableSharePage view="landing" />}
|
||||
/>
|
||||
<Route
|
||||
path="/roundtable-share/:shareCode/report"
|
||||
element={<PublicRoundtableSharePage view="report" />}
|
||||
/>
|
||||
<Route
|
||||
path="/roundtable-share/:shareCode/dashboard"
|
||||
element={<PublicRoundtableSharePage view="dashboard" />}
|
||||
/>
|
||||
{/* Public read-only scheduled-task runs viewer — no authentication. */}
|
||||
<Route
|
||||
path="/public/scheduled-tasks/:taskId"
|
||||
element={<PublicScheduledTaskPage />}
|
||||
/>
|
||||
<Route
|
||||
path="/public/scheduled-tasks/:taskId/runs/:runId/fullscreen"
|
||||
element={<PublicScheduledTaskFullscreenPage />}
|
||||
/>
|
||||
{/* Embedded workspace Q&A — no page/workspace sidebars; parent passes message/username/theme */}
|
||||
<Route
|
||||
path="/embed/chats/:thread_id"
|
||||
element={
|
||||
@@ -156,11 +77,6 @@ export function App() {
|
||||
</ChatRuntime>
|
||||
}
|
||||
/>
|
||||
<Route path="/embed/roundtable" element={<EmbedRoundtablePage />} />
|
||||
<Route
|
||||
path="/roundtable-dashboard"
|
||||
element={<DashboardStandalonePage />}
|
||||
/>
|
||||
<Route path="/page/*" element={<PageRoutes />} />
|
||||
<Route path="*" element={<NotFoundPage />} />
|
||||
</Routes>
|
||||
|
||||
Binary file not shown.
|
Before Width: | Height: | Size: 2.5 KiB |
@@ -1 +0,0 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="1920" height="62" viewBox="0 0 1920 62"><defs><style>.a{fill:url(#a);}.b{fill:rgba(0,98,217,0.1);}.c{fill:#fff;}.d{fill:rgba(0,98,217,0.2);}.e{filter:url(#h);}.f{filter:url(#e);}.g{filter:url(#b);}</style><linearGradient id="a" x1="0.5" x2="0.5" y2="1" gradientUnits="objectBoundingBox"><stop offset="0" stop-color="#0b3776"/><stop offset="1" stop-color="#0c1935"/></linearGradient><filter id="b"><feOffset dy="-3" input="SourceAlpha"/><feGaussianBlur result="c"/><feFlood flood-color="#0062d9" flood-opacity="0.102" result="d"/><feComposite operator="out" in="SourceGraphic" in2="c"/><feComposite operator="in" in="d"/><feComposite operator="in" in2="SourceGraphic"/></filter><filter id="e"><feOffset dy="-3" input="SourceAlpha"/><feGaussianBlur result="f"/><feFlood flood-color="#0062d9" flood-opacity="0.102" result="g"/><feComposite operator="out" in="SourceGraphic" in2="f"/><feComposite operator="in" in="g"/><feComposite operator="in" in2="SourceGraphic"/></filter><filter id="h"><feOffset dy="-3" input="SourceAlpha"/><feGaussianBlur result="i"/><feFlood flood-color="#0062d9" flood-opacity="0.2" result="j"/><feComposite operator="out" in="SourceGraphic" in2="i"/><feComposite operator="in" in="j"/><feComposite operator="in" in2="SourceGraphic"/></filter></defs><g transform="translate(-2009 -3361)"><rect class="a" width="1920" height="62" transform="translate(2009 3361)"/><g data-type="innerShadowGroup"><path class="b" d="M0,0H318L268,62H0Z" transform="translate(2009 3361)"/><g class="g" transform="matrix(1, 0, 0, 1, 2009, 3361)"><path class="c" d="M0,0H318L268,62H0Z"/></g></g><g data-type="innerShadowGroup"><path class="b" d="M0,0H310L260,62H0Z" transform="translate(2009 3361)"/><g class="f" transform="matrix(1, 0, 0, 1, 2009, 3361)"><path class="c" d="M0,0H310L260,62H0Z"/></g></g><g data-type="innerShadowGroup"><path class="d" d="M0,0H302L252,62H0Z" transform="translate(2009 3361)"/><g class="e" transform="matrix(1, 0, 0, 1, 2009, 3361)"><path class="c" d="M0,0H302L252,62H0Z"/></g></g></g></svg>
|
||||
|
Before Width: | Height: | Size: 2.1 KiB |
@@ -1 +0,0 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="1920" height="1080" viewBox="0 0 1920 1080"><image width="1920" height="1080" xlink:href="data:image/jpeg;base64,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 truncated
|
||||
|
Before Width: | Height: | Size: 431 KiB |
@@ -1 +0,0 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="152" height="42.359" viewBox="0 0 152 42.359"><defs><style>.a{fill:url(#a);}</style><pattern id="a" preserveAspectRatio="none" width="100%" height="100%" viewBox="0 0 536 150"><image width="536" height="150" xlink:href="data:image/png;base64,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 truncated
|
||||
|
Before Width: | Height: | Size: 33 KiB |
Binary file not shown.
|
Before Width: | Height: | Size: 3.4 KiB |
@@ -1 +0,0 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="1920" height="64" viewBox="0 0 1920 64"><defs><style>.a{fill:url(#a);}.b{fill:url(#b);}.c{clip-path:url(#c);}.d{fill:url(#d);}</style><linearGradient id="a" x1="0.5" x2="0.5" y2="1" gradientUnits="objectBoundingBox"><stop offset="0" stop-color="#131f3d"/><stop offset="1" stop-color="#094593"/></linearGradient><linearGradient id="b" x1="0.473" x2="0.473" y2="0.95" gradientUnits="objectBoundingBox"><stop offset="0" stop-color="#131f3d"/><stop offset="1" stop-color="#094593" stop-opacity="0"/></linearGradient><clipPath id="c"><rect class="a" width="1920" height="64" transform="translate(-2 5)"/></clipPath><pattern id="d" preserveAspectRatio="xMidYMid slice" width="100%" height="100%" viewBox="0 0 1917 104"><image width="1917" height="104" xlink:href="data:image/png;base64,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 truncated
|
||||
|
Before Width: | Height: | Size: 98 KiB |
@@ -1 +0,0 @@
|
||||
<svg xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" width="32" height="30" viewBox="0 0 32 30"><defs><style>.a{fill:rgba(255,255,255,0.05);}.b{fill:url(#a);}</style><linearGradient id="a" x2="1" y2="1" gradientUnits="objectBoundingBox"><stop offset="0" stop-color="#a2e1f9"/><stop offset="1" stop-color="#45c3f3"/></linearGradient></defs><g transform="translate(1.411 -9.285)"><rect class="a" width="32" height="30" rx="4" transform="translate(-1.411 9.285)"/><path class="b" d="M4618.588,4844.6a1.257,1.257,0,1,1,1.258,1.256A1.258,1.258,0,0,1,4618.588,4844.6Zm4.609.838v-1.676h12.569v1.676Zm-4.609-5.867a1.257,1.257,0,1,1,1.258,1.257A1.257,1.257,0,0,1,4618.588,4839.569Zm4.609.838v-1.676h12.569v1.676Zm-4.609-5.865a1.257,1.257,0,1,1,1.258,1.257A1.257,1.257,0,0,1,4618.588,4834.542Zm4.609.838V4833.7h12.569v1.676Z" transform="translate(-4613 -4815)"/></g></svg>
|
||||
|
Before Width: | Height: | Size: 892 B |
@@ -1,8 +1,7 @@
|
||||
"use client";
|
||||
|
||||
import { Button } from "@/components/ui/button";
|
||||
import { copyTextToClipboard } from "@/core/clipboard";
|
||||
import { cn } from "@/lib/utils";
|
||||
import { Button } from "@/components/ui/button";
|
||||
import { cn } from "@/lib/utils";
|
||||
import { CheckIcon, CopyIcon } from "lucide-react";
|
||||
import {
|
||||
type ComponentProps,
|
||||
@@ -145,13 +144,18 @@ export const CodeBlockCopyButton = ({
|
||||
...props
|
||||
}: CodeBlockCopyButtonProps) => {
|
||||
const [isCopied, setIsCopied] = useState(false);
|
||||
const { code } = useContext(CodeBlockContext);
|
||||
|
||||
const copyToClipboard = async () => {
|
||||
try {
|
||||
await copyTextToClipboard(code);
|
||||
setIsCopied(true);
|
||||
onCopy?.();
|
||||
const { code } = useContext(CodeBlockContext);
|
||||
|
||||
const copyToClipboard = async () => {
|
||||
if (typeof window === "undefined" || !navigator?.clipboard?.writeText) {
|
||||
onError?.(new Error("Clipboard API not available"));
|
||||
return;
|
||||
}
|
||||
|
||||
try {
|
||||
await navigator.clipboard.writeText(code);
|
||||
setIsCopied(true);
|
||||
onCopy?.();
|
||||
setTimeout(() => setIsCopied(false), timeout);
|
||||
} catch (error) {
|
||||
onError?.(error as Error);
|
||||
|
||||
@@ -846,10 +846,7 @@ export const PromptInput = ({
|
||||
ref={formRef}
|
||||
{...props}
|
||||
>
|
||||
{/* Do not rely on InputGroup's :has(textarea) selector here: Chrome
|
||||
versions before 105 ignore :has(), retain h-9, and clip the whole
|
||||
multi-line composer to a single row. */}
|
||||
<InputGroup className="h-auto min-h-16">{children}</InputGroup>
|
||||
<InputGroup>{children}</InputGroup>
|
||||
</form>
|
||||
</PromptInputValidationContext.Provider>
|
||||
);
|
||||
@@ -876,48 +873,16 @@ export type PromptInputTextareaProps = ComponentProps<
|
||||
typeof InputGroupTextarea
|
||||
>;
|
||||
|
||||
export const PromptInputTextarea = ({
|
||||
onChange,
|
||||
onInput,
|
||||
className,
|
||||
placeholder = "What would you like to know?",
|
||||
...props
|
||||
}: PromptInputTextareaProps) => {
|
||||
const controller = useOptionalPromptInputController();
|
||||
const attachments = usePromptInputAttachments();
|
||||
const sanitizeIncomingFiles = usePromptInputValidation();
|
||||
const [isComposing, setIsComposing] = useState(false);
|
||||
const textareaRef = useRef<HTMLTextAreaElement | null>(null);
|
||||
|
||||
const resizeForLegacyChrome = useCallback((textarea: HTMLTextAreaElement) => {
|
||||
if (
|
||||
typeof CSS !== "undefined" &&
|
||||
typeof CSS.supports === "function" &&
|
||||
CSS.supports("field-sizing", "content")
|
||||
) {
|
||||
return;
|
||||
}
|
||||
|
||||
// Chrome 109 ignores field-sizing: content. Preserve the same 4rem–12rem
|
||||
// autosize range used by min-h-16/max-h-48 instead of leaving the prompt
|
||||
// stuck at its minimum height.
|
||||
textarea.style.height = "auto";
|
||||
const height = Math.max(64, Math.min(textarea.scrollHeight, 192));
|
||||
textarea.style.height = `${height}px`;
|
||||
textarea.style.overflowY = textarea.scrollHeight > 192 ? "auto" : "hidden";
|
||||
}, []);
|
||||
|
||||
// onInput does not run for a controlled value filled by a suggested question
|
||||
// or restored draft. Resize once after every render in legacy Chrome too.
|
||||
useEffect(() => {
|
||||
const textarea = textareaRef.current;
|
||||
if (!textarea) return;
|
||||
|
||||
const frame = window.requestAnimationFrame(() => {
|
||||
resizeForLegacyChrome(textarea);
|
||||
});
|
||||
return () => window.cancelAnimationFrame(frame);
|
||||
}, [controller?.textInput.value, resizeForLegacyChrome]);
|
||||
export const PromptInputTextarea = ({
|
||||
onChange,
|
||||
className,
|
||||
placeholder = "What would you like to know?",
|
||||
...props
|
||||
}: PromptInputTextareaProps) => {
|
||||
const controller = useOptionalPromptInputController();
|
||||
const attachments = usePromptInputAttachments();
|
||||
const sanitizeIncomingFiles = usePromptInputValidation();
|
||||
const [isComposing, setIsComposing] = useState(false);
|
||||
|
||||
const handleKeyDown: KeyboardEventHandler<HTMLTextAreaElement> = (e) => {
|
||||
if (e.key === "Enter") {
|
||||
@@ -997,18 +962,12 @@ export const PromptInputTextarea = ({
|
||||
};
|
||||
|
||||
return (
|
||||
<InputGroupTextarea
|
||||
className={cn("field-sizing-content max-h-48 min-h-16", className)}
|
||||
name="message"
|
||||
ref={textareaRef}
|
||||
rows={3}
|
||||
<InputGroupTextarea
|
||||
className={cn("field-sizing-content max-h-48 min-h-16", className)}
|
||||
name="message"
|
||||
onCompositionEnd={() => setIsComposing(false)}
|
||||
onCompositionStart={() => setIsComposing(true)}
|
||||
onKeyDown={handleKeyDown}
|
||||
onInput={(event) => {
|
||||
resizeForLegacyChrome(event.currentTarget);
|
||||
onInput?.(event);
|
||||
}}
|
||||
onKeyDown={handleKeyDown}
|
||||
onPaste={handlePaste}
|
||||
placeholder={placeholder}
|
||||
{...props}
|
||||
|
||||
File diff suppressed because it is too large.
Load diff
@@ -1,20 +0,0 @@
|
||||
import React from 'react';
|
||||
|
||||
const Empty = ({
|
||||
description = '暂无数据',
|
||||
size = 'default'
|
||||
}: {
|
||||
description?: string;
|
||||
size?: 'default' | 'small';
|
||||
}) => {
|
||||
return (
|
||||
<div className={`empty-container ${size === 'small' ? 'empty-small' : ''}`}>
|
||||
<div className="empty-image">
|
||||
<img src="/icon/empty-list.png" alt="暂无数据" />
|
||||
</div>
|
||||
{description && <p className="empty-description">{description}</p>}
|
||||
</div>
|
||||
);
|
||||
};
|
||||
|
||||
export default Empty;
|
||||
@@ -1,627 +0,0 @@
|
||||
"use client";
|
||||
|
||||
import {
|
||||
useEffect,
|
||||
useRef,
|
||||
useState,
|
||||
} from "react";
|
||||
import { Textarea } from "@/components/ui/textarea";
|
||||
import { Loader2, Plus, Search, Trash2 } from "lucide-react";
|
||||
import { Button } from "@/components/ui/button";
|
||||
import { toast } from "sonner";
|
||||
import { getStoredAuth } from "@/core/auth";
|
||||
import { authenticatedFetch } from "@/components/assistant-ui/document-upload-popup";
|
||||
import { readRuntimeEnv } from "@/env";
|
||||
|
||||
|
||||
interface NetworkActorProps {
|
||||
inputValue: string;
|
||||
searchSpaceId: string;
|
||||
onSuccess?: () => void;
|
||||
}
|
||||
|
||||
interface platformItem {
|
||||
title: string;
|
||||
icon: string;
|
||||
checked: boolean;
|
||||
}
|
||||
|
||||
interface languageItem {
|
||||
title: string;
|
||||
checked: boolean;
|
||||
}
|
||||
|
||||
interface PoliticsDirItem {
|
||||
地区: string;
|
||||
语言: string;
|
||||
年龄段: string;
|
||||
性别: string;
|
||||
教育程度: string;
|
||||
关注话题: { 话题: string; 倾向: string; 佐证: string; 风格: string }[];
|
||||
类型: string;
|
||||
对HZZLC: string;
|
||||
对MZZLC: string;
|
||||
对TZZLC: string[];
|
||||
账号描述: string;
|
||||
发文频率: string;
|
||||
活跃时间: string;
|
||||
}
|
||||
|
||||
interface 用户画像 {
|
||||
地区: string;
|
||||
教育程度: string;
|
||||
关注话题: { 倾向: string; 话题: string; 佐证: string; 风格: string }[];
|
||||
语言: string;
|
||||
年龄段: string;
|
||||
账号描述: string;
|
||||
对TZZLC: string[];
|
||||
对HZZLC: string;
|
||||
对MZZLC: string[];
|
||||
性别: string;
|
||||
类型: string;
|
||||
发文频率: string;
|
||||
活跃时间: string;
|
||||
}
|
||||
|
||||
interface hx_info {
|
||||
语言: string;
|
||||
用户画像: 用户画像;
|
||||
}
|
||||
|
||||
interface UserDataItem {
|
||||
comment_count: number;
|
||||
country: string | null;
|
||||
agent_task_id: string | null;
|
||||
education: string | null;
|
||||
groupId: string | null;
|
||||
forward_count: number;
|
||||
political_stand: string;
|
||||
language: string;
|
||||
type: string;
|
||||
concern: string | null;
|
||||
profile_image: string | null;
|
||||
concern_topic: string | null;
|
||||
screen_name: string;
|
||||
school: string | null;
|
||||
is_alive: number;
|
||||
createdTime: string;
|
||||
data_provider: string | null;
|
||||
id: string;
|
||||
text: string | null;
|
||||
group: string | null;
|
||||
read_count: number;
|
||||
group_name: string;
|
||||
area: string | null;
|
||||
image: string | null;
|
||||
updatedTime: string;
|
||||
like_count: number;
|
||||
m_politics_dir: string;
|
||||
camp: string | null;
|
||||
tags: string[];
|
||||
army: string | null;
|
||||
is_sync_a: number;
|
||||
tweets_count: number;
|
||||
groupName: string | null;
|
||||
history_key: string;
|
||||
user_id: string;
|
||||
real_id: string | null;
|
||||
h_politics_dir: string | null;
|
||||
domain: string | null;
|
||||
name: string | null;
|
||||
position: string | null;
|
||||
job: string | null;
|
||||
age_round: string | null;
|
||||
birthday: string | null;
|
||||
is_virtual: number;
|
||||
isMedia: number;
|
||||
hx_info: hx_info;
|
||||
gender: string | null;
|
||||
agentInfo: PoliticsDirItem;
|
||||
profile_url: string | null;
|
||||
tagId: string[];
|
||||
user_name: string;
|
||||
a_update_time: string | null;
|
||||
values: string | null;
|
||||
follows_count: number;
|
||||
title: string | null;
|
||||
is_protected: number;
|
||||
platform: string;
|
||||
importType: string;
|
||||
update_time: string;
|
||||
key: string;
|
||||
introduction: string | null;
|
||||
brief: string | null;
|
||||
creator: string;
|
||||
history_title: string;
|
||||
tweets_ratio: string | null;
|
||||
is_verified: number;
|
||||
t_politics_dir: string[];
|
||||
eventMap: string | null;
|
||||
fans_count: number;
|
||||
a_create_time: string | null;
|
||||
hobbies: string[];
|
||||
location: string | null;
|
||||
average_count: number;
|
||||
ratio: string | null;
|
||||
classification?: string | null;
|
||||
groupDiscussTags?: string[];
|
||||
checked?: boolean;
|
||||
content?: string;
|
||||
url?: string;
|
||||
}
|
||||
|
||||
export function NetworkActor({ inputValue, searchSpaceId, onSuccess }: NetworkActorProps) {
|
||||
const [platforms, setPlatforms] = useState<platformItem[]>([
|
||||
{ title: 'X', icon: 'x.png', checked: false },
|
||||
{ title: 'Facebook', icon: 'facebook.png', checked: false },
|
||||
{ title: 'Reddit', icon: 'reddit.png', checked: false },
|
||||
{ title: 'PTT', icon: 'ptt.png', checked: false },
|
||||
{ title: 'YouTube', icon: 'youtube.png', checked: false },
|
||||
{ title: 'Dcard', icon: 'dcard.png', checked: false },
|
||||
{ title: 'Mobile01', icon: 'mobile01.png', checked: false },
|
||||
{ title: 'Threads', icon: 'threads.png', checked: false },
|
||||
{ title: 'Instagram', icon: 'ins.png', checked: false },
|
||||
]);
|
||||
const platformList = useRef<string[]>([]);
|
||||
const [languages, setLanguages] = useState<languageItem[]>([
|
||||
{ title: '繁体中文', checked: false },
|
||||
{ title: '简体中文', checked: false },
|
||||
{ title: '英文', checked: false },
|
||||
{ title: '日文', checked: false },
|
||||
{ title: '其他', checked: false },
|
||||
]);
|
||||
const languageList = useRef<string[]>([]);
|
||||
const [allPlatforms, setAllPlatforms] = useState<boolean>(true);
|
||||
const [allLanguage, setAllLanguage] = useState<boolean>(true);
|
||||
const [sortType, setSortType] = useState<string>('fans_count');
|
||||
const sortTypeValue = useRef('fans_count');
|
||||
const [fansSort, setFansSort] = useState<string>('desc');
|
||||
const sortValue = useRef('desc');
|
||||
const [relevanceSort, setRelevanceSort] = useState<string>('desc');
|
||||
const [userList, setUserList] = useState<UserDataItem[]>([]);
|
||||
const [searchText, setSearchText] = useState<string>(inputValue);
|
||||
const [isSearching, setIsSearching] = useState<boolean>(false);
|
||||
const [searchType, setSearchType] = useState<string>('network');
|
||||
const searchValue = useRef<string>('network');
|
||||
const [selNetCount, setSelNetCount] = useState<number>(0);
|
||||
const [actorCount, setActorCount] = useState<number>(0);
|
||||
const [directionType, setDirectionType] = useState<string>('TW');
|
||||
const directionValue = useRef<string>('TW');
|
||||
const consumerUrl = readRuntimeEnv("VITE_AIPI_CONSUMER_URL") || 'https://ch1.b.uat.4.cn/consumer';
|
||||
const backendUrl = typeof window !== 'undefined'
|
||||
? (readRuntimeEnv("VITE_BACKEND_BASE_URL") || 'http://localhost:8001')
|
||||
: 'http://localhost:8001';
|
||||
const [selectedSearchSources, setSelectedSearchSources] = useState<UserDataItem[]>([]);
|
||||
const [isImporting, setIsImporting] = useState(false);
|
||||
|
||||
const searchChange = async () => {
|
||||
setIsSearching(true);
|
||||
if (searchValue.current === 'network') {
|
||||
await agentUserSearch();
|
||||
} else {
|
||||
await agentGroupSearch();
|
||||
}
|
||||
};
|
||||
|
||||
const setDirectionChange = async (type: string) => {
|
||||
setDirectionType(type);
|
||||
directionValue.current = type;
|
||||
await searchChange();
|
||||
};
|
||||
|
||||
const searchTypeChange = async (value: string) => {
|
||||
setSearchType(value);
|
||||
searchValue.current = value;
|
||||
await searchChange();
|
||||
};
|
||||
|
||||
const allPlatformChange = async () => {
|
||||
setAllPlatforms(true);
|
||||
setPlatforms((prev) => prev.map((item) => ({ ...item, checked: false })));
|
||||
platformList.current = [];
|
||||
await searchChange();
|
||||
};
|
||||
|
||||
const handleTogglePlatform = async (index: number) => {
|
||||
setPlatforms((prev) => {
|
||||
const newPlatforms = prev.map((item, i) =>
|
||||
i === index ? { ...item, checked: !item.checked } : item
|
||||
);
|
||||
const checkedTitles = newPlatforms.filter((item) => item.checked).map((item) => item.title);
|
||||
platformList.current = checkedTitles;
|
||||
setAllPlatforms(newPlatforms.every((item) => !item.checked));
|
||||
return newPlatforms;
|
||||
});
|
||||
await searchChange();
|
||||
};
|
||||
|
||||
const allLanguageChange = async () => {
|
||||
setAllLanguage(true);
|
||||
setLanguages((prev) => prev.map((item) => ({ ...item, checked: false })));
|
||||
languageList.current = [];
|
||||
await searchChange();
|
||||
};
|
||||
|
||||
const handleToggleLanguage = async (index: number) => {
|
||||
setLanguages((prev) => {
|
||||
const newItems = prev.map((item, i) =>
|
||||
i === index ? { ...item, checked: !item.checked } : item
|
||||
);
|
||||
const checkedTitles = newItems.filter((item) => item.checked).map((item) => item.title);
|
||||
languageList.current = checkedTitles;
|
||||
setAllLanguage(newItems.every((item) => !item.checked));
|
||||
return newItems;
|
||||
});
|
||||
await searchChange();
|
||||
};
|
||||
|
||||
const sortTypeChange = async (type: string) => {
|
||||
setSortType(type);
|
||||
sortTypeValue.current = type;
|
||||
sortValue.current = sortValue.current === 'desc' ? 'asc' : 'desc';
|
||||
if (type === 'fans_count') setFansSort(sortValue.current);
|
||||
else if (type === '_score') setRelevanceSort(sortValue.current);
|
||||
await searchChange();
|
||||
};
|
||||
|
||||
const agentUserSearch = async () => {
|
||||
try {
|
||||
const response = await authenticatedFetch(`${consumerUrl}/portrait/agentUser`, {
|
||||
method: 'POST',
|
||||
headers: { 'Content-Type': 'application/json', Authorization: 'Bearer admin' },
|
||||
body: JSON.stringify({
|
||||
historyKey: [],
|
||||
historyTitle: '',
|
||||
keyword: searchText,
|
||||
language: languageList.current,
|
||||
page: 1,
|
||||
limit: 12,
|
||||
platform: platformList.current,
|
||||
politicalStand: directionValue.current,
|
||||
sort: sortTypeValue.current,
|
||||
asc: sortValue.current,
|
||||
typeKeyword: 9,
|
||||
}),
|
||||
});
|
||||
if (response.ok) {
|
||||
const data = await response.json();
|
||||
setIsSearching(false);
|
||||
const result: UserDataItem[] = data.data.list;
|
||||
result.forEach((item) => {
|
||||
item.profile_image = item.profile_image
|
||||
? `http://minio-proxy.b.sit.4.cn/facebook-proxy?url=${item.profile_image}&is_avatar=true`
|
||||
: 'http://ty.b.uat.4.cn/manage/img/default.png';
|
||||
item.classification = `${item.political_stand}-${item.history_key}-${item.history_title}`;
|
||||
item.title = item.user_name || item.screen_name || item.user_id;
|
||||
item.content = item.agentInfo['账号描述'];
|
||||
item.url = `http://ty.b.uat.4.cn/fromB/toDetail?user_id=${item.user_id}&platform=${item.platform}`;
|
||||
});
|
||||
setUserList(result);
|
||||
setSelNetCount(data.data.total);
|
||||
setActorCount(data.data.zs_total);
|
||||
}
|
||||
} catch (error) {
|
||||
console.error('Search error:', error);
|
||||
}
|
||||
};
|
||||
|
||||
const agentGroupSearch = async () => {
|
||||
try {
|
||||
const response = await authenticatedFetch(`${consumerUrl}/portrait/agentGroup`, {
|
||||
method: 'POST',
|
||||
headers: { 'Content-Type': 'application/json', Authorization: 'Bearer admin' },
|
||||
body: JSON.stringify({
|
||||
historyKey: [],
|
||||
historyTitle: '',
|
||||
keyword: searchText,
|
||||
language: languageList.current,
|
||||
page: 1,
|
||||
limit: 12,
|
||||
platform: platformList.current,
|
||||
politicalStand: directionValue.current,
|
||||
sort: sortTypeValue.current,
|
||||
asc: sortValue.current,
|
||||
typeKeyword: 9,
|
||||
}),
|
||||
});
|
||||
if (response.ok) {
|
||||
const data = await response.json();
|
||||
setIsSearching(false);
|
||||
const result: UserDataItem[] = data.data.data;
|
||||
result.forEach((item) => {
|
||||
item.hobbies = item.groupDiscussTags || [];
|
||||
item.profile_image = item.profile_image
|
||||
? `http://minio-proxy.b.sit.4.cn/facebook-proxy?url=${item.profile_image}&is_avatar=true`
|
||||
: 'http://ty.b.uat.4.cn/manage/img/default.png';
|
||||
item.classification = `${item.political_stand}`;
|
||||
item.title = item.group_name || item.groupName;
|
||||
if (item.hx_info?.['用户画像']) {
|
||||
item.content = item.hx_info['用户画像']['账号描述'];
|
||||
}
|
||||
item.url = '';
|
||||
});
|
||||
setUserList(result);
|
||||
setActorCount(data.data.total);
|
||||
}
|
||||
} catch (error) {
|
||||
console.error('Search error:', error);
|
||||
}
|
||||
};
|
||||
|
||||
const listDataCheckedChange = (index: number, checked: boolean) => {
|
||||
setUserList((prevSources) => {
|
||||
const newSources = prevSources.map((item, aIndex) =>
|
||||
aIndex === index ? { ...item, checked } : item
|
||||
);
|
||||
setSelectedSearchSources(newSources.filter((item) => item.checked));
|
||||
return newSources;
|
||||
});
|
||||
};
|
||||
|
||||
const linkToB2 = (data: UserDataItem) => {
|
||||
const url = `http://ty.b.uat.4.cn/fromB/toDetail?user_id=${data.user_id}&platform=${data.platform}`;
|
||||
window.open(url, '_blank');
|
||||
};
|
||||
|
||||
const handleBatchDelete = () => {
|
||||
setUserList((prev) => prev.map((item) => ({ ...item, checked: false })));
|
||||
setSelectedSearchSources([]);
|
||||
};
|
||||
|
||||
const handleImport = async () => {
|
||||
if (selectedSearchSources.length === 0) {
|
||||
toast.warning('没有可导入的来源');
|
||||
return;
|
||||
}
|
||||
setIsImporting(true);
|
||||
const dataList = selectedSearchSources.map((item) => ({
|
||||
title: item.title,
|
||||
content: item.content,
|
||||
url: item.url,
|
||||
}));
|
||||
try {
|
||||
const response = await authenticatedFetch(`${backendUrl}/api/v1/source/import`, {
|
||||
method: 'POST',
|
||||
headers: { 'Content-Type': 'application/json' },
|
||||
body: JSON.stringify({
|
||||
search_space_id: Number(searchSpaceId),
|
||||
source_list: dataList,
|
||||
}),
|
||||
}, true);
|
||||
if (response.ok) {
|
||||
const data = await response.json();
|
||||
toast.success(`成功导入 ${data.imported_count || dataList.length} 个来源`);
|
||||
setSearchText('');
|
||||
setUserList([]);
|
||||
handleBatchDelete();
|
||||
onSuccess?.();
|
||||
} else {
|
||||
const errorData = await response.json();
|
||||
toast.error(errorData.detail || '导入失败，请重试');
|
||||
}
|
||||
} catch {
|
||||
toast.error('导入失败，请重试');
|
||||
} finally {
|
||||
setIsImporting(false);
|
||||
}
|
||||
};
|
||||
|
||||
useEffect(() => {
|
||||
void searchChange();
|
||||
}, []);
|
||||
|
||||
return (
|
||||
<div>
|
||||
<div className="bg-white rounded-[5px] p-[10px]">
|
||||
<div className="border border-[#e8e8e8] rounded-[5px] p-[10px]">
|
||||
<div className="mb-[10px] flex">
|
||||
<span className="text-[#3D3D3D]">所属平台：</span>
|
||||
<span className="flex-1 min-w-0">
|
||||
<span className="flex items-center gap-[10px] flex-wrap">
|
||||
<span
|
||||
className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${allPlatforms ? 'platform-active' : ''}`}
|
||||
onClick={() => void allPlatformChange()}
|
||||
>全部</span>
|
||||
{platforms.map((item, index) => (
|
||||
<span
|
||||
key={index}
|
||||
className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${item.checked ? 'platform-active' : ''}`}
|
||||
onClick={() => void handleTogglePlatform(index)}
|
||||
>
|
||||
<img src={`/platform/${item.icon}`} className="w-[24px]" alt="" />
|
||||
<span className="w-[2px] h-[12px] ml-[5px] mr-[5px] bg-[#dbdbdb]"></span>
|
||||
<span className="ml-[5px]">{item.title}</span>
|
||||
</span>
|
||||
))}
|
||||
</span>
|
||||
</span>
|
||||
</div>
|
||||
<div className="mb-[10px] flex">
|
||||
<span className="text-[#3D3D3D]">常用语言：</span>
|
||||
<span className="flex-1 min-w-0">
|
||||
<span className="flex items-center gap-[10px] flex-wrap">
|
||||
<span
|
||||
className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${allLanguage ? 'platform-active' : ''}`}
|
||||
onClick={() => void allLanguageChange()}
|
||||
>全部</span>
|
||||
{languages.map((item, index) => (
|
||||
<span
|
||||
key={index}
|
||||
className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${item.checked ? 'platform-active' : ''}`}
|
||||
onClick={() => void handleToggleLanguage(index)}
|
||||
>
|
||||
<span>{item.title}</span>
|
||||
</span>
|
||||
))}
|
||||
</span>
|
||||
</span>
|
||||
</div>
|
||||
<div className="mb-[10px] flex">
|
||||
<span className="text-[#3D3D3D]">方向：</span>
|
||||
<span className="flex-1 min-w-0">
|
||||
<span className="flex items-center gap-[10px] flex-wrap">
|
||||
<span className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${directionType === 'TW' ? 'platform-active' : ''}`} onClick={() => void setDirectionChange('TW')}>W方向</span>
|
||||
<span className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${directionType === 'US' ? 'platform-active' : ''}`} onClick={() => void setDirectionChange('US')}>U方向</span>
|
||||
<span className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${directionType === '其他' ? 'platform-active' : ''}`} onClick={() => void setDirectionChange('其他')}>其他</span>
|
||||
</span>
|
||||
</span>
|
||||
</div>
|
||||
<div className="mb-[10px] flex">
|
||||
<span className="text-[#3D3D3D]">类型：</span>
|
||||
<span className="flex-1 min-w-0">
|
||||
<span className="flex items-center gap-[10px] flex-wrap">
|
||||
<span className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${searchType === 'network' ? 'platform-active' : ''}`} onClick={() => void searchTypeChange('network')}>网络XWT</span>
|
||||
<span className={`flex items-center cursor-pointer border border-[transparent] pl-[10px] pr-[10px] ${searchType === 'group' ? 'platform-active' : ''}`} onClick={() => void searchTypeChange('group')}>群组XWT</span>
|
||||
</span>
|
||||
</span>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
<div className="bg-white rounded-[5px] pt-[15px] mt-[10px]">
|
||||
<div className="flex items-center justify-between pb-[10px] mb-[10px] border-b border-b-[#f1f3f5]">
|
||||
<div className="flex items-center ml-[10px]">
|
||||
<div className="title-bar mr-[8px]"></div>
|
||||
<div className="text-[#454859]">用户表格管理</div>
|
||||
</div>
|
||||
<div className="flex items-center">
|
||||
<div
|
||||
className={`relative flex items-center cursor-pointer border ${sortType === 'fans_count' ? 'border-[#cf1322]' : ''} p-[7px] pl-[10px] pr-[20px] rounded-[5px]`}
|
||||
onClick={() => void sortTypeChange('fans_count')}
|
||||
>
|
||||
<span className="leading-[16px]">粉丝数排序</span>
|
||||
<div className="ml-[10px]">
|
||||
<div className={`sort-up ${fansSort === 'asc' ? 'sort-up-active' : ''}`}></div>
|
||||
<div className={`mt-[4px] sort-down ${fansSort === 'desc' ? 'sort-down-active' : ''}`}></div>
|
||||
</div>
|
||||
</div>
|
||||
<div
|
||||
className={`relative flex items-center ml-[10px] cursor-pointer border ${sortType === '_score' ? 'border-[#cf1322]' : ''} p-[7px] pl-[10px] pr-[20px] rounded-[5px]`}
|
||||
onClick={() => void sortTypeChange('_score')}
|
||||
>
|
||||
<span className="leading-[16px]">相关度排序</span>
|
||||
<div className="ml-[10px]">
|
||||
<div className={`sort-up ${relevanceSort === 'asc' ? 'sort-up-active' : ''}`}></div>
|
||||
<div className={`mt-[4px] sort-down ${relevanceSort === 'desc' ? 'sort-down-active' : ''}`}></div>
|
||||
</div>
|
||||
</div>
|
||||
{searchType === 'network' && (
|
||||
<>
|
||||
<div className="flex items-center ml-[35px]">
|
||||
<img className="w-[24px] h-[24px]" src="/icon/group-yellow-logo.svg" alt="" />
|
||||
<div className="ml-[5px]">
|
||||
<div className="text-[#454859] leading-[16px]">当前选中网络XWT数</div>
|
||||
<div className="text-[#d97b00] leading-[16px] mt-[5px]">{selNetCount}</div>
|
||||
</div>
|
||||
</div>
|
||||
<div className="ml-[5px] mr-[5px] h-[30px] w-[2px] -skew-x-[30deg] bg-[#dbdbdb]"></div>
|
||||
</>
|
||||
)}
|
||||
<div className={`flex items-center ${searchType === 'network' ? 'ml-[5px]' : 'ml-[20px]'} mr-[25px]`}>
|
||||
<img className="w-[24px] h-[24px]" src="/icon/group-red-logo.svg" alt="" />
|
||||
<div className="ml-[5px]">
|
||||
<div className="text-[#454859] leading-[16px]">网络XWT总数</div>
|
||||
<div className="text-[#cf1322] leading-[16px] mt-[5px]">{actorCount}</div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
{selectedSearchSources.length > 0 && (
|
||||
<div className="flex items-center justify-between px-4 py-2 bg-gray-50 border-b border-gray-100">
|
||||
<span className="text-sm text-gray-600">已选中 {selectedSearchSources.length} 项</span>
|
||||
<div className="flex gap-2">
|
||||
<Button size="sm" variant="outline" onClick={handleBatchDelete} className="h-8 px-3 text-sm">
|
||||
<Trash2 className="h-3.5 w-3.5 mr-1" />
|
||||
删除选中
|
||||
</Button>
|
||||
<Button
|
||||
size="sm"
|
||||
onClick={() => void handleImport()}
|
||||
disabled={isImporting}
|
||||
className="h-8 px-3 text-sm bg-blue-50 hover:bg-blue-100 text-blue-600 border border-blue-200"
|
||||
>
|
||||
{isImporting ? <Loader2 className="h-3.5 w-3.5 mr-1 animate-spin" /> : <Plus className="h-3.5 w-3.5 mr-1" />}
|
||||
导入选中
|
||||
</Button>
|
||||
</div>
|
||||
</div>
|
||||
)}
|
||||
<div className="p-[10px]">
|
||||
{isSearching ? (
|
||||
<div className="flex items-center justify-center py-16">
|
||||
<Loader2 className="h-8 w-8 animate-spin text-muted-foreground" />
|
||||
</div>
|
||||
) : null}
|
||||
<div className={`grid grid-cols-4 gap-[10px] user-list${isSearching ? ' hidden' : ''}`}>
|
||||
{userList.map((item, index) => (
|
||||
<div key={index} className="user-item p-[10px] relative">
|
||||
<div className="absolute top-[5px] right-[10px]">
|
||||
<input
|
||||
type="checkbox"
|
||||
checked={item.checked ?? false}
|
||||
onChange={() => listDataCheckedChange(index, !item.checked)}
|
||||
className="w-4 h-4 cursor-pointer accent-primary"
|
||||
/>
|
||||
</div>
|
||||
<div className="flex items-center mb-[10px] cursor-pointer" onClick={() => linkToB2(item)}>
|
||||
<img src={item.profile_image ?? ''} className="w-[97px] h-[97px] rounded-[5px]" alt="" />
|
||||
<div className="flex-1 min-w-0 pl-[10px]">
|
||||
<div className="truncate mb-[10px] text-[18px]">{item.user_name || item.screen_name || item.user_id || item.group_name || item.groupName}</div>
|
||||
<div className="flex justify-between mb-[10px]">
|
||||
<div className="flex items-center">
|
||||
<img src="/icon/group-yellow-logo.svg" className="w-[16px] h-[16px] mr-[5px]" alt="" />
|
||||
<span>粉丝数：{item.fans_count || '-'}</span>
|
||||
</div>
|
||||
<div className="flex items-center">
|
||||
<img src={`/platform/${item.platform}.png`} className="w-[25px] mr-[5px]" alt="" />
|
||||
<span>{item.platform}</span>
|
||||
</div>
|
||||
</div>
|
||||
<div className="flex items-center">
|
||||
<img src="/icon/person-red-logo.svg" className="w-[16px] mr-[5px]" alt="" />
|
||||
<span>实体人:{item.name}</span>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
<div className="cursor-pointer" onClick={() => linkToB2(item)}>
|
||||
<div className="flex items-center mb-[10px]">
|
||||
<div className="w-[70px] flex items-center">
|
||||
<img src="/icon/classification-logo.png" className="w-[16px] mr-[5px]" alt="" />
|
||||
<span>分类：</span>
|
||||
</div>
|
||||
<div className="flex-1 min-w-0">
|
||||
<div className="inline-block h-[28px] pl-[8px] pr-[8px] bg-[#fef6f6] text-[#cf1322]">{item.classification}</div>
|
||||
</div>
|
||||
</div>
|
||||
<div className="flex items-center mb-[10px]">
|
||||
<div className="w-[70px] flex items-center">
|
||||
<img src="/icon/user-green-logo.png" className="w-[16px] mr-[5px]" alt="" />
|
||||
<span>标签：</span>
|
||||
</div>
|
||||
<div className="flex-1 min-w-0">
|
||||
<div className="flex flex-wrap items-center gap-[5px]">
|
||||
{(item.hobbies ?? []).map((hobbie) => (
|
||||
<div key={hobbie} className="h-[28px] pl-[8px] pr-[8px] bg-[#f6fdfb] text-[#1eb07f]">{hobbie}</div>
|
||||
))}
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
<div className="flex items-baseline mb-[10px]">
|
||||
<div className="w-[70px] flex items-center">
|
||||
<img src="/icon/user-green-logo.png" className="w-[16px] mr-[5px]" alt="" />
|
||||
<span>简介：</span>
|
||||
</div>
|
||||
<div className="flex-1 min-w-0">
|
||||
<div className="line-clamp-3 text-ellipsis overflow-hidden">
|
||||
{(item.agentInfo && item.agentInfo['账号描述']) || (item.hx_info?.['用户画像']?.['账号描述']) || ''}
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
))}
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
</div>
|
||||
);
|
||||
}
|
||||
@@ -1,27 +0,0 @@
|
||||
import { useEffect } from "react";
|
||||
|
||||
import { replaceDesensitizeMap } from "@/lib/desensitize";
|
||||
import { useSensitiveWords, wordsToMap } from "@/strategy-components/api/sensitive-words";
|
||||
|
||||
/**
|
||||
* 启动时把后台「敏感词管理」配置加载进运行时脱敏映射表。
|
||||
*
|
||||
* - 后台返回**非空**时：用其中「启用」的词条整体替换运行时映射
|
||||
* （{@link replaceDesensitizeMap}），后台即唯一真源，被删除的词不再生效。
|
||||
* - 后台为空（全新部署还没配置）或拉取失败：什么都不做，保持使用
|
||||
* `src/lib/desensitize.ts` 内置的种子映射。
|
||||
*
|
||||
* 渲染 `null`；挂在 `AppProviders` 里（QueryClientProvider 之内）。与「敏感词管理」
|
||||
* 页共享同一份 React Query 缓存，保存后会自动刷新并回灌运行时映射。
|
||||
*/
|
||||
export function DesensitizeWordsLoader() {
|
||||
const { data } = useSensitiveWords();
|
||||
|
||||
useEffect(() => {
|
||||
if (data && data.length > 0) {
|
||||
replaceDesensitizeMap(wordsToMap(data));
|
||||
}
|
||||
}, [data]);
|
||||
|
||||
return null;
|
||||
}
|
||||
@@ -1,45 +0,0 @@
|
||||
import { useEffect, useState } from "react";
|
||||
|
||||
import { injectInPageAnchorHandler } from "@/core/utils/markdown";
|
||||
|
||||
/**
|
||||
* Sandboxed preview of a generated HTML document.
|
||||
*
|
||||
* Renders via a `blob:` URL rather than `srcDoc`. A srcDoc document's base URL
|
||||
* is `about:srcdoc`, so clicking an in-page anchor (`<a href="#section">`)
|
||||
* navigates the frame to `about:blank#section` and BLANKS it. A `blob:` URL is
|
||||
* a real document URL, so fragment links scroll within the frame as expected —
|
||||
* all without enabling scripts (the sandbox stays script-free for security; the
|
||||
* server already sanitizes script vectors and this is defense-in-depth).
|
||||
*
|
||||
* The `charset=utf-8` on the blob type is required: a blob: URL whose
|
||||
* Content-Type omits the charset is decoded with the OS default (GBK/Latin-1 on
|
||||
* Chinese Windows), garbling CJK text when the HTML lacks a `<meta charset>`.
|
||||
*/
|
||||
export function HtmlPreviewFrame({
|
||||
html,
|
||||
className = "h-[70vh] w-full rounded-md border bg-white",
|
||||
}: {
|
||||
html: string;
|
||||
className?: string;
|
||||
}) {
|
||||
const [url, setUrl] = useState<string>();
|
||||
|
||||
useEffect(() => {
|
||||
const blob = new Blob([injectInPageAnchorHandler(html ?? "")], {
|
||||
type: "text/html;charset=utf-8",
|
||||
});
|
||||
const objectUrl = URL.createObjectURL(blob);
|
||||
setUrl(objectUrl);
|
||||
return () => URL.revokeObjectURL(objectUrl);
|
||||
}, [html]);
|
||||
|
||||
return (
|
||||
<iframe
|
||||
title="HTML 预览"
|
||||
sandbox="allow-scripts allow-forms"
|
||||
src={url}
|
||||
className={className}
|
||||
/>
|
||||
);
|
||||
}
|
||||
@@ -1,956 +0,0 @@
|
||||
import {
|
||||
forwardRef,
|
||||
useCallback,
|
||||
useEffect,
|
||||
useImperativeHandle,
|
||||
useMemo,
|
||||
useRef,
|
||||
type ReactNode,
|
||||
} from "react";
|
||||
|
||||
import type {
|
||||
GraphEdge,
|
||||
GraphNode,
|
||||
LayoutType,
|
||||
SimEdge,
|
||||
SimNode,
|
||||
TooltipState,
|
||||
Transform,
|
||||
} from "./graphTypes";
|
||||
import {
|
||||
applyCircularLayout,
|
||||
ForceSimulation,
|
||||
getEdgeId,
|
||||
getEdgeInfo,
|
||||
getEdgeLabel,
|
||||
getLinkDistance,
|
||||
getNodeKind,
|
||||
getNodeRadius,
|
||||
getNodeTitle,
|
||||
NODE_KIND_META,
|
||||
seedPositions,
|
||||
truncateLabel,
|
||||
} from "./graphUtils";
|
||||
|
||||
const MIN_ZOOM = 0.15;
|
||||
const MAX_ZOOM = 4;
|
||||
|
||||
/** Canvas-drawn colors per app theme (CSS variables don't reach the 2D context). */
|
||||
const CANVAS_PALETTE = {
|
||||
dark: {
|
||||
exportBg: "#0b1020",
|
||||
labelHalo: "rgba(2, 6, 23, 0.9)",
|
||||
nodeRim: "rgba(2, 6, 23, 0.55)",
|
||||
rootRim: "#e0f2fe",
|
||||
expandedRim: "#7dd3fc",
|
||||
rootRing: "rgba(125, 211, 252, 0.9)",
|
||||
expandedDot: "#7dd3fc",
|
||||
expandedDotRim: "rgba(2, 6, 23, 0.8)",
|
||||
selectionRing: "#ffffff",
|
||||
hoverRing: "rgba(255, 255, 255, 0.85)",
|
||||
pin: "rgba(255, 255, 255, 0.8)",
|
||||
labelActive: "#ffffff",
|
||||
label: "#e2e8f0",
|
||||
minimapViewport: "rgba(56, 189, 248, 0.8)",
|
||||
},
|
||||
light: {
|
||||
exportBg: "#f8fafc",
|
||||
labelHalo: "rgba(255, 255, 255, 0.92)",
|
||||
nodeRim: "rgba(255, 255, 255, 0.9)",
|
||||
rootRim: "#075985",
|
||||
expandedRim: "#0284c7",
|
||||
rootRing: "rgba(2, 132, 199, 0.85)",
|
||||
expandedDot: "#0284c7",
|
||||
expandedDotRim: "rgba(255, 255, 255, 0.9)",
|
||||
selectionRing: "#0f172a",
|
||||
hoverRing: "rgba(15, 23, 42, 0.7)",
|
||||
pin: "rgba(15, 23, 42, 0.6)",
|
||||
labelActive: "#0f172a",
|
||||
label: "#334155",
|
||||
minimapViewport: "rgba(2, 132, 199, 0.8)",
|
||||
},
|
||||
} as const;
|
||||
|
||||
export interface GraphCanvasHandle {
|
||||
zoomIn: () => void;
|
||||
zoomOut: () => void;
|
||||
fitView: () => void;
|
||||
relayout: () => void;
|
||||
focusNode: (id: string) => void;
|
||||
exportImage: () => void;
|
||||
}
|
||||
|
||||
interface GraphCanvasProps {
|
||||
nodes: GraphNode[];
|
||||
edges: GraphEdge[];
|
||||
/** Degrees computed on the FULL graph so node sizes stay stable under filters. */
|
||||
degrees: Map<string, number>;
|
||||
selectedNodeId: string | null;
|
||||
/** null = no active search; otherwise the set of matching node ids. */
|
||||
searchMatchIds: Set<string> | null;
|
||||
layoutType: LayoutType;
|
||||
showLabels: boolean;
|
||||
/** App theme; drives the canvas-drawn colors (labels, rims, rings). */
|
||||
theme: "light" | "dark";
|
||||
onSelectNode: (id: string | null) => void;
|
||||
onTooltip: (tooltip: TooltipState | null) => void;
|
||||
onZoomChange: (k: number) => void;
|
||||
/**
|
||||
* Explore-mode rendering context; null/undefined renders the normal full view.
|
||||
* preview = no root picked yet: nodes are shown faded as pickable起点 candidates.
|
||||
*/
|
||||
explore?: {
|
||||
preview: boolean;
|
||||
rootId: string | null;
|
||||
nodeDepthMap: Map<string, number>;
|
||||
expandedNodeIds: Set<string>;
|
||||
} | null;
|
||||
/** Floating pills rendered top-left over the canvas. */
|
||||
toolbar?: ReactNode;
|
||||
/** Small floating hint rendered bottom-center over the canvas. */
|
||||
hint?: ReactNode;
|
||||
/** Centered overlay (empty / loading states). */
|
||||
overlay?: ReactNode;
|
||||
}
|
||||
|
||||
interface DragState {
|
||||
pointerId: number;
|
||||
node: SimNode | null;
|
||||
startX: number;
|
||||
startY: number;
|
||||
originX: number;
|
||||
originY: number;
|
||||
moved: boolean;
|
||||
}
|
||||
|
||||
const easeCubicOut = (t: number) => 1 - Math.pow(1 - t, 3);
|
||||
|
||||
export const GraphCanvas = forwardRef<GraphCanvasHandle, GraphCanvasProps>(
|
||||
function GraphCanvas(props, ref) {
|
||||
const containerRef = useRef<HTMLDivElement>(null);
|
||||
const canvasRef = useRef<HTMLCanvasElement>(null);
|
||||
const minimapRef = useRef<HTMLCanvasElement>(null);
|
||||
|
||||
const simRef = useRef<ForceSimulation | null>(null);
|
||||
const nodeMapRef = useRef<Map<string, SimNode>>(new Map());
|
||||
const adjacencyRef = useRef<Map<string, Set<string>>>(new Map());
|
||||
const labelThresholdRef = useRef(3);
|
||||
|
||||
const transformRef = useRef<Transform>({ x: 0, y: 0, k: 0 });
|
||||
const sizeRef = useRef({ w: 0, h: 0 });
|
||||
const hoverRef = useRef<SimNode | null>(null);
|
||||
const dragRef = useRef<DragState | null>(null);
|
||||
const animRef = useRef<{
|
||||
from: Transform;
|
||||
to: Transform;
|
||||
start: number;
|
||||
duration: number;
|
||||
} | null>(null);
|
||||
const needsDrawRef = useRef(true);
|
||||
const autoFitRef = useRef(false);
|
||||
// While the initial layout is still settling we keep the whole graph in
|
||||
// view every frame; any pan/zoom/drag from the user takes over.
|
||||
const interactedRef = useRef(false);
|
||||
const rafRef = useRef(0);
|
||||
|
||||
// Latest props readable from the rAF loop / event handlers.
|
||||
const propsRef = useRef(props);
|
||||
propsRef.current = props;
|
||||
|
||||
const requestDraw = useCallback(() => {
|
||||
needsDrawRef.current = true;
|
||||
}, []);
|
||||
|
||||
// ---------------------------------------------------------------- build
|
||||
|
||||
useEffect(() => {
|
||||
const previous = nodeMapRef.current;
|
||||
const simNodes: SimNode[] = props.nodes.map((n) => {
|
||||
const degree = props.degrees.get(n.id) ?? 0;
|
||||
const kind = getNodeKind(n);
|
||||
const old = previous.get(n.id);
|
||||
return {
|
||||
id: n.id,
|
||||
node: n,
|
||||
kind,
|
||||
degree,
|
||||
radius: getNodeRadius(degree),
|
||||
color: NODE_KIND_META[kind].color,
|
||||
x: old?.x ?? Number.NaN,
|
||||
y: old?.y ?? Number.NaN,
|
||||
vx: 0,
|
||||
vy: 0,
|
||||
fx: old?.fx ?? null,
|
||||
fy: old?.fy ?? null,
|
||||
};
|
||||
});
|
||||
|
||||
const fresh = simNodes.filter((n) => Number.isNaN(n.x));
|
||||
if (fresh.length === simNodes.length) {
|
||||
seedPositions(simNodes);
|
||||
} else {
|
||||
// Filters re-enabled some nodes: spiral the newcomers in near the center.
|
||||
fresh.forEach((n, i) => {
|
||||
const r = 60 + 16 * Math.sqrt(i + 1);
|
||||
const angle = i * 2.3999632297;
|
||||
n.x = r * Math.cos(angle);
|
||||
n.y = r * Math.sin(angle);
|
||||
});
|
||||
}
|
||||
|
||||
const byId = new Map(simNodes.map((n) => [n.id, n]));
|
||||
const simEdges: SimEdge[] = [];
|
||||
const adjacency = new Map<string, Set<string>>();
|
||||
for (const e of props.edges) {
|
||||
const source = byId.get(e.source);
|
||||
const target = byId.get(e.target);
|
||||
if (!source || !target) continue;
|
||||
const label = getEdgeLabel(e);
|
||||
simEdges.push({
|
||||
id: getEdgeId(e),
|
||||
source,
|
||||
target,
|
||||
label,
|
||||
info: getEdgeInfo(label),
|
||||
weight: e.weight ?? 1,
|
||||
distance: getLinkDistance(label) + source.radius + target.radius,
|
||||
});
|
||||
if (!adjacency.has(source.id)) adjacency.set(source.id, new Set());
|
||||
if (!adjacency.has(target.id)) adjacency.set(target.id, new Set());
|
||||
adjacency.get(source.id)!.add(target.id);
|
||||
adjacency.get(target.id)!.add(source.id);
|
||||
}
|
||||
|
||||
// Label budget: always label hubs; ~20 labels max by default.
|
||||
const sortedDegrees = simNodes.map((n) => n.degree).sort((a, b) => b - a);
|
||||
labelThresholdRef.current = Math.max(
|
||||
2,
|
||||
sortedDegrees[Math.min(19, sortedDegrees.length - 1)] ?? 2,
|
||||
);
|
||||
|
||||
nodeMapRef.current = byId;
|
||||
adjacencyRef.current = adjacency;
|
||||
const sim = new ForceSimulation(simNodes, simEdges);
|
||||
if (props.layoutType === "circular") {
|
||||
applyCircularLayout(simNodes);
|
||||
sim.alpha = 0;
|
||||
} else {
|
||||
sim.alpha = previous.size === 0 ? 1 : 0.5;
|
||||
}
|
||||
simRef.current = sim;
|
||||
if (previous.size === 0) {
|
||||
autoFitRef.current = false;
|
||||
interactedRef.current = false;
|
||||
// Pre-settle a fresh force layout synchronously (time-bounded) so the
|
||||
// first paint is already roughly laid out instead of an exploding
|
||||
// cloud that drifts out of the viewport.
|
||||
if (props.layoutType === "force") {
|
||||
const budget = simNodes.length > 350 ? 280 : 120;
|
||||
const deadline = performance.now() + budget;
|
||||
let guard = 0;
|
||||
while (sim.alpha > 0.25 && guard < 600 && performance.now() < deadline) {
|
||||
sim.tick();
|
||||
guard += 1;
|
||||
}
|
||||
}
|
||||
}
|
||||
requestDraw();
|
||||
// eslint-disable-next-line react-hooks/exhaustive-deps
|
||||
}, [props.nodes, props.edges, props.degrees]);
|
||||
|
||||
// Layout switching.
|
||||
const layoutRef = useRef(props.layoutType);
|
||||
useEffect(() => {
|
||||
if (layoutRef.current === props.layoutType) return;
|
||||
layoutRef.current = props.layoutType;
|
||||
const sim = simRef.current;
|
||||
if (!sim) return;
|
||||
if (props.layoutType === "circular") {
|
||||
applyCircularLayout(sim.nodes);
|
||||
sim.alpha = 0;
|
||||
} else {
|
||||
for (const n of sim.nodes) {
|
||||
n.fx = null;
|
||||
n.fy = null;
|
||||
}
|
||||
sim.reheat(0.9);
|
||||
}
|
||||
autoFitRef.current = false;
|
||||
interactedRef.current = false;
|
||||
requestDraw();
|
||||
}, [props.layoutType, requestDraw]);
|
||||
|
||||
// Redraw on highlight-affecting prop changes.
|
||||
useEffect(() => {
|
||||
requestDraw();
|
||||
}, [
|
||||
props.selectedNodeId,
|
||||
props.searchMatchIds,
|
||||
props.showLabels,
|
||||
props.explore,
|
||||
props.theme,
|
||||
requestDraw,
|
||||
]);
|
||||
|
||||
// ------------------------------------------------------------ transform
|
||||
|
||||
const clampZoom = (k: number) => Math.min(MAX_ZOOM, Math.max(MIN_ZOOM, k));
|
||||
|
||||
// Notify React of zoom changes only when k actually moved — a state
|
||||
// update per animation frame re-renders the whole page while settling.
|
||||
const lastNotifiedZoomRef = useRef(0);
|
||||
const setTransform = useCallback((t: Transform) => {
|
||||
transformRef.current = t;
|
||||
if (Math.abs(t.k - lastNotifiedZoomRef.current) > 0.004) {
|
||||
lastNotifiedZoomRef.current = t.k;
|
||||
propsRef.current.onZoomChange(t.k);
|
||||
}
|
||||
needsDrawRef.current = true;
|
||||
}, []);
|
||||
|
||||
const animateTo = useCallback((to: Transform, duration = 350) => {
|
||||
animRef.current = {
|
||||
from: { ...transformRef.current },
|
||||
to,
|
||||
start: performance.now(),
|
||||
duration,
|
||||
};
|
||||
needsDrawRef.current = true;
|
||||
}, []);
|
||||
|
||||
const zoomBy = useCallback(
|
||||
(factor: number, cx?: number, cy?: number) => {
|
||||
interactedRef.current = true;
|
||||
const t = transformRef.current;
|
||||
const { w, h } = sizeRef.current;
|
||||
const px = cx ?? w / 2;
|
||||
const py = cy ?? h / 2;
|
||||
const k = clampZoom(t.k * factor);
|
||||
const wx = (px - t.x) / t.k;
|
||||
const wy = (py - t.y) / t.k;
|
||||
setTransform({ x: px - wx * k, y: py - wy * k, k });
|
||||
},
|
||||
[setTransform],
|
||||
);
|
||||
|
||||
const computeFitTransform = useCallback((): Transform | null => {
|
||||
const sim = simRef.current;
|
||||
const { w, h } = sizeRef.current;
|
||||
if (!sim || !sim.nodes.length || !w || !h) return null;
|
||||
let minX = Infinity;
|
||||
let minY = Infinity;
|
||||
let maxX = -Infinity;
|
||||
let maxY = -Infinity;
|
||||
for (const n of sim.nodes) {
|
||||
minX = Math.min(minX, n.x - n.radius);
|
||||
minY = Math.min(minY, n.y - n.radius);
|
||||
maxX = Math.max(maxX, n.x + n.radius);
|
||||
maxY = Math.max(maxY, n.y + n.radius);
|
||||
}
|
||||
// A single bad coordinate must not poison the viewport transform.
|
||||
if (!Number.isFinite(minX + minY + maxX + maxY)) return null;
|
||||
const pad = 70;
|
||||
const bw = Math.max(maxX - minX, 1);
|
||||
const bh = Math.max(maxY - minY, 1);
|
||||
const k = clampZoom(Math.min((w - pad * 2) / bw, (h - pad * 2) / bh, 2.2));
|
||||
const cx = (minX + maxX) / 2;
|
||||
const cy = (minY + maxY) / 2;
|
||||
return { x: w / 2 - cx * k, y: h / 2 - cy * k, k };
|
||||
}, []);
|
||||
|
||||
const fitView = useCallback(
|
||||
(animated = true) => {
|
||||
const target = computeFitTransform();
|
||||
if (!target) return;
|
||||
if (animated) animateTo(target);
|
||||
else setTransform(target);
|
||||
},
|
||||
[animateTo, computeFitTransform, setTransform],
|
||||
);
|
||||
|
||||
const focusNode = useCallback(
|
||||
(id: string) => {
|
||||
const node = nodeMapRef.current.get(id);
|
||||
const { w, h } = sizeRef.current;
|
||||
if (!node || !w || !h) return;
|
||||
interactedRef.current = true;
|
||||
const k = clampZoom(Math.max(transformRef.current.k, 1.4));
|
||||
animateTo({ x: w / 2 - node.x * k, y: h / 2 - node.y * k, k });
|
||||
},
|
||||
[animateTo],
|
||||
);
|
||||
|
||||
const relayout = useCallback(() => {
|
||||
const sim = simRef.current;
|
||||
if (!sim) return;
|
||||
if (layoutRef.current === "circular") {
|
||||
applyCircularLayout(sim.nodes);
|
||||
sim.alpha = 0;
|
||||
fitView(true);
|
||||
} else {
|
||||
seedPositions(sim.nodes);
|
||||
sim.alpha = 1;
|
||||
autoFitRef.current = false;
|
||||
interactedRef.current = false;
|
||||
}
|
||||
requestDraw();
|
||||
}, [fitView, requestDraw]);
|
||||
|
||||
const exportImage = useCallback(() => {
|
||||
const main = canvasRef.current;
|
||||
if (!main) return;
|
||||
const out = document.createElement("canvas");
|
||||
out.width = main.width;
|
||||
out.height = main.height;
|
||||
const ctx = out.getContext("2d");
|
||||
if (!ctx) return;
|
||||
ctx.fillStyle = CANVAS_PALETTE[propsRef.current.theme].exportBg;
|
||||
ctx.fillRect(0, 0, out.width, out.height);
|
||||
ctx.drawImage(main, 0, 0);
|
||||
out.toBlob((blob) => {
|
||||
if (!blob) return;
|
||||
const url = URL.createObjectURL(blob);
|
||||
const a = document.createElement("a");
|
||||
a.href = url;
|
||||
a.download = "knowledge-graph.png";
|
||||
a.click();
|
||||
URL.revokeObjectURL(url);
|
||||
}, "image/png");
|
||||
}, []);
|
||||
|
||||
useImperativeHandle(
|
||||
ref,
|
||||
() => ({
|
||||
zoomIn: () => zoomBy(1.25),
|
||||
zoomOut: () => zoomBy(1 / 1.25),
|
||||
fitView: () => fitView(true),
|
||||
relayout,
|
||||
focusNode,
|
||||
exportImage,
|
||||
}),
|
||||
[zoomBy, fitView, relayout, focusNode, exportImage],
|
||||
);
|
||||
|
||||
// ----------------------------------------------------------------- draw
|
||||
|
||||
const draw = useCallback(() => {
|
||||
const canvas = canvasRef.current;
|
||||
const sim = simRef.current;
|
||||
if (!canvas || !sim) return;
|
||||
const ctx = canvas.getContext("2d");
|
||||
if (!ctx) return;
|
||||
const dpr = window.devicePixelRatio || 1;
|
||||
const { w, h } = sizeRef.current;
|
||||
const t = transformRef.current;
|
||||
const p = propsRef.current;
|
||||
|
||||
ctx.setTransform(dpr, 0, 0, dpr, 0, 0);
|
||||
ctx.clearRect(0, 0, w, h);
|
||||
ctx.translate(t.x, t.y);
|
||||
ctx.scale(t.k, t.k);
|
||||
|
||||
const hover = hoverRef.current;
|
||||
const activeId = hover?.id ?? p.selectedNodeId;
|
||||
const neighbors = activeId ? adjacencyRef.current.get(activeId) : undefined;
|
||||
const search = p.searchMatchIds;
|
||||
const k = t.k || 1;
|
||||
const explore = p.explore;
|
||||
const palette = CANVAS_PALETTE[p.theme];
|
||||
|
||||
// ---- edges
|
||||
for (const e of sim.edges) {
|
||||
const touchesActive =
|
||||
activeId !== null &&
|
||||
(e.source.id === activeId || e.target.id === activeId);
|
||||
let alpha: number;
|
||||
if (activeId) alpha = touchesActive ? 0.95 : 0.07;
|
||||
else if (search)
|
||||
alpha = search.has(e.source.id) || search.has(e.target.id) ? 0.75 : 0.06;
|
||||
else alpha = 0.38;
|
||||
|
||||
if (explore && !explore.preview) {
|
||||
const sourceDepth = explore.nodeDepthMap.get(e.source.id) ?? 0;
|
||||
const targetDepth = explore.nodeDepthMap.get(e.target.id) ?? 0;
|
||||
const edgeDepth = Math.max(sourceDepth, targetDepth);
|
||||
alpha *= Math.max(0.48, 1 - edgeDepth * 0.12);
|
||||
}
|
||||
|
||||
ctx.globalAlpha = alpha;
|
||||
ctx.strokeStyle = e.info.color;
|
||||
ctx.lineWidth = (touchesActive ? 2.4 : 1.2) / k;
|
||||
ctx.setLineDash(e.info.dashed ? [6 / k, 5 / k] : []);
|
||||
ctx.beginPath();
|
||||
ctx.moveTo(e.source.x, e.source.y);
|
||||
ctx.lineTo(e.target.x, e.target.y);
|
||||
ctx.stroke();
|
||||
|
||||
// Relation label at the midpoint while a node is focused.
|
||||
if (touchesActive && k >= 0.55) {
|
||||
const mx = (e.source.x + e.target.x) / 2;
|
||||
const my = (e.source.y + e.target.y) / 2;
|
||||
ctx.setLineDash([]);
|
||||
ctx.font = `${10.5 / k}px "Noto Sans SC", system-ui, sans-serif`;
|
||||
ctx.textAlign = "center";
|
||||
ctx.textBaseline = "middle";
|
||||
ctx.lineWidth = 3 / k;
|
||||
ctx.strokeStyle = palette.labelHalo;
|
||||
ctx.strokeText(e.info.zh, mx, my);
|
||||
ctx.fillStyle = e.info.color;
|
||||
ctx.fillText(e.info.zh, mx, my);
|
||||
}
|
||||
}
|
||||
ctx.setLineDash([]);
|
||||
|
||||
// ---- nodes
|
||||
const labelThreshold = labelThresholdRef.current;
|
||||
for (const n of sim.nodes) {
|
||||
const isActive = n.id === activeId;
|
||||
const isSelected = n.id === p.selectedNodeId;
|
||||
const isNeighbor = neighbors?.has(n.id) ?? false;
|
||||
const depth = explore?.nodeDepthMap.get(n.id);
|
||||
const isExploreRoot = explore?.rootId === n.id;
|
||||
const isExpanded = explore?.expandedNodeIds.has(n.id) ?? false;
|
||||
let alpha: number;
|
||||
if (activeId) alpha = isActive || isNeighbor ? 1 : 0.14;
|
||||
else if (search) alpha = search.has(n.id) ? 1 : 0.14;
|
||||
else alpha = 1;
|
||||
|
||||
let baseRadius = n.radius;
|
||||
let labelByExplore = false;
|
||||
let strokeWidth = 1.4 / k;
|
||||
let strokeColor: string = palette.nodeRim;
|
||||
let fillColor = n.color;
|
||||
if (explore?.preview) {
|
||||
alpha *= 0.42;
|
||||
baseRadius = Math.max(8, n.radius * 0.86);
|
||||
labelByExplore = isActive || isSelected;
|
||||
} else if (explore && depth != null) {
|
||||
baseRadius = depth === 0 ? 34 : Math.max(14, 28 - depth * 4);
|
||||
alpha *= Math.max(0.45, 1 - depth * 0.12);
|
||||
labelByExplore = depth <= 2 || isActive || isSelected;
|
||||
strokeWidth = (depth === 0 ? 4 : isExpanded ? 3 : 1.5) / k;
|
||||
strokeColor =
|
||||
depth === 0
|
||||
? palette.rootRim
|
||||
: isExpanded
|
||||
? palette.expandedRim
|
||||
: palette.nodeRim;
|
||||
if (isExploreRoot) fillColor = "#38bdf8";
|
||||
}
|
||||
|
||||
const r = baseRadius * (n.id === hover?.id ? 1.15 : 1);
|
||||
ctx.globalAlpha = alpha;
|
||||
|
||||
// Glow for hubs and the active node. Canvas shadows are expensive,
|
||||
// so on large graphs only the focused nodes keep theirs.
|
||||
const glowEligible =
|
||||
sim.nodes.length <= 450
|
||||
? isActive || isSelected || isExploreRoot || isExpanded || n.degree >= labelThreshold
|
||||
: isActive || isSelected || isExploreRoot;
|
||||
if (alpha > 0.5 && glowEligible) {
|
||||
ctx.shadowColor = fillColor;
|
||||
ctx.shadowBlur = isActive || isSelected || isExploreRoot ? 24 : 12;
|
||||
} else {
|
||||
ctx.shadowBlur = 0;
|
||||
}
|
||||
|
||||
ctx.beginPath();
|
||||
ctx.arc(n.x, n.y, r, 0, Math.PI * 2);
|
||||
ctx.fillStyle = fillColor;
|
||||
ctx.fill();
|
||||
ctx.shadowBlur = 0;
|
||||
|
||||
// Rim.
|
||||
ctx.lineWidth = strokeWidth;
|
||||
ctx.strokeStyle = strokeColor;
|
||||
ctx.stroke();
|
||||
|
||||
if (isExploreRoot) {
|
||||
ctx.beginPath();
|
||||
ctx.arc(n.x, n.y, r + 8 / k, 0, Math.PI * 2);
|
||||
ctx.lineWidth = 1.4 / k;
|
||||
ctx.strokeStyle = palette.rootRing;
|
||||
ctx.stroke();
|
||||
} else if (isExpanded) {
|
||||
ctx.beginPath();
|
||||
ctx.arc(n.x + r * 0.58, n.y - r * 0.58, Math.max(4 / k, 3), 0, Math.PI * 2);
|
||||
ctx.fillStyle = palette.expandedDot;
|
||||
ctx.fill();
|
||||
ctx.lineWidth = 1 / k;
|
||||
ctx.strokeStyle = palette.expandedDotRim;
|
||||
ctx.stroke();
|
||||
}
|
||||
|
||||
// Selection: double ring.
|
||||
if (isSelected) {
|
||||
ctx.beginPath();
|
||||
ctx.arc(n.x, n.y, r + 4 / k, 0, Math.PI * 2);
|
||||
ctx.lineWidth = 2 / k;
|
||||
ctx.strokeStyle = palette.selectionRing;
|
||||
ctx.stroke();
|
||||
ctx.beginPath();
|
||||
ctx.arc(n.x, n.y, r + 8 / k, 0, Math.PI * 2);
|
||||
ctx.lineWidth = 1.2 / k;
|
||||
ctx.strokeStyle = n.color;
|
||||
ctx.stroke();
|
||||
} else if (n.id === hover?.id) {
|
||||
ctx.beginPath();
|
||||
ctx.arc(n.x, n.y, r + 3.5 / k, 0, Math.PI * 2);
|
||||
ctx.lineWidth = 1.6 / k;
|
||||
ctx.strokeStyle = palette.hoverRing;
|
||||
ctx.stroke();
|
||||
}
|
||||
|
||||
// Pin marker for dragged-and-fixed nodes.
|
||||
if (n.fx != null && !isSelected) {
|
||||
ctx.beginPath();
|
||||
ctx.arc(n.x, n.y, Math.max(r * 0.28, 2 / k), 0, Math.PI * 2);
|
||||
ctx.fillStyle = palette.pin;
|
||||
ctx.fill();
|
||||
}
|
||||
|
||||
// Label.
|
||||
const wantsLabel =
|
||||
isActive ||
|
||||
isSelected ||
|
||||
isNeighbor ||
|
||||
labelByExplore ||
|
||||
(search?.has(n.id) ?? false) ||
|
||||
(p.showLabels && n.degree >= labelThreshold && k >= 0.45);
|
||||
if (wantsLabel && alpha > 0.5) {
|
||||
const text = truncateLabel(getNodeTitle(n.node), isActive ? 26 : 14);
|
||||
const fontSize = (isActive || isSelected ? 13 : 12) / k;
|
||||
ctx.font = `600 ${fontSize}px "Noto Sans SC", system-ui, sans-serif`;
|
||||
ctx.textAlign = "center";
|
||||
ctx.textBaseline = "top";
|
||||
ctx.lineWidth = 3.5 / k;
|
||||
ctx.strokeStyle = palette.labelHalo;
|
||||
ctx.strokeText(text, n.x, n.y + r + 5 / k);
|
||||
ctx.fillStyle = isActive || isSelected ? palette.labelActive : palette.label;
|
||||
ctx.fillText(text, n.x, n.y + r + 5 / k);
|
||||
}
|
||||
}
|
||||
ctx.globalAlpha = 1;
|
||||
|
||||
drawMinimap();
|
||||
// eslint-disable-next-line react-hooks/exhaustive-deps
|
||||
}, []);
|
||||
|
||||
const drawMinimap = useCallback(() => {
|
||||
const minimap = minimapRef.current;
|
||||
const sim = simRef.current;
|
||||
if (!minimap || !sim || !sim.nodes.length) return;
|
||||
const ctx = minimap.getContext("2d");
|
||||
if (!ctx) return;
|
||||
const mw = minimap.width;
|
||||
const mh = minimap.height;
|
||||
ctx.clearRect(0, 0, mw, mh);
|
||||
|
||||
let minX = Infinity;
|
||||
let minY = Infinity;
|
||||
let maxX = -Infinity;
|
||||
let maxY = -Infinity;
|
||||
for (const n of sim.nodes) {
|
||||
minX = Math.min(minX, n.x);
|
||||
minY = Math.min(minY, n.y);
|
||||
maxX = Math.max(maxX, n.x);
|
||||
maxY = Math.max(maxY, n.y);
|
||||
}
|
||||
const pad = 10;
|
||||
const scale = Math.min(
|
||||
(mw - pad * 2) / Math.max(maxX - minX, 1),
|
||||
(mh - pad * 2) / Math.max(maxY - minY, 1),
|
||||
);
|
||||
const ox = (mw - (maxX - minX) * scale) / 2 - minX * scale;
|
||||
const oy = (mh - (maxY - minY) * scale) / 2 - minY * scale;
|
||||
|
||||
for (const n of sim.nodes) {
|
||||
ctx.beginPath();
|
||||
ctx.arc(n.x * scale + ox, n.y * scale + oy, 1.6, 0, Math.PI * 2);
|
||||
ctx.fillStyle = n.color;
|
||||
ctx.globalAlpha = 0.9;
|
||||
ctx.fill();
|
||||
}
|
||||
ctx.globalAlpha = 1;
|
||||
|
||||
// Viewport rectangle.
|
||||
const t = transformRef.current;
|
||||
const { w, h } = sizeRef.current;
|
||||
if (t.k > 0 && w && h) {
|
||||
const x0 = ((0 - t.x) / t.k) * scale + ox;
|
||||
const y0 = ((0 - t.y) / t.k) * scale + oy;
|
||||
const x1 = ((w - t.x) / t.k) * scale + ox;
|
||||
const y1 = ((h - t.y) / t.k) * scale + oy;
|
||||
ctx.strokeStyle = CANVAS_PALETTE[propsRef.current.theme].minimapViewport;
|
||||
ctx.lineWidth = 1;
|
||||
ctx.strokeRect(x0, y0, x1 - x0, y1 - y0);
|
||||
}
|
||||
}, []);
|
||||
|
||||
// ----------------------------------------------------------------- loop
|
||||
|
||||
useEffect(() => {
|
||||
const loop = () => {
|
||||
const sim = simRef.current;
|
||||
let active = false;
|
||||
|
||||
const anim = animRef.current;
|
||||
if (anim) {
|
||||
const progress = Math.min(
|
||||
(performance.now() - anim.start) / anim.duration,
|
||||
1,
|
||||
);
|
||||
const e = easeCubicOut(progress);
|
||||
transformRef.current = {
|
||||
x: anim.from.x + (anim.to.x - anim.from.x) * e,
|
||||
y: anim.from.y + (anim.to.y - anim.from.y) * e,
|
||||
k: anim.from.k + (anim.to.k - anim.from.k) * e,
|
||||
};
|
||||
if (progress >= 1) {
|
||||
animRef.current = null;
|
||||
lastNotifiedZoomRef.current = transformRef.current.k;
|
||||
propsRef.current.onZoomChange(transformRef.current.k);
|
||||
}
|
||||
active = true;
|
||||
}
|
||||
|
||||
if (sim && layoutRef.current === "force" && sim.tick()) {
|
||||
active = true;
|
||||
// Until the user pans/zooms/drags, gently track the settling layout
|
||||
// so it stays in view without the viewport snapping every frame.
|
||||
if (!interactedRef.current && sizeRef.current.w) {
|
||||
const target = computeFitTransform();
|
||||
if (target) {
|
||||
const t = transformRef.current;
|
||||
const ease = t.k === 0 ? 1 : 0.16;
|
||||
setTransform({
|
||||
x: t.x + (target.x - t.x) * ease,
|
||||
y: t.y + (target.y - t.y) * ease,
|
||||
k: t.k + (target.k - t.k) * ease,
|
||||
});
|
||||
}
|
||||
autoFitRef.current = true;
|
||||
} else if (!autoFitRef.current && sim.alpha < 0.3) {
|
||||
autoFitRef.current = true;
|
||||
fitView(true);
|
||||
}
|
||||
} else if (sim && !autoFitRef.current && sim.nodes.length && sizeRef.current.w) {
|
||||
autoFitRef.current = true;
|
||||
fitView(false);
|
||||
}
|
||||
|
||||
if (active || needsDrawRef.current) {
|
||||
needsDrawRef.current = false;
|
||||
draw();
|
||||
}
|
||||
rafRef.current = requestAnimationFrame(loop);
|
||||
};
|
||||
rafRef.current = requestAnimationFrame(loop);
|
||||
return () => cancelAnimationFrame(rafRef.current);
|
||||
}, [computeFitTransform, draw, fitView, setTransform]);
|
||||
|
||||
// --------------------------------------------------------------- resize
|
||||
|
||||
useEffect(() => {
|
||||
const container = containerRef.current;
|
||||
const canvas = canvasRef.current;
|
||||
if (!container || !canvas) return;
|
||||
let timer: number | undefined;
|
||||
const apply = () => {
|
||||
const rect = container.getBoundingClientRect();
|
||||
const dpr = window.devicePixelRatio || 1;
|
||||
sizeRef.current = { w: rect.width, h: rect.height };
|
||||
canvas.width = Math.max(1, Math.round(rect.width * dpr));
|
||||
canvas.height = Math.max(1, Math.round(rect.height * dpr));
|
||||
if (transformRef.current.k === 0) {
|
||||
setTransform({ x: rect.width / 2, y: rect.height / 2, k: 0.9 });
|
||||
}
|
||||
requestDraw();
|
||||
};
|
||||
apply();
|
||||
const observer = new ResizeObserver(() => {
|
||||
window.clearTimeout(timer);
|
||||
timer = window.setTimeout(apply, 120);
|
||||
});
|
||||
observer.observe(container);
|
||||
return () => {
|
||||
window.clearTimeout(timer);
|
||||
observer.disconnect();
|
||||
};
|
||||
}, [requestDraw, setTransform]);
|
||||
|
||||
// --------------------------------------------------------------- events
|
||||
|
||||
const screenToWorld = (sx: number, sy: number) => {
|
||||
const t = transformRef.current;
|
||||
return { x: (sx - t.x) / t.k, y: (sy - t.y) / t.k };
|
||||
};
|
||||
|
||||
const hitTest = (wx: number, wy: number): SimNode | null => {
|
||||
const sim = simRef.current;
|
||||
if (!sim) return null;
|
||||
// Topmost (drawn last) first.
|
||||
for (let i = sim.nodes.length - 1; i >= 0; i--) {
|
||||
const n = sim.nodes[i];
|
||||
const dx = wx - n.x;
|
||||
const dy = wy - n.y;
|
||||
const r = n.radius + 3 / (transformRef.current.k || 1);
|
||||
if (dx * dx + dy * dy <= r * r) return n;
|
||||
}
|
||||
return null;
|
||||
};
|
||||
|
||||
const localPoint = (e: { clientX: number; clientY: number }) => {
|
||||
const rect = canvasRef.current!.getBoundingClientRect();
|
||||
return { sx: e.clientX - rect.left, sy: e.clientY - rect.top };
|
||||
};
|
||||
|
||||
const setHover = (node: SimNode | null, e?: React.PointerEvent) => {
|
||||
if (hoverRef.current?.id !== node?.id) {
|
||||
hoverRef.current = node;
|
||||
requestDraw();
|
||||
}
|
||||
const canvas = canvasRef.current;
|
||||
if (canvas) canvas.style.cursor = node ? "pointer" : "grab";
|
||||
if (node && e) {
|
||||
propsRef.current.onTooltip({
|
||||
x: e.clientX,
|
||||
y: e.clientY,
|
||||
node: node.node,
|
||||
kind: node.kind,
|
||||
degree: node.degree,
|
||||
});
|
||||
} else {
|
||||
propsRef.current.onTooltip(null);
|
||||
}
|
||||
};
|
||||
|
||||
const handlePointerDown = (e: React.PointerEvent<HTMLCanvasElement>) => {
|
||||
if (e.button !== 0) return;
|
||||
const { sx, sy } = localPoint(e);
|
||||
const { x: wx, y: wy } = screenToWorld(sx, sy);
|
||||
const node = hitTest(wx, wy);
|
||||
animRef.current = null;
|
||||
dragRef.current = {
|
||||
pointerId: e.pointerId,
|
||||
node,
|
||||
startX: sx,
|
||||
startY: sy,
|
||||
originX: transformRef.current.x,
|
||||
originY: transformRef.current.y,
|
||||
moved: false,
|
||||
};
|
||||
if (node) {
|
||||
node.fx = node.x;
|
||||
node.fy = node.y;
|
||||
}
|
||||
e.currentTarget.setPointerCapture(e.pointerId);
|
||||
};
|
||||
|
||||
const handlePointerMove = (e: React.PointerEvent<HTMLCanvasElement>) => {
|
||||
const { sx, sy } = localPoint(e);
|
||||
const drag = dragRef.current;
|
||||
if (drag && drag.pointerId === e.pointerId) {
|
||||
const dx = sx - drag.startX;
|
||||
const dy = sy - drag.startY;
|
||||
if (Math.abs(dx) + Math.abs(dy) > 3) {
|
||||
drag.moved = true;
|
||||
// Only a real pan/drag takes over the camera; a plain click keeps
|
||||
// the auto-follow so the settling graph can't drift out of view.
|
||||
interactedRef.current = true;
|
||||
}
|
||||
if (drag.node) {
|
||||
const world = screenToWorld(sx, sy);
|
||||
drag.node.fx = world.x;
|
||||
drag.node.fy = world.y;
|
||||
if (layoutRef.current === "force") simRef.current?.reheat(0.3);
|
||||
propsRef.current.onTooltip(null);
|
||||
} else {
|
||||
transformRef.current = {
|
||||
...transformRef.current,
|
||||
x: drag.originX + dx,
|
||||
y: drag.originY + dy,
|
||||
};
|
||||
}
|
||||
requestDraw();
|
||||
return;
|
||||
}
|
||||
const world = screenToWorld(sx, sy);
|
||||
setHover(hitTest(world.x, world.y), e);
|
||||
};
|
||||
|
||||
const handlePointerUp = (e: React.PointerEvent<HTMLCanvasElement>) => {
|
||||
const drag = dragRef.current;
|
||||
if (!drag || drag.pointerId !== e.pointerId) return;
|
||||
dragRef.current = null;
|
||||
if (!drag.moved) {
|
||||
// Plain click: select node / clear selection on empty space.
|
||||
if (drag.node) {
|
||||
drag.node.fx = null;
|
||||
drag.node.fy = null;
|
||||
propsRef.current.onSelectNode(drag.node.id);
|
||||
} else {
|
||||
propsRef.current.onSelectNode(null);
|
||||
}
|
||||
requestDraw();
|
||||
}
|
||||
// A real drag leaves the node pinned at its drop position.
|
||||
};
|
||||
|
||||
const handleDoubleClick = (e: React.MouseEvent<HTMLCanvasElement>) => {
|
||||
const { sx, sy } = localPoint(e);
|
||||
const world = screenToWorld(sx, sy);
|
||||
const node = hitTest(world.x, world.y);
|
||||
if (node) {
|
||||
propsRef.current.onSelectNode(node.id);
|
||||
focusNode(node.id);
|
||||
}
|
||||
};
|
||||
|
||||
const handlePointerLeave = () => {
|
||||
setHover(null);
|
||||
};
|
||||
|
||||
// Wheel zoom needs a non-passive listener to preventDefault.
|
||||
useEffect(() => {
|
||||
const canvas = canvasRef.current;
|
||||
if (!canvas) return;
|
||||
const onWheel = (e: WheelEvent) => {
|
||||
e.preventDefault();
|
||||
interactedRef.current = true;
|
||||
const rect = canvas.getBoundingClientRect();
|
||||
zoomBy(
|
||||
Math.exp(-e.deltaY * 0.0012),
|
||||
e.clientX - rect.left,
|
||||
e.clientY - rect.top,
|
||||
);
|
||||
};
|
||||
canvas.addEventListener("wheel", onWheel, { passive: false });
|
||||
return () => canvas.removeEventListener("wheel", onWheel);
|
||||
}, [zoomBy]);
|
||||
|
||||
const minimapVisible = useMemo(() => props.nodes.length > 0, [props.nodes]);
|
||||
|
||||
return (
|
||||
<div className="kg-canvas-wrap" ref={containerRef}>
|
||||
<canvas
|
||||
ref={canvasRef}
|
||||
className="kg-canvas"
|
||||
onPointerDown={handlePointerDown}
|
||||
onPointerMove={handlePointerMove}
|
||||
onPointerUp={handlePointerUp}
|
||||
onPointerLeave={handlePointerLeave}
|
||||
onDoubleClick={handleDoubleClick}
|
||||
/>
|
||||
{props.toolbar ? (
|
||||
<div className="kg-canvas-toolbar">{props.toolbar}</div>
|
||||
) : null}
|
||||
{props.hint}
|
||||
{minimapVisible ? (
|
||||
<canvas ref={minimapRef} className="kg-minimap" width={180} height={120} />
|
||||
) : null}
|
||||
{props.overlay}
|
||||
</div>
|
||||
);
|
||||
},
|
||||
);
|
||||
@@ -1,287 +0,0 @@
|
||||
import {
|
||||
Ban,
|
||||
Compass,
|
||||
Crosshair,
|
||||
EyeOff,
|
||||
FileText,
|
||||
Loader2,
|
||||
MousePointerClick,
|
||||
Waypoints,
|
||||
X,
|
||||
} from "lucide-react";
|
||||
|
||||
import type { KnowledgeNote } from "@/core/knowledge/types";
|
||||
|
||||
import type { GraphNode } from "./graphTypes";
|
||||
import {
|
||||
getEdgeInfo,
|
||||
getNodeKind,
|
||||
getNodeTitle,
|
||||
NODE_KIND_META,
|
||||
} from "./graphUtils";
|
||||
|
||||
const SOURCE_TYPE_ZH: Record<string, string> = {
|
||||
thread: "对话沉淀",
|
||||
search: "搜索沉淀",
|
||||
tool: "工具结果",
|
||||
file: "文件导入",
|
||||
manual: "手动创建",
|
||||
report: "报告",
|
||||
};
|
||||
|
||||
const STATUS_ZH: Record<string, string> = {
|
||||
draft: "草稿",
|
||||
approved: "已审核",
|
||||
archived: "已归档",
|
||||
};
|
||||
|
||||
export interface NeighborItem {
|
||||
node: GraphNode;
|
||||
relation: string;
|
||||
}
|
||||
|
||||
interface GraphDetailPanelProps {
|
||||
node: GraphNode | null;
|
||||
degree: number;
|
||||
neighbors: NeighborItem[];
|
||||
/** Note detail (only for note-type nodes; undefined while absent/loading). */
|
||||
note: KnowledgeNote | undefined;
|
||||
noteLoading: boolean;
|
||||
/** Explore-mode context; null when in full mode. */
|
||||
explore: {
|
||||
/** Depth of this node in the explore graph; null = not yet visible. */
|
||||
depth: number | null;
|
||||
expanded: boolean;
|
||||
isRoot: boolean;
|
||||
} | null;
|
||||
onViewSource: (noteId: string) => void;
|
||||
onLocate: (id: string) => void;
|
||||
onHide: (id: string) => void;
|
||||
onBlacklist: (id: string) => void;
|
||||
onPickNeighbor: (id: string) => void;
|
||||
onExpandNode: (id: string) => void;
|
||||
onSetExploreRoot: (id: string) => void;
|
||||
onClose: () => void;
|
||||
}
|
||||
|
||||
function formatTime(value: string | null | undefined): string {
|
||||
if (!value) return "—";
|
||||
const date = new Date(value);
|
||||
if (Number.isNaN(date.getTime())) return value;
|
||||
return date.toLocaleString("zh-CN", { hour12: false });
|
||||
}
|
||||
|
||||
export function GraphDetailPanel(props: GraphDetailPanelProps) {
|
||||
const { node } = props;
|
||||
|
||||
return (
|
||||
<aside className="kg-panel kg-detail">
|
||||
<div className="kg-panel-header">
|
||||
<span className="kg-panel-title">节点详情</span>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-panel-close-btn"
|
||||
title="收起详情面板"
|
||||
onClick={props.onClose}
|
||||
>
|
||||
<X size={15} />
|
||||
</button>
|
||||
</div>
|
||||
{!node ? (
|
||||
<div className="kg-detail-empty">
|
||||
<MousePointerClick size={28} />
|
||||
<div className="kg-detail-empty-title">选择一个节点查看详情</div>
|
||||
<div>
|
||||
点击图谱中的文档、主题或实体节点,
|
||||
<br />
|
||||
查看摘要、标签、关联关系和相邻节点。
|
||||
</div>
|
||||
</div>
|
||||
) : (
|
||||
<DetailBody {...props} node={node} />
|
||||
)}
|
||||
</aside>
|
||||
);
|
||||
}
|
||||
|
||||
function DetailBody(props: GraphDetailPanelProps & { node: GraphNode }) {
|
||||
const { node, note } = props;
|
||||
const kind = getNodeKind(node);
|
||||
const kindMeta = NODE_KIND_META[kind];
|
||||
const isNote = (node.type || "").toLowerCase() === "note";
|
||||
|
||||
return (
|
||||
<div className="kg-panel-content">
|
||||
<h2 className="kg-detail-title">{getNodeTitle(node)}</h2>
|
||||
<div>
|
||||
<span
|
||||
className="kg-detail-type"
|
||||
style={{
|
||||
background: `${kindMeta.color}26`,
|
||||
color: kindMeta.color,
|
||||
}}
|
||||
>
|
||||
<span className="kg-legend-dot" style={{ background: kindMeta.color }} />
|
||||
{kindMeta.label}
|
||||
{node.category && node.category !== kindMeta.label
|
||||
? ` · ${node.category}`
|
||||
: ""}
|
||||
</span>
|
||||
<span className="kg-detail-type kg-detail-type-neutral">
|
||||
度数 {props.degree}
|
||||
</span>
|
||||
{props.explore && props.explore.depth !== null ? (
|
||||
<>
|
||||
<span className="kg-detail-type kg-detail-type-primary">
|
||||
{props.explore.isRoot
|
||||
? "探索起点"
|
||||
: `第 ${props.explore.depth} 层`}
|
||||
</span>
|
||||
<span
|
||||
className={`kg-detail-type ${
|
||||
props.explore.expanded
|
||||
? "kg-detail-type-success"
|
||||
: "kg-detail-type-warn"
|
||||
}`}
|
||||
>
|
||||
{props.explore.expanded ? "已展开" : "可展开"}
|
||||
</span>
|
||||
</>
|
||||
) : null}
|
||||
</div>
|
||||
|
||||
{props.noteLoading ? (
|
||||
<div className="kg-detail-summary" style={{ display: "flex", gap: 8 }}>
|
||||
<Loader2 size={14} className="animate-spin" /> 加载笔记详情...
|
||||
</div>
|
||||
) : null}
|
||||
|
||||
{note?.summary ? <p className="kg-detail-summary">{note.summary}</p> : null}
|
||||
|
||||
{note?.tags?.length ? (
|
||||
<div className="kg-tag-list">
|
||||
{note.tags.map((tag) => (
|
||||
<span key={tag} className="kg-tag">
|
||||
#{tag}
|
||||
</span>
|
||||
))}
|
||||
</div>
|
||||
) : null}
|
||||
|
||||
{note ? (
|
||||
<div className="kg-detail-meta">
|
||||
<span className="kg-detail-meta-label">来源</span>
|
||||
<span className="kg-detail-meta-value">
|
||||
{SOURCE_TYPE_ZH[note.source_type] ?? note.source_type}
|
||||
</span>
|
||||
<span className="kg-detail-meta-label">状态</span>
|
||||
<span className="kg-detail-meta-value">
|
||||
{STATUS_ZH[note.status] ?? note.status}
|
||||
</span>
|
||||
<span className="kg-detail-meta-label">创建时间</span>
|
||||
<span className="kg-detail-meta-value">{formatTime(note.created_at)}</span>
|
||||
<span className="kg-detail-meta-label">更新时间</span>
|
||||
<span className="kg-detail-meta-value">{formatTime(note.updated_at)}</span>
|
||||
{note.created_by_name ? (
|
||||
<>
|
||||
<span className="kg-detail-meta-label">创建人</span>
|
||||
<span className="kg-detail-meta-value">{note.created_by_name}</span>
|
||||
</>
|
||||
) : null}
|
||||
</div>
|
||||
) : null}
|
||||
|
||||
<div className="kg-detail-actions">
|
||||
{isNote ? (
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
onClick={() => props.onViewSource(node.id)}
|
||||
>
|
||||
<FileText size={14} /> 查看原文
|
||||
</button>
|
||||
) : null}
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
onClick={() => props.onLocate(node.id)}
|
||||
>
|
||||
<Crosshair size={14} /> 定位节点
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
onClick={() => props.onHide(node.id)}
|
||||
>
|
||||
<EyeOff size={14} /> 隐藏节点
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="加入黑名单后该节点将不再出现在图谱中,可在左侧「图谱黑名单」中移除"
|
||||
onClick={() => props.onBlacklist(node.id)}
|
||||
>
|
||||
<Ban size={14} /> 加入黑名单
|
||||
</button>
|
||||
{props.explore ? (
|
||||
<>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
disabled={props.explore.depth === null || props.explore.expanded}
|
||||
title={
|
||||
props.explore.expanded
|
||||
? "该节点的一跳邻居已全部展开"
|
||||
: "展开该节点的一跳邻居"
|
||||
}
|
||||
onClick={() => props.onExpandNode(node.id)}
|
||||
>
|
||||
<Waypoints size={14} /> 展开一跳邻居
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="清空当前探索路径,以此节点为新起点"
|
||||
onClick={() => props.onSetExploreRoot(node.id)}
|
||||
>
|
||||
<Compass size={14} /> 以此为新起点
|
||||
</button>
|
||||
</>
|
||||
) : null}
|
||||
</div>
|
||||
|
||||
<h3 className="kg-detail-section-title">
|
||||
关联节点({props.neighbors.length})
|
||||
</h3>
|
||||
{props.neighbors.length === 0 ? (
|
||||
<div className="kg-search-hint">该节点暂无可见的关联节点</div>
|
||||
) : (
|
||||
<div className="kg-neighbor-list">
|
||||
{props.neighbors.map(({ node: neighbor, relation }) => {
|
||||
const neighborKind = NODE_KIND_META[getNodeKind(neighbor)];
|
||||
const relationInfo = getEdgeInfo(relation);
|
||||
return (
|
||||
<div
|
||||
key={`${neighbor.id}:${relation}`}
|
||||
className="kg-neighbor-item"
|
||||
onClick={() => props.onPickNeighbor(neighbor.id)}
|
||||
>
|
||||
<span
|
||||
className="kg-legend-dot"
|
||||
style={{ background: neighborKind.color, flexShrink: 0 }}
|
||||
/>
|
||||
<span className="kg-neighbor-title">{getNodeTitle(neighbor)}</span>
|
||||
<span
|
||||
className="kg-neighbor-relation"
|
||||
style={{ color: relationInfo.color }}
|
||||
>
|
||||
{relationInfo.zh}
|
||||
</span>
|
||||
</div>
|
||||
);
|
||||
})}
|
||||
</div>
|
||||
)}
|
||||
</div>
|
||||
);
|
||||
}
|
||||
@@ -1,57 +0,0 @@
|
||||
import type { NodeKind } from "./graphTypes";
|
||||
import { getEdgeInfo, NODE_KIND_META, NODE_KIND_ORDER } from "./graphUtils";
|
||||
|
||||
interface GraphFooterProps {
|
||||
nodeKindCounts: Map<NodeKind, number>;
|
||||
edgeLabels: string[];
|
||||
zoomLevel: number;
|
||||
selectedTitle: string | null;
|
||||
visibleNodes: number;
|
||||
visibleEdges: number;
|
||||
}
|
||||
|
||||
export function GraphFooter(props: GraphFooterProps) {
|
||||
const kinds = NODE_KIND_ORDER.filter((k) => (props.nodeKindCounts.get(k) ?? 0) > 0);
|
||||
|
||||
return (
|
||||
<footer className="kg-footer">
|
||||
<div className="kg-legend">
|
||||
{kinds.map((kind) => (
|
||||
<span key={kind} className="kg-legend-item">
|
||||
<span
|
||||
className="kg-legend-dot"
|
||||
style={{ background: NODE_KIND_META[kind].color }}
|
||||
/>
|
||||
{NODE_KIND_META[kind].label}
|
||||
</span>
|
||||
))}
|
||||
{props.edgeLabels.slice(0, 6).map((label) => {
|
||||
const info = getEdgeInfo(label);
|
||||
return (
|
||||
<span key={label} className="kg-legend-item">
|
||||
<span
|
||||
className="kg-legend-line"
|
||||
style={{
|
||||
borderTopColor: info.color,
|
||||
borderTopStyle: info.dashed ? "dashed" : "solid",
|
||||
}}
|
||||
/>
|
||||
{info.zh}
|
||||
</span>
|
||||
);
|
||||
})}
|
||||
</div>
|
||||
<div className="kg-footer-status">
|
||||
{props.selectedTitle ? (
|
||||
<span className="kg-footer-selected" title={props.selectedTitle}>
|
||||
选中:{props.selectedTitle}
|
||||
</span>
|
||||
) : null}
|
||||
<span>
|
||||
显示 {props.visibleNodes} 节点 · {props.visibleEdges} 关系
|
||||
</span>
|
||||
<span>缩放 {Math.round(props.zoomLevel * 100)}%</span>
|
||||
</div>
|
||||
</footer>
|
||||
);
|
||||
}
|
||||
@@ -1,124 +0,0 @@
|
||||
import {
|
||||
ArrowLeft,
|
||||
Download,
|
||||
Loader2,
|
||||
Maximize2,
|
||||
RefreshCw,
|
||||
Scan,
|
||||
Shuffle,
|
||||
ZoomIn,
|
||||
ZoomOut,
|
||||
} from "lucide-react";
|
||||
|
||||
interface GraphHeaderProps {
|
||||
nodeCount: number;
|
||||
edgeCount: number;
|
||||
totalNodes: number;
|
||||
totalEdges: number;
|
||||
refreshing: boolean;
|
||||
onBack: () => void;
|
||||
onRelayout: () => void;
|
||||
onFitView: () => void;
|
||||
onZoomIn: () => void;
|
||||
onZoomOut: () => void;
|
||||
onExport: () => void;
|
||||
onFullscreen: () => void;
|
||||
onRefresh: () => void;
|
||||
}
|
||||
|
||||
export function GraphHeader(props: GraphHeaderProps) {
|
||||
const filteredNodes =
|
||||
props.nodeCount === props.totalNodes
|
||||
? `${props.totalNodes}`
|
||||
: `${props.nodeCount}/${props.totalNodes}`;
|
||||
const filteredEdges =
|
||||
props.edgeCount === props.totalEdges
|
||||
? `${props.totalEdges}`
|
||||
: `${props.edgeCount}/${props.totalEdges}`;
|
||||
|
||||
return (
|
||||
<header className="kg-header">
|
||||
<div className="kg-header-left">
|
||||
<button type="button" className="kg-back" onClick={props.onBack}>
|
||||
<ArrowLeft size={15} />
|
||||
<span className="kg-button-text">返回</span>
|
||||
</button>
|
||||
<div className="kg-title-block">
|
||||
<h1 className="kg-title">知识图谱</h1>
|
||||
<span className="kg-subtitle">知识关系可视化</span>
|
||||
</div>
|
||||
<div className="kg-stat-pills">
|
||||
<span className="kg-stat-pill">{filteredNodes} 个节点</span>
|
||||
<span className="kg-stat-pill kg-stat-pill-violet">
|
||||
{filteredEdges} 条关系
|
||||
</span>
|
||||
</div>
|
||||
</div>
|
||||
<div className="kg-header-actions">
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="重新布局"
|
||||
onClick={props.onRelayout}
|
||||
>
|
||||
<Shuffle size={15} />
|
||||
<span className="kg-button-text">重新布局</span>
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="适应视图"
|
||||
onClick={props.onFitView}
|
||||
>
|
||||
<Scan size={15} />
|
||||
<span className="kg-button-text">适应视图</span>
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="放大"
|
||||
onClick={props.onZoomIn}
|
||||
>
|
||||
<ZoomIn size={15} />
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="缩小"
|
||||
onClick={props.onZoomOut}
|
||||
>
|
||||
<ZoomOut size={15} />
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="导出图片"
|
||||
onClick={props.onExport}
|
||||
>
|
||||
<Download size={15} />
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="全屏"
|
||||
onClick={props.onFullscreen}
|
||||
>
|
||||
<Maximize2 size={15} />
|
||||
</button>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
title="刷新图谱数据"
|
||||
onClick={props.onRefresh}
|
||||
disabled={props.refreshing}
|
||||
>
|
||||
{props.refreshing ? (
|
||||
<Loader2 size={15} className="animate-spin" />
|
||||
) : (
|
||||
<RefreshCw size={15} />
|
||||
)}
|
||||
</button>
|
||||
</div>
|
||||
</header>
|
||||
);
|
||||
}
|
||||
@@ -1,292 +0,0 @@
|
||||
import { ChevronDown, ChevronRight, Eraser, Search, X } from "lucide-react";
|
||||
import { useState } from "react";
|
||||
|
||||
import type { GraphBlacklistItem } from "@/core/knowledge/api";
|
||||
|
||||
import type { GraphNode, GraphStats, NodeKind } from "./graphTypes";
|
||||
import {
|
||||
getEdgeInfo,
|
||||
getNodeTitle,
|
||||
NODE_KIND_META,
|
||||
NODE_KIND_ORDER,
|
||||
} from "./graphUtils";
|
||||
|
||||
interface GraphSidebarProps {
|
||||
searchKeyword: string;
|
||||
searchMatchCount: number | null;
|
||||
nodeKindCounts: Map<NodeKind, number>;
|
||||
edgeLabelCounts: Map<string, number>;
|
||||
disabledNodeKinds: Set<NodeKind>;
|
||||
disabledEdgeLabels: Set<string>;
|
||||
stats: GraphStats;
|
||||
topNodes: Array<{ node: GraphNode; degree: number }>;
|
||||
hiddenCount: number;
|
||||
blacklistItems: GraphBlacklistItem[];
|
||||
blacklistPending: boolean;
|
||||
junkCandidateCount: number;
|
||||
onSearchChange: (keyword: string) => void;
|
||||
onSearchEnter: () => void;
|
||||
onToggleNodeKind: (kind: NodeKind) => void;
|
||||
onToggleEdgeLabel: (label: string) => void;
|
||||
onRestoreHidden: () => void;
|
||||
onRemoveBlacklist: (entryId: string) => void;
|
||||
onCleanJunk: () => void;
|
||||
onPickNode: (id: string) => void;
|
||||
onClose: () => void;
|
||||
}
|
||||
|
||||
type SectionKey =
|
||||
| "nodeTypes"
|
||||
| "edgeTypes"
|
||||
| "graphStats"
|
||||
| "coreNodes"
|
||||
| "blacklist";
|
||||
|
||||
interface FilterSectionProps {
|
||||
title: string;
|
||||
extra?: string;
|
||||
collapsed: boolean;
|
||||
onToggle: () => void;
|
||||
children: React.ReactNode;
|
||||
}
|
||||
|
||||
function FilterSection(props: FilterSectionProps) {
|
||||
return (
|
||||
<section className="kg-filter-section">
|
||||
<button
|
||||
type="button"
|
||||
className="kg-filter-section-header"
|
||||
onClick={props.onToggle}
|
||||
aria-expanded={!props.collapsed}
|
||||
>
|
||||
<span>{props.title}</span>
|
||||
<span className="kg-filter-section-extra">
|
||||
{props.extra}
|
||||
{props.collapsed ? <ChevronRight size={14} /> : <ChevronDown size={14} />}
|
||||
</span>
|
||||
</button>
|
||||
{!props.collapsed ? (
|
||||
<div className="kg-filter-section-body">{props.children}</div>
|
||||
) : null}
|
||||
</section>
|
||||
);
|
||||
}
|
||||
|
||||
export function GraphSidebar(props: GraphSidebarProps) {
|
||||
// 节点类型 / 关系类型默认收起,统计与核心节点默认展开。
|
||||
const [collapsedSections, setCollapsedSections] = useState<
|
||||
Record<SectionKey, boolean>
|
||||
>({
|
||||
nodeTypes: true,
|
||||
edgeTypes: true,
|
||||
graphStats: false,
|
||||
coreNodes: false,
|
||||
blacklist: true,
|
||||
});
|
||||
|
||||
const toggleFilterSection = (key: SectionKey) => {
|
||||
setCollapsedSections((prev) => ({ ...prev, [key]: !prev[key] }));
|
||||
};
|
||||
|
||||
const kinds = NODE_KIND_ORDER.filter((k) => (props.nodeKindCounts.get(k) ?? 0) > 0);
|
||||
const edgeLabels = [...props.edgeLabelCounts.entries()].sort((a, b) => b[1] - a[1]);
|
||||
|
||||
return (
|
||||
<aside className="kg-panel kg-sidebar">
|
||||
<div className="kg-panel-header">
|
||||
<span className="kg-panel-title">搜索与筛选</span>
|
||||
<div className="kg-panel-header-actions">
|
||||
{props.hiddenCount > 0 ? (
|
||||
<button
|
||||
type="button"
|
||||
className="kg-text-button"
|
||||
onClick={props.onRestoreHidden}
|
||||
>
|
||||
恢复 {props.hiddenCount} 个隐藏节点
|
||||
</button>
|
||||
) : null}
|
||||
<button
|
||||
type="button"
|
||||
className="kg-panel-close-btn"
|
||||
title="收起筛选面板"
|
||||
onClick={props.onClose}
|
||||
>
|
||||
<X size={15} />
|
||||
</button>
|
||||
</div>
|
||||
</div>
|
||||
<div className="kg-panel-content">
|
||||
<div className="kg-search-wrap">
|
||||
<Search size={14} />
|
||||
<input
|
||||
className="kg-search"
|
||||
placeholder="搜索节点标题..."
|
||||
value={props.searchKeyword}
|
||||
onChange={(e) => props.onSearchChange(e.target.value)}
|
||||
onKeyDown={(e) => {
|
||||
if (e.key === "Enter") props.onSearchEnter();
|
||||
}}
|
||||
/>
|
||||
</div>
|
||||
{props.searchMatchCount !== null ? (
|
||||
<div
|
||||
className={`kg-search-hint${props.searchMatchCount === 0 ? " kg-search-hint-warn" : ""}`}
|
||||
>
|
||||
{props.searchMatchCount === 0
|
||||
? "没有匹配的节点,换个关键词试试"
|
||||
: `匹配 ${props.searchMatchCount} 个节点,回车定位`}
|
||||
</div>
|
||||
) : null}
|
||||
|
||||
<FilterSection
|
||||
title="节点类型"
|
||||
extra={`${kinds.length} 类`}
|
||||
collapsed={collapsedSections.nodeTypes}
|
||||
onToggle={() => toggleFilterSection("nodeTypes")}
|
||||
>
|
||||
{kinds.map((kind) => (
|
||||
<label key={kind} className="kg-check-row">
|
||||
<span className="kg-check-row-main">
|
||||
<input
|
||||
type="checkbox"
|
||||
checked={!props.disabledNodeKinds.has(kind)}
|
||||
onChange={() => props.onToggleNodeKind(kind)}
|
||||
/>
|
||||
<span
|
||||
className="kg-legend-dot"
|
||||
style={{ background: NODE_KIND_META[kind].color }}
|
||||
/>
|
||||
<span className="kg-check-label">{NODE_KIND_META[kind].label}</span>
|
||||
</span>
|
||||
<span className="kg-count">{props.nodeKindCounts.get(kind)}</span>
|
||||
</label>
|
||||
))}
|
||||
</FilterSection>
|
||||
|
||||
{edgeLabels.length > 0 ? (
|
||||
<FilterSection
|
||||
title="关系类型"
|
||||
extra={`${edgeLabels.length} 类`}
|
||||
collapsed={collapsedSections.edgeTypes}
|
||||
onToggle={() => toggleFilterSection("edgeTypes")}
|
||||
>
|
||||
{edgeLabels.map(([label, count]) => {
|
||||
const info = getEdgeInfo(label);
|
||||
return (
|
||||
<label key={label} className="kg-check-row">
|
||||
<span className="kg-check-row-main">
|
||||
<input
|
||||
type="checkbox"
|
||||
checked={!props.disabledEdgeLabels.has(label)}
|
||||
onChange={() => props.onToggleEdgeLabel(label)}
|
||||
/>
|
||||
<span
|
||||
className="kg-legend-line"
|
||||
style={{
|
||||
borderTopColor: info.color,
|
||||
borderTopStyle: info.dashed ? "dashed" : "solid",
|
||||
}}
|
||||
/>
|
||||
<span className="kg-check-label">{info.zh}</span>
|
||||
</span>
|
||||
<span className="kg-count">{count}</span>
|
||||
</label>
|
||||
);
|
||||
})}
|
||||
</FilterSection>
|
||||
) : null}
|
||||
|
||||
<FilterSection
|
||||
title="图谱统计"
|
||||
collapsed={collapsedSections.graphStats}
|
||||
onToggle={() => toggleFilterSection("graphStats")}
|
||||
>
|
||||
<div className="kg-stats-grid">
|
||||
<div className="kg-stat-card">
|
||||
<div className="kg-stat-card-value">{props.stats.nodes}</div>
|
||||
<div className="kg-stat-card-label">节点</div>
|
||||
</div>
|
||||
<div className="kg-stat-card">
|
||||
<div className="kg-stat-card-value">{props.stats.edges}</div>
|
||||
<div className="kg-stat-card-label">关系</div>
|
||||
</div>
|
||||
<div className="kg-stat-card">
|
||||
<div className="kg-stat-card-value">
|
||||
{props.stats.avgDegree.toFixed(2)}
|
||||
</div>
|
||||
<div className="kg-stat-card-label">平均度</div>
|
||||
</div>
|
||||
<div className="kg-stat-card">
|
||||
<div className="kg-stat-card-value">{props.stats.largestComponent}</div>
|
||||
<div className="kg-stat-card-label">最大连通分量</div>
|
||||
</div>
|
||||
</div>
|
||||
</FilterSection>
|
||||
|
||||
{props.topNodes.length > 0 ? (
|
||||
<FilterSection
|
||||
title="核心节点"
|
||||
extra={`${props.topNodes.length} 个`}
|
||||
collapsed={collapsedSections.coreNodes}
|
||||
onToggle={() => toggleFilterSection("coreNodes")}
|
||||
>
|
||||
{props.topNodes.map(({ node, degree }, i) => (
|
||||
<div
|
||||
key={node.id}
|
||||
className="kg-rank-item"
|
||||
onClick={() => props.onPickNode(node.id)}
|
||||
>
|
||||
<span className="kg-rank-index">{i + 1}</span>
|
||||
<span className="kg-rank-title">{getNodeTitle(node)}</span>
|
||||
<span className="kg-rank-degree">{degree} 度</span>
|
||||
</div>
|
||||
))}
|
||||
</FilterSection>
|
||||
) : null}
|
||||
|
||||
<FilterSection
|
||||
title="图谱黑名单"
|
||||
extra={`${props.blacklistItems.length} 个`}
|
||||
collapsed={collapsedSections.blacklist}
|
||||
onToggle={() => toggleFilterSection("blacklist")}
|
||||
>
|
||||
<div className="kg-search-hint">
|
||||
黑名单中的主题 / 节点不会出现在图谱中。可在节点详情里加入黑名单。
|
||||
</div>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-icon-button"
|
||||
disabled={props.blacklistPending || props.junkCandidateCount === 0}
|
||||
title="将单字符、纯数字符号、通用泛词等低价值主题批量加入黑名单"
|
||||
onClick={props.onCleanJunk}
|
||||
>
|
||||
<Eraser size={13} /> 清理低价值主题
|
||||
{props.junkCandidateCount > 0 ? `(${props.junkCandidateCount})` : ""}
|
||||
</button>
|
||||
{props.blacklistItems.length === 0 ? (
|
||||
<div className="kg-search-hint">暂无黑名单条目。</div>
|
||||
) : (
|
||||
props.blacklistItems.map((item) => (
|
||||
<div key={item.id} className="kg-check-row">
|
||||
<span className="kg-check-row-main">
|
||||
<span className="kg-check-label" title={item.label}>
|
||||
{item.label}
|
||||
</span>
|
||||
</span>
|
||||
<button
|
||||
type="button"
|
||||
className="kg-text-button"
|
||||
disabled={props.blacklistPending}
|
||||
title="移出黑名单,重新显示该节点"
|
||||
onClick={() => props.onRemoveBlacklist(item.id)}
|
||||
>
|
||||
移除
|
||||
</button>
|
||||
</div>
|
||||
))
|
||||
)}
|
||||
</FilterSection>
|
||||
</div>
|
||||
</aside>
|
||||
);
|
||||
}
|
||||
@@ -1,29 +0,0 @@
|
||||
import type { TooltipState } from "./graphTypes";
|
||||
import { getNodeTitle, NODE_KIND_META } from "./graphUtils";
|
||||
|
||||
interface GraphTooltipProps {
|
||||
tooltip: TooltipState | null;
|
||||
}
|
||||
|
||||
export function GraphTooltip({ tooltip }: GraphTooltipProps) {
|
||||
if (!tooltip) return null;
|
||||
const meta = NODE_KIND_META[tooltip.kind];
|
||||
// Keep the tooltip inside the viewport.
|
||||
const left = Math.min(tooltip.x + 14, window.innerWidth - 336);
|
||||
const top = Math.min(tooltip.y + 14, window.innerHeight - 120);
|
||||
|
||||
return (
|
||||
<div className="kg-tooltip" style={{ left, top }}>
|
||||
<div className="kg-tooltip-title">{getNodeTitle(tooltip.node)}</div>
|
||||
<div className="kg-tooltip-desc">
|
||||
<span className="kg-legend-item">
|
||||
<span className="kg-legend-dot" style={{ background: meta.color }} />
|
||||
{meta.label}
|
||||
</span>
|
||||
{tooltip.node.category ? <span>· {tooltip.node.category}</span> : null}
|
||||
<span>· 度数 {tooltip.degree}</span>
|
||||
</div>
|
||||
<div className="kg-tooltip-hint">单击查看详情 · 双击聚焦 · 拖拽固定位置</div>
|
||||
</div>
|
||||
);
|
||||
}
|
||||
File diff suppressed because it is too large.
Load diff
@@ -1,111 +0,0 @@
|
||||
/** Shared types for the knowledge-graph workbench. */
|
||||
|
||||
/** Raw node as returned by `GET /api/knowledge/graph` (plus compat fields). */
|
||||
export interface GraphNode {
|
||||
id: string;
|
||||
/** "note" | "entity" on the current backend; kept loose for compat. */
|
||||
type?: string;
|
||||
label?: string;
|
||||
title?: string;
|
||||
name?: string;
|
||||
/** note: source_type (thread/file/manual/...); entity: entity_type. */
|
||||
category?: string | null;
|
||||
}
|
||||
|
||||
/** Raw edge as returned by `GET /api/knowledge/graph` (plus compat fields). */
|
||||
export interface GraphEdge {
|
||||
id?: string;
|
||||
source: string;
|
||||
target: string;
|
||||
/** Relation type: "links" | "related" | LLM relation_type. */
|
||||
label?: string;
|
||||
type?: string;
|
||||
relation?: string;
|
||||
weight?: number;
|
||||
directed?: boolean;
|
||||
}
|
||||
|
||||
export interface GraphData {
|
||||
nodes: GraphNode[];
|
||||
edges: GraphEdge[];
|
||||
}
|
||||
|
||||
/** Normalized visual bucket for a node. */
|
||||
export type NodeKind =
|
||||
| "document"
|
||||
| "concept"
|
||||
| "person"
|
||||
| "org"
|
||||
| "location"
|
||||
| "event"
|
||||
| "tag"
|
||||
| "other";
|
||||
|
||||
export type LayoutType = "force" | "circular";
|
||||
|
||||
/** Graph display mode: full graph vs. click-to-expand exploration. */
|
||||
export type GraphMode = "full" | "explore";
|
||||
|
||||
/** One incremental expansion step in explore mode (for undo / rebuild). */
|
||||
export interface ExploreStep {
|
||||
nodeId: string;
|
||||
depth: number;
|
||||
addedNodeIds: string[];
|
||||
addedEdgeIds: string[];
|
||||
timestamp: number;
|
||||
}
|
||||
|
||||
/** Style metadata of an edge label. */
|
||||
export interface EdgeInfo {
|
||||
color: string;
|
||||
zh: string;
|
||||
dashed: boolean;
|
||||
}
|
||||
|
||||
/** A node inside the force simulation (world coordinates). */
|
||||
export interface SimNode {
|
||||
id: string;
|
||||
node: GraphNode;
|
||||
kind: NodeKind;
|
||||
degree: number;
|
||||
radius: number;
|
||||
color: string;
|
||||
x: number;
|
||||
y: number;
|
||||
vx: number;
|
||||
vy: number;
|
||||
fx: number | null;
|
||||
fy: number | null;
|
||||
}
|
||||
|
||||
export interface SimEdge {
|
||||
id: string;
|
||||
source: SimNode;
|
||||
target: SimNode;
|
||||
label: string;
|
||||
info: EdgeInfo;
|
||||
weight: number;
|
||||
distance: number;
|
||||
}
|
||||
|
||||
export interface Transform {
|
||||
x: number;
|
||||
y: number;
|
||||
k: number;
|
||||
}
|
||||
|
||||
export interface TooltipState {
|
||||
/** Viewport (client) coordinates. */
|
||||
x: number;
|
||||
y: number;
|
||||
node: GraphNode;
|
||||
kind: NodeKind;
|
||||
degree: number;
|
||||
}
|
||||
|
||||
export interface GraphStats {
|
||||
nodes: number;
|
||||
edges: number;
|
||||
avgDegree: number;
|
||||
largestComponent: number;
|
||||
}
|
||||
@@ -1,434 +0,0 @@
|
||||
import type {
|
||||
EdgeInfo,
|
||||
GraphEdge,
|
||||
GraphNode,
|
||||
GraphStats,
|
||||
NodeKind,
|
||||
SimEdge,
|
||||
SimNode,
|
||||
} from "./graphTypes";
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Node helpers
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
export const getNodeTitle = (node: GraphNode): string =>
|
||||
node.title || node.label || node.name || node.id;
|
||||
|
||||
const KIND_PATTERNS: Array<[NodeKind, RegExp]> = [
|
||||
["person", /^(person|people|human|人物|人名|人员)$/i],
|
||||
["org", /^(org|organization|organisation|company|institution|party|组织|机构|公司|团体|政党|国会)$/i],
|
||||
["location", /^(location|place|country|city|region|gpe|地点|地区|国家|城市|地理位置)$/i],
|
||||
["event", /^(event|事件|活动|会议|战争|选举|讲话)$/i],
|
||||
["tag", /^(tag|keyword|标签|关键词)$/i],
|
||||
["concept", /^(concept|topic|term|technology|theory|product|project|主题|概念|术语|技术|理论|产品|项目|政策)$/i],
|
||||
];
|
||||
|
||||
/** Map a raw node (note source_type / entity entity_type) to a visual bucket. */
|
||||
export function getNodeKind(node: GraphNode): NodeKind {
|
||||
const type = (node.type || "").toLowerCase();
|
||||
if (type === "note" || type === "document" || type === "file") return "document";
|
||||
const category = (node.category || "").trim();
|
||||
if (category) {
|
||||
for (const [kind, re] of KIND_PATTERNS) {
|
||||
if (re.test(category)) return kind;
|
||||
}
|
||||
}
|
||||
// Unmatched LLM entity types still read better as "concept" than as a grey blob.
|
||||
if (type === "entity") return category ? "other" : "concept";
|
||||
return "other";
|
||||
}
|
||||
|
||||
export const NODE_KIND_META: Record<NodeKind, { label: string; color: string }> = {
|
||||
document: { label: "文档", color: "#38bdf8" },
|
||||
concept: { label: "主题", color: "#a78bfa" },
|
||||
person: { label: "人物", color: "#fbbf24" },
|
||||
org: { label: "组织", color: "#2dd4bf" },
|
||||
location: { label: "地点", color: "#34d399" },
|
||||
event: { label: "事件", color: "#fb7185" },
|
||||
tag: { label: "标签", color: "#a3e635" },
|
||||
other: { label: "其他", color: "#94a3b8" },
|
||||
};
|
||||
|
||||
/** Stable ordering for filters / legend. */
|
||||
export const NODE_KIND_ORDER: NodeKind[] = [
|
||||
"document",
|
||||
"concept",
|
||||
"person",
|
||||
"org",
|
||||
"location",
|
||||
"event",
|
||||
"tag",
|
||||
"other",
|
||||
];
|
||||
|
||||
export const getNodeRadius = (degree: number): number =>
|
||||
Math.min(30, 9 + Math.sqrt(Math.max(degree, 0)) * 4);
|
||||
|
||||
export function truncateLabel(text: string, max = 14): string {
|
||||
return text.length > max ? `${text.slice(0, max)}…` : text;
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Edge helpers
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
export const getEdgeLabel = (edge: GraphEdge): string =>
|
||||
edge.label || edge.type || edge.relation || "related";
|
||||
|
||||
/** Stable edge identity shared by explore-mode visibility sets and the canvas. */
|
||||
export const getEdgeId = (edge: GraphEdge): string =>
|
||||
edge.id ?? `${edge.source}->${edge.target}:${getEdgeLabel(edge)}`;
|
||||
|
||||
const KNOWN_EDGE_INFO: Record<string, EdgeInfo> = {
|
||||
related: { color: "#94a3b8", zh: "相关", dashed: false },
|
||||
links: { color: "#a78bfa", zh: "引用", dashed: true },
|
||||
references: { color: "#a78bfa", zh: "引用", dashed: true },
|
||||
contains: { color: "#34d399", zh: "包含", dashed: false },
|
||||
belongs_to: { color: "#2dd4bf", zh: "属于", dashed: false },
|
||||
similar: { color: "#fbbf24", zh: "相似", dashed: true },
|
||||
mentions: { color: "#38bdf8", zh: "提及", dashed: false },
|
||||
};
|
||||
|
||||
const EDGE_PALETTE = [
|
||||
"#38bdf8",
|
||||
"#a78bfa",
|
||||
"#34d399",
|
||||
"#fbbf24",
|
||||
"#fb7185",
|
||||
"#2dd4bf",
|
||||
"#fb923c",
|
||||
"#e879f9",
|
||||
];
|
||||
|
||||
function hashString(text: string): number {
|
||||
let h = 0;
|
||||
for (let i = 0; i < text.length; i++) {
|
||||
h = (h * 31 + text.charCodeAt(i)) | 0;
|
||||
}
|
||||
return Math.abs(h);
|
||||
}
|
||||
|
||||
/** Color / display name / line style of an edge label (stable for unknown labels). */
|
||||
export function getEdgeInfo(label: string): EdgeInfo {
|
||||
const known = KNOWN_EDGE_INFO[label.toLowerCase()];
|
||||
if (known) return known;
|
||||
return {
|
||||
color: EDGE_PALETTE[hashString(label) % EDGE_PALETTE.length],
|
||||
zh: label,
|
||||
dashed: false,
|
||||
};
|
||||
}
|
||||
|
||||
export function getLinkDistance(label: string): number {
|
||||
switch (label.toLowerCase()) {
|
||||
case "contains":
|
||||
return 100;
|
||||
case "links":
|
||||
case "references":
|
||||
return 130;
|
||||
case "similar":
|
||||
return 170;
|
||||
default:
|
||||
return 145;
|
||||
}
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Graph statistics
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
export function computeDegrees(
|
||||
nodes: GraphNode[],
|
||||
edges: GraphEdge[],
|
||||
): Map<string, number> {
|
||||
const degrees = new Map<string, number>();
|
||||
for (const n of nodes) degrees.set(n.id, 0);
|
||||
for (const e of edges) {
|
||||
if (degrees.has(e.source)) degrees.set(e.source, (degrees.get(e.source) ?? 0) + 1);
|
||||
if (degrees.has(e.target)) degrees.set(e.target, (degrees.get(e.target) ?? 0) + 1);
|
||||
}
|
||||
return degrees;
|
||||
}
|
||||
|
||||
export function computeGraphStats(
|
||||
nodes: GraphNode[],
|
||||
edges: GraphEdge[],
|
||||
): GraphStats {
|
||||
const adjacency = new Map<string, string[]>();
|
||||
for (const n of nodes) adjacency.set(n.id, []);
|
||||
let validEdges = 0;
|
||||
for (const e of edges) {
|
||||
if (!adjacency.has(e.source) || !adjacency.has(e.target)) continue;
|
||||
validEdges += 1;
|
||||
adjacency.get(e.source)!.push(e.target);
|
||||
adjacency.get(e.target)!.push(e.source);
|
||||
}
|
||||
// Largest connected component via BFS.
|
||||
const visited = new Set<string>();
|
||||
let largest = 0;
|
||||
for (const n of nodes) {
|
||||
if (visited.has(n.id)) continue;
|
||||
let size = 0;
|
||||
const queue = [n.id];
|
||||
visited.add(n.id);
|
||||
while (queue.length) {
|
||||
const cur = queue.pop()!;
|
||||
size += 1;
|
||||
for (const next of adjacency.get(cur) ?? []) {
|
||||
if (!visited.has(next)) {
|
||||
visited.add(next);
|
||||
queue.push(next);
|
||||
}
|
||||
}
|
||||
}
|
||||
largest = Math.max(largest, size);
|
||||
}
|
||||
return {
|
||||
nodes: nodes.length,
|
||||
edges: validEdges,
|
||||
avgDegree: nodes.length ? (validEdges * 2) / nodes.length : 0,
|
||||
largestComponent: largest,
|
||||
};
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Layouts
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
const GOLDEN_ANGLE = 2.3999632297286533;
|
||||
|
||||
/** Phyllotaxis seeding around the origin — hubs (high degree) near the center. */
|
||||
export function seedPositions(nodes: SimNode[]): void {
|
||||
const sorted = [...nodes].sort((a, b) => b.degree - a.degree);
|
||||
sorted.forEach((n, i) => {
|
||||
const r = 22 * Math.sqrt(i + 0.5);
|
||||
const angle = i * GOLDEN_ANGLE;
|
||||
n.x = r * Math.cos(angle);
|
||||
n.y = r * Math.sin(angle);
|
||||
n.vx = 0;
|
||||
n.vy = 0;
|
||||
n.fx = null;
|
||||
n.fy = null;
|
||||
});
|
||||
}
|
||||
|
||||
/** Static circular layout, grouped by kind then degree. */
|
||||
export function applyCircularLayout(nodes: SimNode[]): void {
|
||||
const sorted = [...nodes].sort((a, b) => {
|
||||
const ka = NODE_KIND_ORDER.indexOf(a.kind);
|
||||
const kb = NODE_KIND_ORDER.indexOf(b.kind);
|
||||
return ka !== kb ? ka - kb : b.degree - a.degree;
|
||||
});
|
||||
const radius = Math.max(180, (sorted.length * 56) / (2 * Math.PI));
|
||||
sorted.forEach((n, i) => {
|
||||
const angle = (i / sorted.length) * Math.PI * 2 - Math.PI / 2;
|
||||
n.x = radius * Math.cos(angle);
|
||||
n.y = radius * Math.sin(angle);
|
||||
n.vx = 0;
|
||||
n.vy = 0;
|
||||
n.fx = null;
|
||||
n.fy = null;
|
||||
});
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Force simulation (d3-force-like, dependency-free, fine for <500 nodes)
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
export class ForceSimulation {
|
||||
alpha = 1;
|
||||
readonly alphaMin = 0.004;
|
||||
private readonly alphaDecay: number;
|
||||
private readonly velocityDecay = 0.38;
|
||||
private readonly chargeStrength: number;
|
||||
private readonly chargeCutoff: number;
|
||||
private readonly chargeCutoff2: number;
|
||||
/** Hard cap on per-tick movement so dense graphs can't oscillate or fly off. */
|
||||
private readonly maxSpeed = 30;
|
||||
|
||||
constructor(
|
||||
public nodes: SimNode[],
|
||||
public edges: SimEdge[],
|
||||
) {
|
||||
// Scale forces down and cool faster as the graph grows — with the
|
||||
// small-graph constants a few hundred nodes keep shoving each other
|
||||
// around for minutes and the layout never settles.
|
||||
const n = nodes.length;
|
||||
if (n > 600) {
|
||||
this.chargeStrength = -160;
|
||||
this.chargeCutoff = 260;
|
||||
this.alphaDecay = 0.05;
|
||||
} else if (n > 300) {
|
||||
this.chargeStrength = -240;
|
||||
this.chargeCutoff = 340;
|
||||
this.alphaDecay = 0.038;
|
||||
} else if (n > 150) {
|
||||
this.chargeStrength = -320;
|
||||
this.chargeCutoff = 440;
|
||||
this.alphaDecay = 0.03;
|
||||
} else {
|
||||
this.chargeStrength = -420;
|
||||
this.chargeCutoff = 600;
|
||||
this.alphaDecay = 0.0228;
|
||||
}
|
||||
this.chargeCutoff2 = this.chargeCutoff * this.chargeCutoff;
|
||||
}
|
||||
|
||||
reheat(alpha = 0.9): void {
|
||||
this.alpha = Math.max(this.alpha, alpha);
|
||||
}
|
||||
|
||||
/** Advance one step; returns false when the layout has cooled down. */
|
||||
tick(): boolean {
|
||||
if (this.alpha < this.alphaMin) return false;
|
||||
this.alpha += (0 - this.alpha) * this.alphaDecay;
|
||||
const a = this.alpha;
|
||||
const { nodes, edges } = this;
|
||||
|
||||
// Link spring force.
|
||||
for (const e of edges) {
|
||||
const s = e.source;
|
||||
const t = e.target;
|
||||
let dx = t.x - s.x;
|
||||
let dy = t.y - s.y;
|
||||
const dist = Math.sqrt(dx * dx + dy * dy) || 1e-3;
|
||||
const f = ((dist - e.distance) / dist) * 0.45 * a;
|
||||
dx *= f;
|
||||
dy *= f;
|
||||
t.vx -= dx * 0.5;
|
||||
t.vy -= dy * 0.5;
|
||||
s.vx += dx * 0.5;
|
||||
s.vy += dy * 0.5;
|
||||
}
|
||||
|
||||
// Repulsion + collision over nearby pairs only (uniform grid).
|
||||
this.applyPairForces(a);
|
||||
|
||||
// Gentle pull toward the origin keeps disconnected clusters on screen.
|
||||
for (const n of nodes) {
|
||||
n.vx += (0 - n.x) * 0.05 * a;
|
||||
n.vy += (0 - n.y) * 0.05 * a;
|
||||
}
|
||||
|
||||
// Integrate velocities.
|
||||
for (let i = 0; i < nodes.length; i++) {
|
||||
const n = nodes[i];
|
||||
if (n.fx != null && n.fy != null) {
|
||||
n.x = n.fx;
|
||||
n.y = n.fy;
|
||||
n.vx = 0;
|
||||
n.vy = 0;
|
||||
} else {
|
||||
n.vx *= 1 - this.velocityDecay;
|
||||
n.vy *= 1 - this.velocityDecay;
|
||||
const speed = Math.sqrt(n.vx * n.vx + n.vy * n.vy);
|
||||
if (speed > this.maxSpeed) {
|
||||
const s = this.maxSpeed / speed;
|
||||
n.vx *= s;
|
||||
n.vy *= s;
|
||||
}
|
||||
n.x += n.vx;
|
||||
n.y += n.vy;
|
||||
}
|
||||
if (!Number.isFinite(n.x) || !Number.isFinite(n.y)) {
|
||||
// A NaN/Infinity coordinate would blank the whole canvas (the
|
||||
// bounding box and fit transform become NaN); park the node instead.
|
||||
const r = 40 + 18 * Math.sqrt(i + 1);
|
||||
n.x = r * Math.cos(i * GOLDEN_ANGLE);
|
||||
n.y = r * Math.sin(i * GOLDEN_ANGLE);
|
||||
n.vx = 0;
|
||||
n.vy = 0;
|
||||
}
|
||||
}
|
||||
return true;
|
||||
}
|
||||
|
||||
/**
|
||||
* Many-body repulsion + collision resolution, restricted to pairs within
|
||||
* one grid cell (cell size = charge cutoff). Two nodes farther apart than
|
||||
* the cutoff never interact, so binning by cell keeps the scan close to
|
||||
* O(n) for spread-out layouts instead of O(n²) over all pairs.
|
||||
*/
|
||||
private applyPairForces(a: number): void {
|
||||
const cellSize = this.chargeCutoff;
|
||||
const grid = new Map<string, { cx: number; cy: number; items: SimNode[] }>();
|
||||
for (const n of this.nodes) {
|
||||
const cx = Math.floor(n.x / cellSize);
|
||||
const cy = Math.floor(n.y / cellSize);
|
||||
const key = `${cx},${cy}`;
|
||||
let bucket = grid.get(key);
|
||||
if (!bucket) grid.set(key, (bucket = { cx, cy, items: [] }));
|
||||
bucket.items.push(n);
|
||||
}
|
||||
// Half of the 8-neighborhood so every cell pair is visited exactly once.
|
||||
const offsets = [
|
||||
[0, 0],
|
||||
[1, 0],
|
||||
[0, 1],
|
||||
[1, 1],
|
||||
[1, -1],
|
||||
] as const;
|
||||
for (const bucket of grid.values()) {
|
||||
for (const [ox, oy] of offsets) {
|
||||
const other =
|
||||
ox === 0 && oy === 0
|
||||
? bucket
|
||||
: grid.get(`${bucket.cx + ox},${bucket.cy + oy}`);
|
||||
if (!other) continue;
|
||||
const left = bucket.items;
|
||||
const right = other.items;
|
||||
if (other === bucket) {
|
||||
for (let i = 0; i < left.length; i++) {
|
||||
for (let j = i + 1; j < left.length; j++) {
|
||||
this.pairForce(left[i], left[j], i, j, a);
|
||||
}
|
||||
}
|
||||
} else {
|
||||
for (let i = 0; i < left.length; i++) {
|
||||
for (let j = 0; j < right.length; j++) {
|
||||
this.pairForce(left[i], right[j], i, j, a);
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
private pairForce(ni: SimNode, nj: SimNode, i: number, j: number, a: number): void {
|
||||
let dx = nj.x - ni.x;
|
||||
let dy = nj.y - ni.y;
|
||||
let d2 = dx * dx + dy * dy;
|
||||
if (d2 > this.chargeCutoff2) return;
|
||||
if (d2 < 1) {
|
||||
// Deterministic nudge for coincident nodes.
|
||||
dx = ((i % 3) - 1) * 0.5 || 0.5;
|
||||
dy = ((j % 3) - 1) * 0.5 || -0.5;
|
||||
d2 = dx * dx + dy * dy;
|
||||
}
|
||||
const d = Math.sqrt(d2);
|
||||
const f = (this.chargeStrength * a) / d2;
|
||||
const fx = (dx / d) * f;
|
||||
const fy = (dy / d) * f;
|
||||
ni.vx += fx;
|
||||
ni.vy += fy;
|
||||
nj.vx -= fx;
|
||||
nj.vy -= fy;
|
||||
|
||||
// Positional collision resolution (prevents node overlap).
|
||||
const minDist = ni.radius + nj.radius + 8;
|
||||
if (d < minDist) {
|
||||
const push = ((minDist - d) / d) * 0.5;
|
||||
const px = dx * push;
|
||||
const py = dy * push;
|
||||
if (ni.fx == null) {
|
||||
ni.x -= px;
|
||||
ni.y -= py;
|
||||
}
|
||||
if (nj.fx == null) {
|
||||
nj.x += px;
|
||||
nj.y += py;
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -508,9 +508,8 @@ export default function ProgressiveSkillsAnimation() {
|
||||
"deploying",
|
||||
"done",
|
||||
].includes(phase) && (
|
||||
<div className="flex items-center gap-2 text-base text-zinc-300">
|
||||
<Sparkles size={16} className="text-purple-400" />
|
||||
<span>Found 3 skills</span>
|
||||
<div className="text-base text-zinc-300">
|
||||
<span className="text-purple-400">✨</span> Found 3 skills
|
||||
</div>
|
||||
)}
|
||||
|
||||
@@ -527,9 +526,8 @@ export default function ProgressiveSkillsAnimation() {
|
||||
].includes(phase) && (
|
||||
<div className="mt-4">
|
||||
<hr className="mb-3 border-zinc-700" />
|
||||
<div className="mb-3 inline-flex items-center gap-2 text-zinc-300">
|
||||
<Search size={16} className="text-blue-400" />
|
||||
<span>Researching...</span>
|
||||
<div className="mb-3 text-zinc-300">
|
||||
🔬 Researching...
|
||||
</div>
|
||||
<div className="mb-3 space-y-2">
|
||||
{/* Loading SKILL.md */}
|
||||
@@ -625,10 +623,7 @@ export default function ProgressiveSkillsAnimation() {
|
||||
className="mt-4"
|
||||
>
|
||||
<hr className="mb-3 border-zinc-700" />
|
||||
<div className="mb-3 inline-flex items-center gap-2 text-zinc-300">
|
||||
<FileText size={16} className="text-zinc-400" />
|
||||
<span>Building...</span>
|
||||
</div>
|
||||
<div className="mb-3 text-zinc-300">🔨 Building...</div>
|
||||
<div className="mb-3 flex items-center gap-2 pl-4 text-zinc-400">
|
||||
<FileText size={16} />
|
||||
<span>Loading frontend-design/SKILL.md...</span>
|
||||
@@ -658,10 +653,7 @@ export default function ProgressiveSkillsAnimation() {
|
||||
className="mt-4"
|
||||
>
|
||||
<hr className="mb-3 border-zinc-700" />
|
||||
<div className="mb-3 inline-flex items-center gap-2 text-zinc-300">
|
||||
<Terminal size={16} className="text-zinc-400" />
|
||||
<span>Deploying...</span>
|
||||
</div>
|
||||
<div className="mb-3 text-zinc-300">🚀 Deploying...</div>
|
||||
<div className="mb-3 space-y-2">
|
||||
<div className="flex items-center gap-2 pl-4 text-zinc-400">
|
||||
<FileText size={16} />
|
||||
@@ -684,9 +676,8 @@ export default function ProgressiveSkillsAnimation() {
|
||||
animate={{ opacity: 1, scale: 1 }}
|
||||
className="mt-4 rounded-xl border border-green-500/30 bg-green-500/10 p-4"
|
||||
>
|
||||
<div className="inline-flex items-center gap-2 text-lg font-medium text-green-500">
|
||||
<Check size={18} />
|
||||
<span>Live at biotech-startup.vercel.app</span>
|
||||
<div className="text-lg font-medium text-green-500">
|
||||
✅ Live at biotech-startup.vercel.app
|
||||
</div>
|
||||
</motion.div>
|
||||
)}
|
||||
|
||||
@@ -35,27 +35,27 @@ export function SandboxSection({ className }: { className?: string }) {
|
||||
$ pip install -r requirements.txt
|
||||
</TypingAnimation>
|
||||
<AnimatedSpan delay={2000} className="text-green-500">
|
||||
[ok] Installed pygame
|
||||
✔ Installed pygame
|
||||
</AnimatedSpan>
|
||||
|
||||
<TypingAnimation delay={2400}>
|
||||
$ write game.py --lines 156
|
||||
</TypingAnimation>
|
||||
<AnimatedSpan delay={3200} className="text-blue-500">
|
||||
[ok] Written 156 lines
|
||||
✔ Written 156 lines
|
||||
</AnimatedSpan>
|
||||
|
||||
<TypingAnimation delay={3600}>
|
||||
$ python game.py --test
|
||||
</TypingAnimation>
|
||||
<AnimatedSpan delay={4200} className="text-green-500">
|
||||
[ok] All sprites loaded
|
||||
✔ All sprites loaded
|
||||
</AnimatedSpan>
|
||||
<AnimatedSpan delay={4500} className="text-green-500">
|
||||
[ok] Physics engine OK
|
||||
✔ Physics engine OK
|
||||
</AnimatedSpan>
|
||||
<AnimatedSpan delay={4800} className="text-green-500">
|
||||
[ok] 60 FPS stable
|
||||
✔ 60 FPS stable
|
||||
</AnimatedSpan>
|
||||
|
||||
{/* Scene 2: Data Analysis */}
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
.sidebar-contnet {
|
||||
position: absolute;
|
||||
position: fixed;
|
||||
left: 0;
|
||||
top: 4rem;
|
||||
z-index: 40;
|
||||
@@ -43,8 +43,8 @@
|
||||
width: 100%;
|
||||
height: 75px;
|
||||
margin-top: 13px;
|
||||
font-family: "Source Han Sans SC", "Noto Sans SC", "Noto Sans S Chinese", "Microsoft YaHei", sans-serif;
|
||||
font-weight: 700;
|
||||
font-family: Noto Sans S Chinese;
|
||||
font-weight: 500;
|
||||
color: #fff;
|
||||
font-size: 20px;
|
||||
display: flex;
|
||||
@@ -111,42 +111,6 @@
|
||||
word-break: break-all;
|
||||
}
|
||||
|
||||
.sidebar-contnet--top-submenu .sidebar-contents {
|
||||
overflow-y: auto;
|
||||
overflow-x: hidden;
|
||||
}
|
||||
|
||||
.sidebar-contnet--top-submenu .item-nav {
|
||||
height: auto;
|
||||
min-height: 75px;
|
||||
padding: 10px 2px;
|
||||
font-size: 19px;
|
||||
line-height: 1.35;
|
||||
}
|
||||
|
||||
.sidebar-contnet--top-submenu .item-nav__label {
|
||||
width: 3.1em;
|
||||
max-width: 100%;
|
||||
padding-left: 0;
|
||||
letter-spacing: 0;
|
||||
line-height: 1.35;
|
||||
text-align: center;
|
||||
word-break: normal;
|
||||
overflow-wrap: normal;
|
||||
white-space: normal;
|
||||
}
|
||||
|
||||
.sidebar-contnet--top-submenu .item-nav__label--four-char {
|
||||
width: 2.2em;
|
||||
}
|
||||
|
||||
.sidebar-contnet--top-submenu .item-nav:hover::after,
|
||||
.sidebar-contnet--top-submenu .item-nav-active::after,
|
||||
.sidebar-contnet--top-submenu .item-nav-hovered::after {
|
||||
top: auto;
|
||||
bottom: 0;
|
||||
}
|
||||
|
||||
.item-nav-content {
|
||||
width: 155px;
|
||||
height: calc(100vh - 4rem - 20px);
|
||||
@@ -155,8 +119,7 @@
|
||||
border: 1px solid;
|
||||
border-color: rgba(43, 143, 254, 0.6);
|
||||
border-radius: 4px;
|
||||
backdrop-filter: blur(20px);
|
||||
-webkit-backdrop-filter: blur(20px);
|
||||
backdrop-filter: blur(2px);
|
||||
position: absolute;
|
||||
left: 108px;
|
||||
top: 0;
|
||||
@@ -230,10 +193,10 @@
|
||||
border: 0;
|
||||
background: transparent;
|
||||
padding: 3px 0;
|
||||
font-family: "Source Han Sans SC", "Noto Sans SC", "Noto Sans S Chinese", "Microsoft YaHei", sans-serif;
|
||||
font-family: Noto Sans S Chinese;
|
||||
font-size: 18px;
|
||||
line-height: 16px;
|
||||
font-weight: 700;
|
||||
font-weight: 500;
|
||||
letter-spacing: 0;
|
||||
color: rgba(255, 255, 255, 0.8);
|
||||
text-align: left;
|
||||
@@ -356,390 +319,4 @@
|
||||
max-width: 64px;
|
||||
opacity: 1;
|
||||
transform: translateX(0);
|
||||
}
|
||||
|
||||
.sidebar-contnet--top-submenu .item-nav:hover::before,
|
||||
.sidebar-contnet--top-submenu .item-nav-active::before,
|
||||
.sidebar-contnet--top-submenu .item-nav-hovered::before {
|
||||
display: none;
|
||||
}
|
||||
|
||||
.top-submenu-bar {
|
||||
position: relative;
|
||||
z-index: 35;
|
||||
height: 64px;
|
||||
min-width: 0;
|
||||
flex: 1 1 auto;
|
||||
align-self: stretch;
|
||||
margin-left: 34px;
|
||||
background: transparent;
|
||||
border: 0;
|
||||
box-shadow: none;
|
||||
overflow-x: auto;
|
||||
overflow-y: hidden;
|
||||
}
|
||||
|
||||
/* 左右分组：不撑满、不裁切选中高光 */
|
||||
.top-submenu-bar--group {
|
||||
flex: 0 0 auto;
|
||||
margin-left: 0;
|
||||
overflow: visible;
|
||||
}
|
||||
|
||||
.top-submenu-list {
|
||||
min-width: max-content;
|
||||
height: 64px;
|
||||
display: flex;
|
||||
align-items: stretch;
|
||||
gap: 0;
|
||||
padding: 0;
|
||||
}
|
||||
|
||||
.top-submenu-item {
|
||||
position: relative;
|
||||
min-width: 122px;
|
||||
height: 64px;
|
||||
display: inline-flex;
|
||||
align-items: center;
|
||||
justify-content: center;
|
||||
border: 0;
|
||||
padding: 0 18px;
|
||||
background: transparent;
|
||||
color: rgba(255, 255, 255, 0.8);
|
||||
font-family: "Source Han Sans SC", "Noto Sans SC", "Noto Sans S Chinese", "Microsoft YaHei", sans-serif;
|
||||
font-size: 18px;
|
||||
line-height: 1;
|
||||
font-weight: 700;
|
||||
letter-spacing: 0;
|
||||
white-space: nowrap;
|
||||
cursor: pointer;
|
||||
overflow: visible;
|
||||
transition: color 0.18s ease, text-shadow 0.18s ease;
|
||||
}
|
||||
|
||||
.top-submenu-item:hover {
|
||||
color: #9fe3ff;
|
||||
text-shadow: 0 0 10px rgba(91, 210, 255, 0.4);
|
||||
}
|
||||
|
||||
/* 选中态：设计稿提供的 nav-active.svg 发光高光 + 青色文字 */
|
||||
.top-submenu-item--active {
|
||||
color: #64cdf5;
|
||||
text-shadow: 0 0 12px rgba(100, 205, 245, 0.5);
|
||||
background-image: url("../assets/img/nav-active.svg");
|
||||
background-repeat: no-repeat;
|
||||
background-position: center calc(50% + 7px);
|
||||
background-size: 152px 42px;
|
||||
}
|
||||
|
||||
.third-submenu-bar {
|
||||
position: relative;
|
||||
z-index: 34;
|
||||
height: 42px;
|
||||
margin-left: 118px;
|
||||
flex-shrink: 0;
|
||||
background:
|
||||
linear-gradient(180deg, rgba(4, 34, 78, 0.96) 0%, rgba(2, 24, 58, 0.96) 100%),
|
||||
radial-gradient(ellipse at top left, rgba(45, 149, 255, 0.18) 0%, rgba(45, 149, 255, 0) 42%);
|
||||
border-top: 1px solid rgba(79, 154, 255, 0.18);
|
||||
border-bottom: 1px solid rgba(43, 143, 254, 0.26);
|
||||
box-shadow: inset 0 1px 0 rgba(115, 194, 255, 0.08), inset 0 -10px 18px rgba(0, 26, 70, 0.22);
|
||||
overflow-x: auto;
|
||||
overflow-y: hidden;
|
||||
}
|
||||
|
||||
.third-submenu-list {
|
||||
min-width: max-content;
|
||||
height: 100%;
|
||||
display: flex;
|
||||
align-items: center;
|
||||
gap: 10px;
|
||||
padding: 0 24px;
|
||||
}
|
||||
|
||||
.third-submenu-item {
|
||||
position: relative;
|
||||
height: 28px;
|
||||
min-width: 86px;
|
||||
display: inline-flex;
|
||||
align-items: center;
|
||||
justify-content: center;
|
||||
border: 1px solid rgba(90, 169, 255, 0.2);
|
||||
border-radius: 2px;
|
||||
padding: 0 16px;
|
||||
background: rgba(2, 28, 65, 0.36);
|
||||
color: rgba(221, 242, 255, 0.78);
|
||||
font-family: "Source Han Sans SC", "Noto Sans SC", "Noto Sans S Chinese", "Microsoft YaHei", sans-serif;
|
||||
font-size: 14px;
|
||||
line-height: 1;
|
||||
font-weight: 700;
|
||||
letter-spacing: 0;
|
||||
white-space: nowrap;
|
||||
cursor: pointer;
|
||||
transition: color 0.18s ease, border-color 0.18s ease, background 0.18s ease, box-shadow 0.18s ease;
|
||||
}
|
||||
|
||||
.third-submenu-item:hover,
|
||||
.third-submenu-item--active {
|
||||
color: #ffffff;
|
||||
border-color: rgba(82, 205, 255, 0.55);
|
||||
background: linear-gradient(180deg, rgba(14, 84, 165, 0.54) 0%, rgba(4, 45, 104, 0.58) 100%);
|
||||
box-shadow: inset 0 0 12px rgba(0, 185, 255, 0.14), 0 0 10px rgba(35, 167, 255, 0.14);
|
||||
text-shadow: 0 0 8px rgba(87, 215, 255, 0.34);
|
||||
}
|
||||
|
||||
.third-submenu-item--active::after {
|
||||
content: '';
|
||||
position: absolute;
|
||||
left: 12px;
|
||||
right: 12px;
|
||||
bottom: -1px;
|
||||
height: 2px;
|
||||
background: linear-gradient(90deg, rgba(62, 178, 255, 0) 0%, #4ee6ff 50%, rgba(62, 178, 255, 0) 100%);
|
||||
box-shadow: 0 0 10px rgba(78, 230, 255, 0.72);
|
||||
}
|
||||
|
||||
/* ══════════════════════════════════════════════════════════════════════
|
||||
浅色主题导航（红/蓝两套，颜色经 --brand-* 变量自动切换）
|
||||
逻辑同深色：一级导航在顶栏、二级导航在侧边栏、三级导航在顶栏下方横条。
|
||||
但一级导航不做深色那种围绕基座的特殊处理，正常横排即可。
|
||||
══════════════════════════════════════════════════════════════════════ */
|
||||
|
||||
/* ── 顶栏一级导航（白字落在 red/blue-header banner 上）──────────────────── */
|
||||
.light-topnav {
|
||||
display: flex;
|
||||
align-items: center;
|
||||
height: 64px;
|
||||
min-width: 0;
|
||||
gap: 6px;
|
||||
/* 整体往右移（用百分比，随容器宽度自适应；约等于宽屏下的 200px 量级） */
|
||||
margin-left: 10%;
|
||||
overflow-x: auto;
|
||||
overflow-y: hidden;
|
||||
}
|
||||
|
||||
.light-topnav-item {
|
||||
position: relative;
|
||||
flex: 0 0 auto;
|
||||
display: inline-flex;
|
||||
align-items: center;
|
||||
justify-content: center;
|
||||
height: 60px;
|
||||
padding: 0 18px;
|
||||
border: 0;
|
||||
background: transparent;
|
||||
color: rgba(255, 255, 255, 0.85);
|
||||
/* 复用全局字体栈（与「多智能体会商」等页面一致）：Source Han Sans SC 在 Windows 上
|
||||
常按字重分家族，放栈首会导致 font-weight 选不到粗体；用 --font-sans（Noto Sans SC
|
||||
领先）才能正常加粗。 */
|
||||
font-family: var(--font-sans);
|
||||
font-size: 16px;
|
||||
/* 普通项用 500（雅黑兜底下渲染为常规字面），选中项 700 才能看出粗体对比 */
|
||||
font-weight: 500;
|
||||
line-height: 1;
|
||||
white-space: nowrap;
|
||||
cursor: pointer;
|
||||
transition: color 0.18s ease;
|
||||
}
|
||||
|
||||
.light-topnav-item:hover {
|
||||
color: #ffffff;
|
||||
}
|
||||
|
||||
/* 选中态：不改背景，文字加粗变亮 + 贴近导航栏底部一条粗白线 */
|
||||
.light-topnav-item--active,
|
||||
.light-topnav-item--active:hover {
|
||||
color: #ffffff;
|
||||
font-weight: 700;
|
||||
}
|
||||
|
||||
.light-topnav-item--active::after {
|
||||
content: '';
|
||||
position: absolute;
|
||||
left: 12px;
|
||||
right: 12px;
|
||||
bottom: 0;
|
||||
height: 4px;
|
||||
border-radius: 3px 3px 0 0;
|
||||
background: #ffffff;
|
||||
}
|
||||
|
||||
/* ── 侧边栏：二级图标卡片列（96px）+ 右侧三级列（展开时整体变宽，带过渡）──────── */
|
||||
.light-sidebar {
|
||||
position: fixed;
|
||||
left: 0;
|
||||
top: 4rem;
|
||||
z-index: 40;
|
||||
width: 96px;
|
||||
height: calc(100vh - 4rem);
|
||||
display: flex;
|
||||
flex-direction: row;
|
||||
background: #ffffff;
|
||||
border-right: 1px solid var(--border);
|
||||
box-sizing: border-box;
|
||||
overflow: hidden;
|
||||
/* 展开三级列时整体变宽的过渡动画 */
|
||||
transition: width 0.3s ease;
|
||||
}
|
||||
|
||||
/* 展开三级列：96(二级) + 140(三级) = 236px */
|
||||
.light-sidebar--expanded {
|
||||
width: 236px;
|
||||
}
|
||||
|
||||
/* 二级列：固定 96px，竖排图标卡片 + 底部设置 */
|
||||
.light-sidebar__primary {
|
||||
flex: 0 0 96px;
|
||||
width: 96px;
|
||||
height: 100%;
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
/* 横向 padding 置 0：让选中态浅红背景能铺满与列宽一致 */
|
||||
padding: 10px 0;
|
||||
box-sizing: border-box;
|
||||
}
|
||||
|
||||
/* 三级列：固定 140px，随侧边栏展开淡入 */
|
||||
.light-sidebar__secondary {
|
||||
flex: 0 0 140px;
|
||||
width: 140px;
|
||||
height: 100%;
|
||||
border-left: 1px solid var(--border);
|
||||
/* 横向 padding 置 0：让三级选中背景能铺满与列宽一致（文字内缩由 subitem 自身 padding 保证） */
|
||||
padding: 12px 0;
|
||||
box-sizing: border-box;
|
||||
overflow-y: auto;
|
||||
overflow-x: hidden;
|
||||
background: #ffffff;
|
||||
opacity: 0;
|
||||
transition: opacity 0.2s ease;
|
||||
}
|
||||
|
||||
.light-sidebar--expanded .light-sidebar__secondary {
|
||||
opacity: 1;
|
||||
}
|
||||
|
||||
.light-sidebar__list {
|
||||
flex: 1;
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
gap: 6px;
|
||||
overflow-y: auto;
|
||||
overflow-x: hidden;
|
||||
}
|
||||
|
||||
.light-sidebar-item {
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
align-items: center;
|
||||
justify-content: center;
|
||||
gap: 8px;
|
||||
width: 100%;
|
||||
padding: 14px 4px;
|
||||
border: 0;
|
||||
border-radius: 0;
|
||||
background: transparent;
|
||||
color: #222222;
|
||||
/* 复用全局字体栈，确保加粗生效（详见 .light-topnav-item 注释） */
|
||||
font-family: var(--font-sans);
|
||||
font-size: 16px;
|
||||
font-weight: 500;
|
||||
line-height: 1.3;
|
||||
text-align: center;
|
||||
cursor: pointer;
|
||||
transition: background-color 0.18s ease, color 0.18s ease, box-shadow 0.18s ease;
|
||||
}
|
||||
|
||||
.light-sidebar-item__icon {
|
||||
flex-shrink: 0;
|
||||
/* 图标统一用主题色（红主题红、蓝主题蓝） */
|
||||
color: var(--brand-900);
|
||||
transition: color 0.18s ease, transform 0.18s ease;
|
||||
}
|
||||
|
||||
.light-sidebar-item__label {
|
||||
display: block;
|
||||
width: 100%;
|
||||
word-break: break-word;
|
||||
}
|
||||
|
||||
.light-sidebar-item:hover {
|
||||
background: var(--brand-200);
|
||||
}
|
||||
|
||||
.light-sidebar-item:hover .light-sidebar-item__icon {
|
||||
color: var(--brand-900);
|
||||
transform: scale(1.05);
|
||||
}
|
||||
|
||||
.light-sidebar-item--active {
|
||||
background: var(--brand-200);
|
||||
color: var(--brand-900);
|
||||
font-weight: 600;
|
||||
/* 选中态：去圆角去边框，浅红底铺满父元素宽度，右侧一条粗主题色边（用 inset 阴影避免挤压内容） */
|
||||
box-shadow: inset -4px 0 0 0 var(--brand-900);
|
||||
}
|
||||
|
||||
.light-sidebar-item--active .light-sidebar-item__icon {
|
||||
color: var(--brand-900);
|
||||
}
|
||||
|
||||
/* 侧边栏底部：设置入口（浅色模式不支持收起，故无收起按钮） */
|
||||
.light-sidebar__footer {
|
||||
margin-top: auto;
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
gap: 4px;
|
||||
padding-top: 8px;
|
||||
border-top: 1px solid var(--border);
|
||||
}
|
||||
|
||||
/* ── 三级列条目（左对齐图标 + 文字，参考 xiugai-1.png）─────────────────────── */
|
||||
.light-sidebar-subnav {
|
||||
display: flex;
|
||||
flex-direction: column;
|
||||
gap: 2px;
|
||||
}
|
||||
|
||||
.light-sidebar-subitem {
|
||||
position: relative;
|
||||
display: flex;
|
||||
align-items: center;
|
||||
gap: 8px;
|
||||
width: 100%;
|
||||
padding: 10px;
|
||||
border: 0;
|
||||
background: transparent;
|
||||
color: #222222;
|
||||
font-family: var(--font-sans);
|
||||
font-size: 14px;
|
||||
font-weight: 500;
|
||||
line-height: 1.3;
|
||||
text-align: left;
|
||||
cursor: pointer;
|
||||
transition: background-color 0.18s ease, color 0.18s ease, box-shadow 0.18s ease;
|
||||
}
|
||||
|
||||
.light-sidebar-subitem__icon {
|
||||
flex-shrink: 0;
|
||||
/* 图标用主题色（红主题红、蓝主题蓝） */
|
||||
color: var(--brand-900);
|
||||
}
|
||||
|
||||
.light-sidebar-subitem__label {
|
||||
flex: 1;
|
||||
overflow: hidden;
|
||||
white-space: nowrap;
|
||||
text-overflow: ellipsis;
|
||||
}
|
||||
|
||||
.light-sidebar-subitem:hover {
|
||||
background: var(--brand-200);
|
||||
}
|
||||
|
||||
.light-sidebar-subitem--active {
|
||||
background: var(--brand-200);
|
||||
color: var(--brand-900);
|
||||
font-weight: 700;
|
||||
box-shadow: inset -4px 0 0 0 var(--brand-900);
|
||||
}
|
||||
}
|
||||
@@ -1,16 +1,14 @@
|
||||
import { useEffect, useRef, useState } from "react";
|
||||
import { useEffect, useMemo, useRef, useState } from "react";
|
||||
import { useNavigate, useLocation } from "react-router-dom";
|
||||
import { PanelLeftOpen, PanelLeftClose } from "lucide-react";
|
||||
|
||||
import { getStoredAuth } from "@/core/auth";
|
||||
import { isWujieEnv } from "@/core/embed/wujie";
|
||||
import {
|
||||
defaultExpandedIds,
|
||||
menuItemContainsPath,
|
||||
menuItemMatchesPath,
|
||||
filterMenuItemsByRole,
|
||||
sidebarMenuItems,
|
||||
type MenuItem,
|
||||
} from "@/core/page-layout/sidebar-menu";
|
||||
import { useSidebarMenuItems } from "@/core/page-layout/use-sidebar-menu";
|
||||
import './page-sdebar.css'
|
||||
|
||||
type PageSidebarDarkProps = {
|
||||
@@ -21,8 +19,10 @@ type PageSidebarDarkProps = {
|
||||
|
||||
export function PageSidebarDark({ collapsed, onToggleCollapsed }: PageSidebarDarkProps) {
|
||||
const isAdmin = getStoredAuth()?.system_role === "admin";
|
||||
const visibleMenuItems = useSidebarMenuItems(!!isAdmin);
|
||||
const inWujieEnv = isWujieEnv();
|
||||
const visibleMenuItems = useMemo(
|
||||
() => filterMenuItemsByRole(sidebarMenuItems, !!isAdmin),
|
||||
[isAdmin],
|
||||
);
|
||||
// defaultExpandedIds 是共享的 ReadonlySet，复制一份给 useState（避免外部突变）
|
||||
const [expandedIds, setExpandedIds] = useState<Set<string>>(() => new Set(defaultExpandedIds));
|
||||
const navigate = useNavigate();
|
||||
@@ -55,7 +55,21 @@ export function PageSidebarDark({ collapsed, onToggleCollapsed }: PageSidebarDar
|
||||
}, 180);
|
||||
};
|
||||
|
||||
const hasActiveChild = (item: MenuItem): boolean => menuItemContainsPath(item, pathname);
|
||||
const closeNavPanel = () => {
|
||||
if (hoverCloseTimerRef.current !== null) {
|
||||
window.clearTimeout(hoverCloseTimerRef.current);
|
||||
hoverCloseTimerRef.current = null;
|
||||
}
|
||||
setHoveredTopId(null);
|
||||
};
|
||||
|
||||
const hasActiveChild = (item: MenuItem): boolean => {
|
||||
if (item.path && pathname === item.path) {
|
||||
return true;
|
||||
}
|
||||
|
||||
return item.children?.some(hasActiveChild) ?? false;
|
||||
};
|
||||
|
||||
// lazy init：仅在首次 render 计算一次 active id；之后由下面的 useEffect 跟随 pathname 同步
|
||||
const [activeTopId, setActiveTopId] = useState<string>(
|
||||
@@ -69,9 +83,7 @@ export function PageSidebarDark({ collapsed, onToggleCollapsed }: PageSidebarDar
|
||||
if (matchedTopItem) {
|
||||
setActiveTopId(matchedTopItem.id);
|
||||
}
|
||||
// 不在导航时强制关闭 flyout：flyout 绝对定位在 group 盒子之外，强制关闭后鼠标
|
||||
// 仍停在 flyout 区域却收不到 group 的 mouseenter，导致连续点击第二个任务时面板
|
||||
// 消失且不回来。开合只交给 hover（mouseleave 有 180ms 宽限），即可连续点选。
|
||||
closeNavPanel();
|
||||
}, [pathname, visibleMenuItems]);
|
||||
|
||||
const toggleExpanded = (id: string) => {
|
||||
@@ -110,23 +122,18 @@ export function PageSidebarDark({ collapsed, onToggleCollapsed }: PageSidebarDar
|
||||
if (item.children?.length) {
|
||||
const firstPath = findFirstRoutePath(item.children);
|
||||
if (firstPath) {
|
||||
closeNavPanel();
|
||||
navigate(firstPath);
|
||||
}
|
||||
return;
|
||||
}
|
||||
|
||||
if (item.path) {
|
||||
closeNavPanel();
|
||||
navigate(item.path);
|
||||
}
|
||||
};
|
||||
|
||||
const handleSecondItemClick = (item: MenuItem) => {
|
||||
const path = item.path ?? findFirstRoutePath(item.children);
|
||||
if (path) {
|
||||
navigate(path);
|
||||
}
|
||||
};
|
||||
|
||||
const handleMenuItemClick = (item: MenuItem) => {
|
||||
if (item.children?.length) {
|
||||
toggleExpanded(item.id);
|
||||
@@ -134,47 +141,17 @@ export function PageSidebarDark({ collapsed, onToggleCollapsed }: PageSidebarDar
|
||||
}
|
||||
|
||||
if (item.path) {
|
||||
closeNavPanel();
|
||||
navigate(item.path);
|
||||
}
|
||||
};
|
||||
|
||||
const activeTopItem =
|
||||
visibleMenuItems.find((item) => item.id === activeTopId) ??
|
||||
visibleMenuItems.find(hasActiveChild) ??
|
||||
visibleMenuItems[0];
|
||||
const sidebarMenuItemsToRender = inWujieEnv
|
||||
? visibleMenuItems
|
||||
: activeTopItem?.children ?? [];
|
||||
const isSidebarItemActive = (item: MenuItem) =>
|
||||
inWujieEnv
|
||||
? item.id === activeTopId
|
||||
: item.path || item.aliasPaths?.length
|
||||
? menuItemMatchesPath(item, pathname)
|
||||
: menuItemContainsPath(item, pathname);
|
||||
const handleSidebarItemClick = (item: MenuItem) => {
|
||||
if (inWujieEnv) {
|
||||
handleTopItemClick(item);
|
||||
return;
|
||||
}
|
||||
|
||||
handleSecondItemClick(item);
|
||||
};
|
||||
const getSidebarLabelClassName = (label: string) => {
|
||||
if (!inWujieEnv && Array.from(label.trim()).length === 4) {
|
||||
return "item-nav__label item-nav__label--four-char";
|
||||
}
|
||||
|
||||
return "item-nav__label";
|
||||
};
|
||||
|
||||
const renderFlyoutMenu = (items: MenuItem[], level = 0) => (
|
||||
<div className={`sidebar-flyout-list sidebar-flyout-list--level-${level}`}>
|
||||
{items.map((item) => {
|
||||
const hasChildren = !!item.children?.length;
|
||||
const isExpanded = expandedIds.has(item.id);
|
||||
const isSelected = item.path || item.aliasPaths?.length
|
||||
? menuItemMatchesPath(item, pathname)
|
||||
: hasActiveChild(item);
|
||||
const isSelected = item.path ? pathname === item.path : hasActiveChild(item);
|
||||
|
||||
return (
|
||||
<div
|
||||
@@ -195,7 +172,6 @@ export function PageSidebarDark({ collapsed, onToggleCollapsed }: PageSidebarDar
|
||||
})}
|
||||
</div>
|
||||
);
|
||||
|
||||
// collapsed 状态：sidebar 主体完全消失（包括占据的宽度），只渲染左下角浮动入口
|
||||
if (collapsed) {
|
||||
return (
|
||||
@@ -213,16 +189,16 @@ export function PageSidebarDark({ collapsed, onToggleCollapsed }: PageSidebarDar
|
||||
}
|
||||
|
||||
return (
|
||||
<aside className={`sidebar-contnet ${inWujieEnv ? "" : "sidebar-contnet--top-submenu"}`}>
|
||||
<aside className="sidebar-contnet">
|
||||
<div className="sidebar-contents">
|
||||
{sidebarMenuItemsToRender.map((item) => (
|
||||
{visibleMenuItems.map((item) => (
|
||||
<div
|
||||
key={item.id}
|
||||
className="item-nav-group"
|
||||
onMouseEnter={inWujieEnv ? () => handleNavHoverEnter(item.id) : undefined}
|
||||
onMouseLeave={inWujieEnv ? scheduleNavPanelClose : undefined}
|
||||
onMouseEnter={() => handleNavHoverEnter(item.id)}
|
||||
onMouseLeave={scheduleNavPanelClose}
|
||||
>
|
||||
{inWujieEnv && item.children?.length ? (
|
||||
{item.children?.length ? (
|
||||
<div
|
||||
className={`item-nav-content ${hoveredTopId === item.id ? "item-nav-content--visible" : ""}`}
|
||||
>
|
||||
@@ -232,12 +208,12 @@ export function PageSidebarDark({ collapsed, onToggleCollapsed }: PageSidebarDar
|
||||
<div className="item-nav-wrap">
|
||||
<button
|
||||
type="button"
|
||||
className={`item-nav ${isSidebarItemActive(item) ? "item-nav-active" : ""} ${hoveredTopId === item.id ? "item-nav-hovered" : ""}`}
|
||||
onClick={() => handleSidebarItemClick(item)}
|
||||
className={`item-nav ${item.id === activeTopId ? "item-nav-active" : ""} ${hoveredTopId === item.id ? "item-nav-hovered" : ""}`}
|
||||
onClick={() => handleTopItemClick(item)}
|
||||
>
|
||||
<span className={getSidebarLabelClassName(item.label)}>{item.label}</span>
|
||||
<span className="item-nav__label">{item.label}</span>
|
||||
</button>
|
||||
{inWujieEnv && item.children?.length && hoveredTopId === item.id ? (
|
||||
{item.children?.length && hoveredTopId === item.id ? (
|
||||
<span className="item-nav-hover-bridge" aria-hidden="true" />
|
||||
) : null}
|
||||
</div>
|
||||
|
||||
@@ -1,303 +0,0 @@
|
||||
import { useEffect, useState } from "react";
|
||||
import { useNavigate, useLocation } from "react-router-dom";
|
||||
import {
|
||||
Bell,
|
||||
ChevronLeft,
|
||||
ChevronRight,
|
||||
ChevronsUpDown,
|
||||
Folder,
|
||||
FolderOpen,
|
||||
Settings,
|
||||
User,
|
||||
Wrench,
|
||||
} from "lucide-react";
|
||||
|
||||
import {
|
||||
DropdownMenu,
|
||||
DropdownMenuContent,
|
||||
DropdownMenuItem,
|
||||
DropdownMenuTrigger,
|
||||
} from "@/components/ui/dropdown-menu";
|
||||
import { NotificationDialog } from "@/components/workspace/notification-dialog";
|
||||
import { SettingsDialog } from "@/components/workspace/settings";
|
||||
import { useTheme } from "@/contexts/ThemeContext";
|
||||
import { getStoredAuth } from "@/core/auth";
|
||||
import { useI18n } from "@/core/i18n/hooks";
|
||||
import { useUnreadCount } from "@/core/notifications";
|
||||
import {
|
||||
defaultExpandedIds,
|
||||
type MenuItem,
|
||||
} from "@/core/page-layout/sidebar-menu";
|
||||
import { useSidebarMenuItems } from "@/core/page-layout/use-sidebar-menu";
|
||||
|
||||
type PageSidebarProps = {
|
||||
collapsed: boolean;
|
||||
onToggleCollapsed: () => void;
|
||||
};
|
||||
|
||||
type SettingsSection = "account" | "appearance" | "memory" | "tools" | "about";
|
||||
|
||||
export function PageSidebar({
|
||||
collapsed,
|
||||
onToggleCollapsed,
|
||||
}: PageSidebarProps) {
|
||||
const { theme } = useTheme();
|
||||
const isDark = theme === "dark";
|
||||
const isAdmin = getStoredAuth()?.system_role === "admin";
|
||||
const visibleMenuItems = useSidebarMenuItems(!!isAdmin);
|
||||
const { t } = useI18n();
|
||||
const { data: unreadCount = 0 } = useUnreadCount();
|
||||
// defaultExpandedIds 是共享的 ReadonlySet，复制一份给 useState（避免外部突变）
|
||||
const [expandedIds, setExpandedIds] = useState<Set<string>>(() => new Set(defaultExpandedIds));
|
||||
const [settingsOpen, setSettingsOpen] = useState(false);
|
||||
const [settingsDefaultSection, setSettingsDefaultSection] =
|
||||
useState<SettingsSection>("appearance");
|
||||
const [notificationOpen, setNotificationOpen] = useState(false);
|
||||
const [mounted, setMounted] = useState(false);
|
||||
const navigate = useNavigate();
|
||||
const { pathname } = useLocation();
|
||||
|
||||
useEffect(() => {
|
||||
setMounted(true);
|
||||
}, []);
|
||||
|
||||
const openSettings = (section: SettingsSection) => {
|
||||
setSettingsDefaultSection(section);
|
||||
setSettingsOpen(true);
|
||||
};
|
||||
|
||||
const iconButtonClass = isDark
|
||||
? "rounded-lg p-2 text-slate-400 transition-all duration-300 ease-out hover:scale-105 hover:bg-slate-800/50 hover:text-white"
|
||||
: "rounded-lg p-2 text-gray-500 transition-all duration-300 ease-out hover:scale-105 hover:bg-red-50 hover:text-red-600";
|
||||
const panelClass = isDark
|
||||
? "ml-4 space-y-1 border-l border-blue-800/30 pl-4"
|
||||
: "ml-4 space-y-1 border-l border-red-100 pl-4";
|
||||
const labelClass = (compact: boolean, small = false) =>
|
||||
`overflow-hidden whitespace-nowrap transition-all duration-300 ease-out ${
|
||||
compact ? "max-w-0 opacity-0" : "max-w-[220px] opacity-100"
|
||||
} ${small ? "text-xs" : "text-sm"} flex-1 font-medium`;
|
||||
const menuButtonClass = (isActive: boolean, depth = 0) =>
|
||||
`group relative flex w-full items-center rounded-[2px] py-2 text-left transition-all duration-300 ease-out theme-transition ${
|
||||
collapsed && depth === 0 ? "justify-center px-2" : "gap-3 px-3"
|
||||
} ${
|
||||
isActive
|
||||
? isDark
|
||||
? "bg-blue-500/15 text-cyan-300 font-semibold"
|
||||
: "bg-red-50 text-red-600 font-semibold"
|
||||
: isDark
|
||||
? "text-slate-400 hover:bg-slate-800/50 hover:text-white"
|
||||
: "text-gray-600 hover:bg-red-50 hover:text-red-600"
|
||||
}`;
|
||||
const footerButtonClass = menuButtonClass(false, 0);
|
||||
|
||||
const toggleExpanded = (id: string) => {
|
||||
setExpandedIds((prev) => {
|
||||
const next = new Set(prev);
|
||||
if (next.has(id)) {
|
||||
next.delete(id);
|
||||
} else {
|
||||
next.add(id);
|
||||
}
|
||||
return next;
|
||||
});
|
||||
};
|
||||
|
||||
const renderLeaf = (item: MenuItem, depth = 0) => {
|
||||
const Icon = item.icon;
|
||||
const isActive = !!item.path && pathname === item.path;
|
||||
|
||||
return (
|
||||
<button
|
||||
key={item.id}
|
||||
type="button"
|
||||
title={collapsed && depth === 0 ? item.label : undefined}
|
||||
onClick={() => item.path && navigate(item.path)}
|
||||
className={menuButtonClass(isActive, depth)}
|
||||
>
|
||||
{isActive && (
|
||||
<span
|
||||
className={`absolute left-0 inset-y-0 w-[3px] rounded-r ${
|
||||
isDark ? "bg-cyan-300" : "bg-red-600"
|
||||
}`}
|
||||
/>
|
||||
)}
|
||||
<Icon className={`${depth === 0 ? "h-[18px] w-[18px]" : "h-4 w-4"} flex-shrink-0 transition-transform duration-300 ease-out group-hover:scale-105`} />
|
||||
{(!collapsed || depth > 0) && (
|
||||
<span className={labelClass(false, depth > 0)}>
|
||||
{item.label}
|
||||
</span>
|
||||
)}
|
||||
</button>
|
||||
);
|
||||
};
|
||||
|
||||
const renderBranch = (item: MenuItem, depth = 0) => {
|
||||
const expanded = expandedIds.has(item.id);
|
||||
const isTopLevel = depth === 0;
|
||||
const Icon = depth > 0 && item.icon === Folder && expanded ? FolderOpen : item.icon;
|
||||
|
||||
return (
|
||||
<div key={item.id} className="space-y-1">
|
||||
<button
|
||||
type="button"
|
||||
onClick={() => {
|
||||
if (collapsed && isTopLevel) {
|
||||
onToggleCollapsed();
|
||||
return;
|
||||
}
|
||||
toggleExpanded(item.id);
|
||||
}}
|
||||
title={collapsed && isTopLevel ? item.label : undefined}
|
||||
className={menuButtonClass(false, depth)}
|
||||
>
|
||||
<Icon
|
||||
className={`${isTopLevel ? "h-[18px] w-[18px]" : "h-4 w-4"} flex-shrink-0 transition-transform duration-300 ease-out group-hover:scale-105 ${
|
||||
expanded && depth > 0 ? "scale-105" : ""
|
||||
}`}
|
||||
/>
|
||||
{(!collapsed || !isTopLevel) && (
|
||||
<span className={labelClass(false, depth > 0)}>
|
||||
{item.label}
|
||||
</span>
|
||||
)}
|
||||
{(!collapsed || !isTopLevel) && (
|
||||
<ChevronRight
|
||||
className={`h-4 w-4 transition-transform duration-300 ease-out ${expanded ? "rotate-90" : ""}`}
|
||||
/>
|
||||
)}
|
||||
</button>
|
||||
|
||||
{(!collapsed || !isTopLevel) && item.children && (
|
||||
<div
|
||||
className={`${panelClass} overflow-hidden transition-all duration-300 ease-out ${
|
||||
expanded ? "max-h-[1200px] opacity-100" : "max-h-0 opacity-0"
|
||||
}`}
|
||||
>
|
||||
{item.children.map((child) =>
|
||||
child.children ? renderBranch(child, depth + 1) : renderLeaf(child, depth + 1),
|
||||
)}
|
||||
</div>
|
||||
)}
|
||||
</div>
|
||||
);
|
||||
};
|
||||
|
||||
return (
|
||||
<>
|
||||
<SettingsDialog
|
||||
open={settingsOpen}
|
||||
onOpenChange={setSettingsOpen}
|
||||
defaultSection={settingsDefaultSection}
|
||||
/>
|
||||
<NotificationDialog
|
||||
open={notificationOpen}
|
||||
onOpenChange={setNotificationOpen}
|
||||
/>
|
||||
<aside
|
||||
className={`fixed left-0 top-16 z-40 flex h-[calc(100vh-4rem)] flex-col overflow-hidden shadow-sm theme-transition transition-all duration-300 ${
|
||||
isDark
|
||||
? "border-r border-blue-800/30 bg-slate-900"
|
||||
: "border-r border-red-100 bg-white"
|
||||
} ${collapsed ? "w-16" : "w-64"}`}
|
||||
>
|
||||
<button
|
||||
type="button"
|
||||
onClick={onToggleCollapsed}
|
||||
title={collapsed ? "展开侧边栏" : "收起侧边栏"}
|
||||
className={`absolute -right-3 top-6 z-50 flex h-6 w-6 items-center justify-center rounded-full shadow-md theme-transition ${
|
||||
isDark
|
||||
? "bg-slate-700 text-slate-300 hover:bg-slate-600"
|
||||
: "bg-white text-gray-500 hover:bg-red-50 hover:text-red-600 border border-red-100"
|
||||
}`}
|
||||
>
|
||||
{collapsed ? <ChevronRight className="h-3 w-3" /> : <ChevronLeft className="h-3 w-3" />}
|
||||
</button>
|
||||
|
||||
<nav className={`flex-1 overflow-y-auto ${collapsed ? "p-2" : "p-4"} space-y-2 transition-all duration-300 ease-out`}>
|
||||
{visibleMenuItems.map((item) =>
|
||||
item.children ? renderBranch(item) : renderLeaf(item),
|
||||
)}
|
||||
</nav>
|
||||
|
||||
<div
|
||||
className={`mt-auto shrink-0 theme-transition ${
|
||||
isDark ? "border-t border-blue-800/30" : "border-t border-red-100"
|
||||
} ${collapsed ? "p-2" : "p-4"}`}
|
||||
>
|
||||
{mounted ? (
|
||||
<DropdownMenu>
|
||||
<DropdownMenuTrigger asChild>
|
||||
<button
|
||||
type="button"
|
||||
title={collapsed ? t.workspace.settingsAndMore : undefined}
|
||||
className={footerButtonClass}
|
||||
>
|
||||
<Settings className="h-5 w-5 flex-shrink-0" />
|
||||
{!collapsed && (
|
||||
<>
|
||||
<span className={labelClass(false)}>
|
||||
{t.workspace.settingsAndMore}
|
||||
</span>
|
||||
<ChevronsUpDown className="h-4 w-4 flex-shrink-0" />
|
||||
</>
|
||||
)}
|
||||
</button>
|
||||
</DropdownMenuTrigger>
|
||||
<DropdownMenuContent
|
||||
align={collapsed ? "start" : "end"}
|
||||
side={collapsed ? "right" : "top"}
|
||||
sideOffset={4}
|
||||
className="min-w-56 rounded-lg"
|
||||
>
|
||||
<DropdownMenuItem onClick={() => setNotificationOpen(true)}>
|
||||
<Bell className="size-4" />
|
||||
{t.notifications.title}
|
||||
{unreadCount > 0 && (
|
||||
<span
|
||||
className={`ml-auto flex h-4 min-w-4 items-center justify-center rounded-full px-1 text-[10px] font-medium leading-none text-white ${
|
||||
isDark ? "bg-cyan-500" : "bg-red-500"
|
||||
}`}
|
||||
>
|
||||
{unreadCount > 99 ? "99+" : unreadCount}
|
||||
</span>
|
||||
)}
|
||||
</DropdownMenuItem>
|
||||
<DropdownMenuItem onClick={() => openSettings("tools")}>
|
||||
<Wrench className="size-4" />
|
||||
{t.settings.sections.tools}
|
||||
</DropdownMenuItem>
|
||||
<DropdownMenuItem onClick={() => openSettings("account")}>
|
||||
<User className="size-4" />
|
||||
个人信息
|
||||
</DropdownMenuItem>
|
||||
</DropdownMenuContent>
|
||||
</DropdownMenu>
|
||||
) : (
|
||||
<button
|
||||
type="button"
|
||||
className={`${footerButtonClass} pointer-events-none`}
|
||||
>
|
||||
<Settings className="h-5 w-5 flex-shrink-0" />
|
||||
{!collapsed && (
|
||||
<span className={labelClass(false)}>
|
||||
{t.workspace.settingsAndMore}
|
||||
</span>
|
||||
)}
|
||||
</button>
|
||||
)}
|
||||
</div>
|
||||
|
||||
<div
|
||||
className={`pointer-events-none absolute right-0 top-0 h-32 w-32 rounded-full blur-3xl theme-transition ${
|
||||
isDark ? "bg-blue-500/10" : "bg-red-500/5"
|
||||
}`}
|
||||
/>
|
||||
<div
|
||||
className={`pointer-events-none absolute bottom-20 right-0 h-24 w-24 rounded-full blur-3xl theme-transition ${
|
||||
isDark ? "bg-cyan-400/5" : "bg-red-400/5"
|
||||
}`}
|
||||
/>
|
||||
</aside>
|
||||
</>
|
||||
);
|
||||
}
|
||||
@@ -7,12 +7,7 @@ import {
|
||||
ChevronsUpDown,
|
||||
Folder,
|
||||
FolderOpen,
|
||||
Gauge,
|
||||
ListTree,
|
||||
MousePointerClick,
|
||||
Settings,
|
||||
ShieldAlert,
|
||||
ShieldCheck,
|
||||
User,
|
||||
Wrench,
|
||||
} from "lucide-react";
|
||||
@@ -26,25 +21,15 @@ import {
|
||||
import { NotificationDialog } from "@/components/workspace/notification-dialog";
|
||||
import { SettingsDialog } from "@/components/workspace/settings";
|
||||
import { useTheme } from "@/contexts/ThemeContext";
|
||||
import { getAccountDisplayName, getStoredAuth } from "@/core/auth";
|
||||
|
||||
/** 敏感词管理仅对用户名为 lqq 的账号可见（用户名 = 设置→个人信息页展示名）。 */
|
||||
const SENSITIVE_WORDS_USERNAME = "lqq";
|
||||
const canManageSensitiveWords = () =>
|
||||
getAccountDisplayName().toLowerCase() === SENSITIVE_WORDS_USERNAME;
|
||||
import { isWujieEnv } from "@/core/embed/wujie";
|
||||
import { getStoredAuth } from "@/core/auth";
|
||||
import { useI18n } from "@/core/i18n/hooks";
|
||||
import { useUnreadCount } from "@/core/notifications";
|
||||
import {
|
||||
defaultExpandedIds,
|
||||
getActiveThirdLevelItems,
|
||||
menuItemContainsPath,
|
||||
menuItemMatchesPath,
|
||||
filterMenuItemsByRole,
|
||||
sidebarMenuItems,
|
||||
type MenuItem,
|
||||
} from "@/core/page-layout/sidebar-menu";
|
||||
import { useSidebarMenuItems } from "@/core/page-layout/use-sidebar-menu";
|
||||
|
||||
import "@/components/page-sdebar.css";
|
||||
|
||||
type PageSidebarProps = {
|
||||
collapsed: boolean;
|
||||
@@ -53,245 +38,14 @@ type PageSidebarProps = {
|
||||
|
||||
type SettingsSection = "account" | "appearance" | "memory" | "tools" | "about";
|
||||
|
||||
const hasActiveChild = (item: MenuItem, pathname: string): boolean =>
|
||||
menuItemContainsPath(item, pathname);
|
||||
|
||||
const findFirstRoutePath = (items?: MenuItem[]): string | null => {
|
||||
if (!items?.length) {
|
||||
return null;
|
||||
}
|
||||
for (const child of items) {
|
||||
if (child.path) {
|
||||
return child.path;
|
||||
}
|
||||
const nested = findFirstRoutePath(child.children);
|
||||
if (nested) {
|
||||
return nested;
|
||||
}
|
||||
}
|
||||
return null;
|
||||
};
|
||||
|
||||
/**
|
||||
* 浅色侧边栏入口。
|
||||
*
|
||||
* - 普通桌面环境：渲染「二级导航图标卡片」（一级在顶栏、二级在侧边栏，参考 导航-1.png）。
|
||||
* - wujie 嵌入环境：顶栏被宿主隐藏，一级导航无处安放，故回落到原手风琴整树菜单
|
||||
* （保证嵌入模式下所有层级仍可达，零回归）。
|
||||
*/
|
||||
export function PageSidebar({ collapsed, onToggleCollapsed }: PageSidebarProps) {
|
||||
if (isWujieEnv()) {
|
||||
return (
|
||||
<LegacyAccordionSidebar
|
||||
collapsed={collapsed}
|
||||
onToggleCollapsed={onToggleCollapsed}
|
||||
/>
|
||||
);
|
||||
}
|
||||
return <SecondaryNavSidebar />;
|
||||
}
|
||||
|
||||
/**
|
||||
* 浅色二级导航侧边栏（图标卡片，红/蓝主题色经 --brand-* 变量自动切换）。
|
||||
* 浅色模式不支持侧边栏收起，故无收起按钮、不接收 collapsed 相关 props。
|
||||
*/
|
||||
function SecondaryNavSidebar() {
|
||||
const isAdmin = getStoredAuth()?.system_role === "admin";
|
||||
const visibleMenuItems = useSidebarMenuItems(!!isAdmin);
|
||||
const { t } = useI18n();
|
||||
const { data: unreadCount = 0 } = useUnreadCount();
|
||||
const [settingsOpen, setSettingsOpen] = useState(false);
|
||||
const [settingsDefaultSection, setSettingsDefaultSection] =
|
||||
useState<SettingsSection>("appearance");
|
||||
const [notificationOpen, setNotificationOpen] = useState(false);
|
||||
const [mounted, setMounted] = useState(false);
|
||||
const navigate = useNavigate();
|
||||
const { pathname } = useLocation();
|
||||
|
||||
useEffect(() => {
|
||||
setMounted(true);
|
||||
}, []);
|
||||
|
||||
const openSettings = (section: SettingsSection) => {
|
||||
setSettingsDefaultSection(section);
|
||||
setSettingsOpen(true);
|
||||
};
|
||||
|
||||
// 当前激活的一级菜单（与顶栏一级导航联动）；其 children 即侧边栏要渲染的二级导航。
|
||||
const activeTopItem =
|
||||
visibleMenuItems.find((item) => hasActiveChild(item, pathname)) ??
|
||||
visibleMenuItems[0];
|
||||
const secondaryItems = activeTopItem?.children ?? [];
|
||||
|
||||
const isItemActive = (item: MenuItem) =>
|
||||
item.path || item.aliasPaths?.length
|
||||
? menuItemMatchesPath(item, pathname)
|
||||
: menuItemContainsPath(item, pathname);
|
||||
|
||||
const handleItemClick = (item: MenuItem) => {
|
||||
const path = item.path ?? findFirstRoutePath(item.children);
|
||||
if (path) {
|
||||
navigate(path);
|
||||
}
|
||||
};
|
||||
|
||||
// 当前激活二级项的三级子项；非空时侧边栏在右侧展开「三级列」并整体变宽（见 css 过渡）。
|
||||
const thirdItems = getActiveThirdLevelItems(visibleMenuItems, pathname);
|
||||
const expanded = thirdItems.length > 0;
|
||||
|
||||
return (
|
||||
<>
|
||||
<SettingsDialog
|
||||
open={settingsOpen}
|
||||
onOpenChange={setSettingsOpen}
|
||||
defaultSection={settingsDefaultSection}
|
||||
/>
|
||||
<NotificationDialog
|
||||
open={notificationOpen}
|
||||
onOpenChange={setNotificationOpen}
|
||||
/>
|
||||
<aside className={`light-sidebar ${expanded ? "light-sidebar--expanded" : ""}`}>
|
||||
{/* 一级列：二级图标卡片 + 底部设置 */}
|
||||
<div className="light-sidebar__primary">
|
||||
<nav className="light-sidebar__list" aria-label="二级菜单">
|
||||
{secondaryItems.map((item) => {
|
||||
const Icon = item.icon;
|
||||
const active = isItemActive(item);
|
||||
return (
|
||||
<button
|
||||
key={item.id}
|
||||
type="button"
|
||||
title={item.label}
|
||||
onClick={() => handleItemClick(item)}
|
||||
className={`light-sidebar-item ${active ? "light-sidebar-item--active" : ""}`}
|
||||
>
|
||||
<Icon className="light-sidebar-item__icon h-7 w-7" />
|
||||
<span className="light-sidebar-item__label">{item.label}</span>
|
||||
</button>
|
||||
);
|
||||
})}
|
||||
</nav>
|
||||
|
||||
<div className="light-sidebar__footer">
|
||||
{mounted ? (
|
||||
<DropdownMenu>
|
||||
<DropdownMenuTrigger asChild>
|
||||
<button
|
||||
type="button"
|
||||
title={t.workspace.settingsAndMore}
|
||||
className="light-sidebar-item"
|
||||
>
|
||||
<Settings className="light-sidebar-item__icon h-7 w-7" />
|
||||
<span className="light-sidebar-item__label">设置</span>
|
||||
</button>
|
||||
</DropdownMenuTrigger>
|
||||
<DropdownMenuContent
|
||||
align="start"
|
||||
side="right"
|
||||
sideOffset={8}
|
||||
className="min-w-56 rounded-lg"
|
||||
>
|
||||
<DropdownMenuItem onClick={() => setNotificationOpen(true)}>
|
||||
<Bell className="size-4" />
|
||||
{t.notifications.title}
|
||||
{unreadCount > 0 && (
|
||||
<span className="ml-auto flex h-4 min-w-4 items-center justify-center rounded-full bg-[var(--brand-900)] px-1 text-[10px] font-medium leading-none text-white">
|
||||
{unreadCount > 99 ? "99+" : unreadCount}
|
||||
</span>
|
||||
)}
|
||||
</DropdownMenuItem>
|
||||
<DropdownMenuItem onClick={() => openSettings("tools")}>
|
||||
<Wrench className="size-4" />
|
||||
{t.settings.sections.tools}
|
||||
</DropdownMenuItem>
|
||||
<DropdownMenuItem onClick={() => openSettings("account")}>
|
||||
<User className="size-4" />
|
||||
个人信息
|
||||
</DropdownMenuItem>
|
||||
{isAdmin && (
|
||||
<DropdownMenuItem
|
||||
onClick={() => navigate("/page/strategy/admin/menu-management")}
|
||||
>
|
||||
<ListTree className="size-4" />
|
||||
菜单管理
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
{isAdmin && (
|
||||
<DropdownMenuItem
|
||||
onClick={() => navigate("/page/strategy/admin/task-buttons")}
|
||||
>
|
||||
<MousePointerClick className="size-4" />
|
||||
按钮管理
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
{isAdmin && (
|
||||
<DropdownMenuItem
|
||||
onClick={() => navigate("/page/strategy/admin/login-whitelist")}
|
||||
>
|
||||
<ShieldCheck className="size-4" />
|
||||
白名单管理
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
{isAdmin && (
|
||||
<DropdownMenuItem
|
||||
onClick={() => navigate("/page/strategy/settings/model-concurrency")}
|
||||
>
|
||||
<Gauge className="size-4" />
|
||||
模型并发限制
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
{canManageSensitiveWords() && (
|
||||
<DropdownMenuItem
|
||||
onClick={() => navigate("/page/strategy/admin/sensitive-words")}
|
||||
>
|
||||
<ShieldAlert className="size-4" />
|
||||
敏感词管理
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
</DropdownMenuContent>
|
||||
</DropdownMenu>
|
||||
) : null}
|
||||
</div>
|
||||
</div>
|
||||
|
||||
{/* 二级列：三级导航（激活二级项有三级子项时展开，整体动画见 css width transition） */}
|
||||
<div className="light-sidebar__secondary" aria-hidden={!expanded}>
|
||||
<nav className="light-sidebar-subnav" aria-label="三级菜单">
|
||||
{thirdItems.map((child) => {
|
||||
const ChildIcon = child.icon;
|
||||
const childActive = menuItemMatchesPath(child, pathname)
|
||||
|| (child.children?.length ? menuItemContainsPath(child, pathname) : false);
|
||||
return (
|
||||
<button
|
||||
key={child.id}
|
||||
type="button"
|
||||
title={child.label}
|
||||
onClick={() => handleItemClick(child)}
|
||||
className={`light-sidebar-subitem ${childActive ? "light-sidebar-subitem--active" : ""}`}
|
||||
>
|
||||
<ChildIcon className="light-sidebar-subitem__icon h-4 w-4" />
|
||||
<span className="light-sidebar-subitem__label">{child.label}</span>
|
||||
</button>
|
||||
);
|
||||
})}
|
||||
</nav>
|
||||
</div>
|
||||
</aside>
|
||||
</>
|
||||
);
|
||||
}
|
||||
|
||||
/**
|
||||
* 旧版浅色整树手风琴侧边栏。仅用于 wujie 嵌入环境（顶栏被宿主隐藏时保证所有层级可达）。
|
||||
*/
|
||||
function LegacyAccordionSidebar({
|
||||
export function PageSidebar({
|
||||
collapsed,
|
||||
onToggleCollapsed,
|
||||
}: PageSidebarProps) {
|
||||
const { theme } = useTheme();
|
||||
const isDark = theme === "dark";
|
||||
const isAdmin = getStoredAuth()?.system_role === "admin";
|
||||
const visibleMenuItems = useSidebarMenuItems(!!isAdmin);
|
||||
const visibleMenuItems = filterMenuItemsByRole(sidebarMenuItems, !!isAdmin);
|
||||
const { t } = useI18n();
|
||||
const { data: unreadCount = 0 } = useUnreadCount();
|
||||
// defaultExpandedIds 是共享的 ReadonlySet，复制一份给 useState（避免外部突变）
|
||||
@@ -313,6 +67,9 @@ function LegacyAccordionSidebar({
|
||||
setSettingsOpen(true);
|
||||
};
|
||||
|
||||
const iconButtonClass = isDark
|
||||
? "rounded-lg p-2 text-slate-400 transition-all duration-300 ease-out hover:scale-105 hover:bg-slate-800/50 hover:text-white"
|
||||
: "rounded-lg p-2 text-gray-500 transition-all duration-300 ease-out hover:scale-105 hover:bg-red-50 hover:text-red-600";
|
||||
const panelClass = isDark
|
||||
? "ml-4 space-y-1 border-l border-blue-800/30 pl-4"
|
||||
: "ml-4 space-y-1 border-l border-red-100 pl-4";
|
||||
@@ -321,8 +78,8 @@ function LegacyAccordionSidebar({
|
||||
compact ? "max-w-0 opacity-0" : "max-w-[220px] opacity-100"
|
||||
} ${small ? "text-xs" : "text-sm"} flex-1 font-medium`;
|
||||
const menuButtonClass = (isActive: boolean, depth = 0) =>
|
||||
`group relative flex w-full items-center rounded-[2px] py-2 text-left transition-all duration-300 ease-out theme-transition ${
|
||||
collapsed && depth === 0 ? "justify-center px-2" : "gap-3 px-3"
|
||||
`flex w-full items-center rounded-lg py-2 text-left transition-all duration-300 ease-out hover:-translate-y-0.5 theme-transition ${
|
||||
collapsed && depth === 0 ? "justify-center px-2" : "gap-2 px-3"
|
||||
} ${
|
||||
isActive
|
||||
? isDark
|
||||
@@ -348,7 +105,7 @@ function LegacyAccordionSidebar({
|
||||
|
||||
const renderLeaf = (item: MenuItem, depth = 0) => {
|
||||
const Icon = item.icon;
|
||||
const isActive = menuItemMatchesPath(item, pathname);
|
||||
const isActive = !!item.path && pathname === item.path;
|
||||
|
||||
return (
|
||||
<button
|
||||
@@ -358,14 +115,7 @@ function LegacyAccordionSidebar({
|
||||
onClick={() => item.path && navigate(item.path)}
|
||||
className={menuButtonClass(isActive, depth)}
|
||||
>
|
||||
{isActive && (
|
||||
<span
|
||||
className={`absolute left-0 inset-y-0 w-[3px] rounded-r ${
|
||||
isDark ? "bg-cyan-300" : "bg-red-600"
|
||||
}`}
|
||||
/>
|
||||
)}
|
||||
<Icon className={`${depth === 0 ? "h-[18px] w-[18px]" : "h-4 w-4"} flex-shrink-0 transition-transform duration-300 ease-out group-hover:scale-105`} />
|
||||
<Icon className={`${depth === 0 ? "h-5 w-5" : "h-4 w-4"} flex-shrink-0 transition-transform duration-300 ease-out`} />
|
||||
{(!collapsed || depth > 0) && (
|
||||
<span className={labelClass(false, depth > 0)}>
|
||||
{item.label}
|
||||
@@ -395,7 +145,7 @@ function LegacyAccordionSidebar({
|
||||
className={menuButtonClass(false, depth)}
|
||||
>
|
||||
<Icon
|
||||
className={`${isTopLevel ? "h-[18px] w-[18px]" : "h-4 w-4"} flex-shrink-0 transition-transform duration-300 ease-out group-hover:scale-105 ${
|
||||
className={`${isTopLevel ? "h-5 w-5" : "h-4 w-4"} flex-shrink-0 transition-transform duration-300 ease-out ${
|
||||
expanded && depth > 0 ? "scale-105" : ""
|
||||
}`}
|
||||
/>
|
||||
@@ -514,30 +264,6 @@ function LegacyAccordionSidebar({
|
||||
<User className="size-4" />
|
||||
个人信息
|
||||
</DropdownMenuItem>
|
||||
{isAdmin && (
|
||||
<DropdownMenuItem onClick={() => navigate("/page/strategy/admin/menu-management")}>
|
||||
<ListTree className="size-4" />
|
||||
菜单管理
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
{isAdmin && (
|
||||
<DropdownMenuItem onClick={() => navigate("/page/strategy/admin/task-buttons")}>
|
||||
<MousePointerClick className="size-4" />
|
||||
按钮管理
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
{isAdmin && (
|
||||
<DropdownMenuItem onClick={() => navigate("/page/strategy/admin/login-whitelist")}>
|
||||
<ShieldCheck className="size-4" />
|
||||
白名单管理
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
{canManageSensitiveWords() && (
|
||||
<DropdownMenuItem onClick={() => navigate("/page/strategy/admin/sensitive-words")}>
|
||||
<ShieldAlert className="size-4" />
|
||||
敏感词管理
|
||||
</DropdownMenuItem>
|
||||
)}
|
||||
</DropdownMenuContent>
|
||||
</DropdownMenu>
|
||||
) : (
|
||||
|
||||
@@ -5,7 +5,7 @@ import {
|
||||
QueryClientProvider as TanStackQueryClientProvider,
|
||||
} from "@tanstack/react-query";
|
||||
|
||||
export const queryClient = new QueryClient();
|
||||
const queryClient = new QueryClient();
|
||||
|
||||
export function QueryClientProvider({
|
||||
children,
|
||||
|
||||
@@ -1,59 +0,0 @@
|
||||
import * as React from "react";
|
||||
|
||||
export interface AutoGrowTextareaProps
|
||||
extends React.TextareaHTMLAttributes<HTMLTextAreaElement> {
|
||||
/** 内容增长高度上限(px)，超过后框内滚动。默认 200。 */
|
||||
maxHeight?: number;
|
||||
}
|
||||
|
||||
/**
|
||||
* 自适应高度的多行输入框。
|
||||
*
|
||||
* 内容增多时自动扩展高度（输入框处于底部布局时即「向上扩展」），到 `maxHeight` 后框内滚动。
|
||||
* 体验对齐通用问答页（`/page/strategy/qa/general`）的 tdesign `autosize` 输入框，但保留原生
|
||||
* `<textarea>`，从而沿用各调用处自带的 Tailwind 样式 / 边框 / 占位色，无需 TDesign 包裹层。
|
||||
*
|
||||
* 用法：当成普通受控 `<textarea>` 用（传 `value` + `onChange`）。高度在布局阶段随 `value` 重算，
|
||||
* 也在 `onChange` 时即时重算（兼顾非受控/输入法场景）。最小高度交给调用方的 `min-h-*` className
|
||||
* 或 `rows` 控制——本组件不强加最小高度。换行/发送（Enter/Shift+Enter）逻辑由调用方的
|
||||
* `onKeyDown` 决定。
|
||||
*/
|
||||
export const AutoGrowTextarea = React.forwardRef<
|
||||
HTMLTextAreaElement,
|
||||
AutoGrowTextareaProps
|
||||
>(function AutoGrowTextarea(
|
||||
{ maxHeight = 200, value, onChange, onInput, style, rows = 1, ...rest },
|
||||
ref,
|
||||
) {
|
||||
const innerRef = React.useRef<HTMLTextAreaElement | null>(null);
|
||||
React.useImperativeHandle(ref, () => innerRef.current as HTMLTextAreaElement);
|
||||
|
||||
const resize = React.useCallback(() => {
|
||||
const el = innerRef.current;
|
||||
if (!el) return;
|
||||
el.style.height = "auto";
|
||||
const next = Math.min(el.scrollHeight, maxHeight);
|
||||
el.style.height = `${next}px`;
|
||||
el.style.overflowY = el.scrollHeight > maxHeight ? "auto" : "hidden";
|
||||
}, [maxHeight]);
|
||||
|
||||
// 受控值变化（含外部清空、粘贴大段文本）后在布局阶段重算，避免闪烁。
|
||||
React.useLayoutEffect(() => {
|
||||
resize();
|
||||
}, [resize, value]);
|
||||
|
||||
return (
|
||||
<textarea
|
||||
ref={innerRef}
|
||||
rows={rows}
|
||||
value={value}
|
||||
onChange={onChange}
|
||||
onInput={(e) => {
|
||||
onInput?.(e);
|
||||
resize();
|
||||
}}
|
||||
style={{ resize: "none", ...style }}
|
||||
{...rest}
|
||||
/>
|
||||
);
|
||||
});
|
||||
@@ -1,46 +0,0 @@
|
||||
"use client";
|
||||
|
||||
import * as React from "react";
|
||||
import { Check } from "lucide-react";
|
||||
|
||||
import { cn } from "@/lib/utils";
|
||||
|
||||
export interface CheckboxProps
|
||||
extends Omit<
|
||||
React.ButtonHTMLAttributes<HTMLButtonElement>,
|
||||
"onChange" | "checked" | "type"
|
||||
> {
|
||||
checked?: boolean;
|
||||
onCheckedChange?: (checked: boolean) => void;
|
||||
}
|
||||
|
||||
/**
|
||||
* 轻量级 Checkbox（不依赖 @radix-ui/react-checkbox，适配离线部署）。
|
||||
* API 与 Radix/shadcn 对齐：受控 `checked` + `onCheckedChange(next)`。
|
||||
*/
|
||||
export const Checkbox = React.forwardRef<HTMLButtonElement, CheckboxProps>(
|
||||
({ className, checked = false, onCheckedChange, disabled, ...props }, ref) => {
|
||||
return (
|
||||
<button
|
||||
ref={ref}
|
||||
type="button"
|
||||
role="checkbox"
|
||||
aria-checked={checked}
|
||||
disabled={disabled}
|
||||
data-state={checked ? "checked" : "unchecked"}
|
||||
onClick={() => onCheckedChange?.(!checked)}
|
||||
className={cn(
|
||||
"peer flex size-4 shrink-0 items-center justify-center rounded-[4px] border border-input shadow-sm outline-none transition-colors",
|
||||
"focus-visible:ring-2 focus-visible:ring-ring focus-visible:ring-offset-2",
|
||||
"disabled:cursor-not-allowed disabled:opacity-50",
|
||||
checked && "border-primary bg-primary text-primary-foreground",
|
||||
className,
|
||||
)}
|
||||
{...props}
|
||||
>
|
||||
{checked && <Check className="size-3.5" strokeWidth={3} />}
|
||||
</button>
|
||||
);
|
||||
},
|
||||
);
|
||||
Checkbox.displayName = "Checkbox";
|
||||
@@ -4,38 +4,28 @@ import * as React from "react";
|
||||
import * as DialogPrimitive from "@radix-ui/react-dialog";
|
||||
import { XIcon } from "lucide-react";
|
||||
|
||||
import { isWujieEnv } from "@/core/embed/wujie";
|
||||
import { cn } from "@/lib/utils";
|
||||
|
||||
function Dialog({
|
||||
modal,
|
||||
...props
|
||||
}: React.ComponentProps<typeof DialogPrimitive.Root>) {
|
||||
// 无界嵌入时 Radix modal 会给 body 注入异常 margin-right，导致右侧空白。
|
||||
// 默认降为 non-modal；显式传 modal={true} 仍可强制 modal。
|
||||
return (
|
||||
<DialogPrimitive.Root
|
||||
data-slot="dialog"
|
||||
modal={modal ?? !isWujieEnv()}
|
||||
{...props}
|
||||
/>
|
||||
);
|
||||
return <DialogPrimitive.Root data-slot="dialog" {...props} />;
|
||||
}
|
||||
|
||||
const DialogTrigger = React.forwardRef<
|
||||
React.ElementRef<typeof DialogPrimitive.Trigger>,
|
||||
React.ComponentPropsWithoutRef<typeof DialogPrimitive.Trigger>
|
||||
>(({ ...props }, ref) => {
|
||||
return (
|
||||
<DialogPrimitive.Trigger
|
||||
ref={ref}
|
||||
data-slot="dialog-trigger"
|
||||
{...props}
|
||||
/>
|
||||
);
|
||||
});
|
||||
|
||||
DialogTrigger.displayName = DialogPrimitive.Trigger.displayName;
|
||||
const DialogTrigger = React.forwardRef<
|
||||
React.ElementRef<typeof DialogPrimitive.Trigger>,
|
||||
React.ComponentPropsWithoutRef<typeof DialogPrimitive.Trigger>
|
||||
>(({ ...props }, ref) => {
|
||||
return (
|
||||
<DialogPrimitive.Trigger
|
||||
ref={ref}
|
||||
data-slot="dialog-trigger"
|
||||
{...props}
|
||||
/>
|
||||
);
|
||||
});
|
||||
|
||||
DialogTrigger.displayName = DialogPrimitive.Trigger.displayName;
|
||||
|
||||
function DialogPortal({
|
||||
...props
|
||||
@@ -43,68 +33,68 @@ function DialogPortal({
|
||||
return <DialogPrimitive.Portal data-slot="dialog-portal" {...props} />;
|
||||
}
|
||||
|
||||
const DialogClose = React.forwardRef<
|
||||
React.ElementRef<typeof DialogPrimitive.Close>,
|
||||
React.ComponentPropsWithoutRef<typeof DialogPrimitive.Close>
|
||||
>(({ ...props }, ref) => {
|
||||
return <DialogPrimitive.Close ref={ref} data-slot="dialog-close" {...props} />;
|
||||
});
|
||||
const DialogClose = React.forwardRef<
|
||||
React.ElementRef<typeof DialogPrimitive.Close>,
|
||||
React.ComponentPropsWithoutRef<typeof DialogPrimitive.Close>
|
||||
>(({ ...props }, ref) => {
|
||||
return <DialogPrimitive.Close ref={ref} data-slot="dialog-close" {...props} />;
|
||||
});
|
||||
|
||||
DialogClose.displayName = DialogPrimitive.Close.displayName;
|
||||
|
||||
DialogClose.displayName = DialogPrimitive.Close.displayName;
|
||||
const DialogOverlay = React.forwardRef<
|
||||
React.ElementRef<typeof DialogPrimitive.Overlay>,
|
||||
React.ComponentPropsWithoutRef<typeof DialogPrimitive.Overlay>
|
||||
>(({ className, ...props }, ref) => {
|
||||
return (
|
||||
<DialogPrimitive.Overlay
|
||||
ref={ref}
|
||||
data-slot="dialog-overlay"
|
||||
className={cn(
|
||||
"data-[state=open]:animate-in data-[state=closed]:animate-out data-[state=closed]:fade-out-0 data-[state=open]:fade-in-0 fixed inset-0 z-50 bg-black/50",
|
||||
className,
|
||||
)}
|
||||
{...props}
|
||||
/>
|
||||
);
|
||||
});
|
||||
|
||||
DialogOverlay.displayName = DialogPrimitive.Overlay.displayName;
|
||||
|
||||
const DialogOverlay = React.forwardRef<
|
||||
React.ElementRef<typeof DialogPrimitive.Overlay>,
|
||||
React.ComponentPropsWithoutRef<typeof DialogPrimitive.Overlay>
|
||||
>(({ className, ...props }, ref) => {
|
||||
return (
|
||||
<DialogPrimitive.Overlay
|
||||
ref={ref}
|
||||
data-slot="dialog-overlay"
|
||||
className={cn(
|
||||
"data-[state=open]:animate-in data-[state=closed]:animate-out data-[state=closed]:fade-out-0 data-[state=open]:fade-in-0 fixed inset-0 z-50 bg-black/50",
|
||||
className,
|
||||
)}
|
||||
{...props}
|
||||
/>
|
||||
);
|
||||
});
|
||||
|
||||
DialogOverlay.displayName = DialogPrimitive.Overlay.displayName;
|
||||
|
||||
const DialogContent = React.forwardRef<
|
||||
React.ElementRef<typeof DialogPrimitive.Content>,
|
||||
React.ComponentPropsWithoutRef<typeof DialogPrimitive.Content> & {
|
||||
showCloseButton?: boolean;
|
||||
}
|
||||
>(({ className, children, showCloseButton = true, ...props }, ref) => {
|
||||
return (
|
||||
<DialogPortal data-slot="dialog-portal">
|
||||
<DialogOverlay />
|
||||
<DialogPrimitive.Content
|
||||
ref={ref}
|
||||
data-slot="dialog-content"
|
||||
className={cn(
|
||||
"bg-background data-[state=open]:animate-in data-[state=closed]:animate-out data-[state=closed]:fade-out-0 data-[state=open]:fade-in-0 data-[state=closed]:zoom-out-95 data-[state=open]:zoom-in-95 fixed top-[50%] left-[50%] z-50 grid w-full max-w-[calc(100%-2rem)] translate-x-[-50%] translate-y-[-50%] gap-4 rounded-lg border p-6 shadow-lg duration-200 outline-none sm:max-w-lg",
|
||||
className,
|
||||
)}
|
||||
{...props}
|
||||
>
|
||||
{children}
|
||||
{showCloseButton && (
|
||||
<DialogPrimitive.Close
|
||||
data-slot="dialog-close"
|
||||
className="ring-offset-background focus:ring-ring data-[state=open]:bg-accent data-[state=open]:text-muted-foreground absolute top-4 right-4 rounded-xs opacity-70 transition-opacity hover:opacity-100 focus:ring-2 focus:ring-offset-2 focus:outline-hidden disabled:pointer-events-none [&_svg]:pointer-events-none [&_svg]:shrink-0 [&_svg:not([class*='size-'])]:size-4"
|
||||
>
|
||||
<XIcon />
|
||||
<span className="sr-only">Close</span>
|
||||
</DialogPrimitive.Close>
|
||||
)}
|
||||
</DialogPrimitive.Content>
|
||||
</DialogPortal>
|
||||
);
|
||||
});
|
||||
|
||||
DialogContent.displayName = DialogPrimitive.Content.displayName;
|
||||
const DialogContent = React.forwardRef<
|
||||
React.ElementRef<typeof DialogPrimitive.Content>,
|
||||
React.ComponentPropsWithoutRef<typeof DialogPrimitive.Content> & {
|
||||
showCloseButton?: boolean;
|
||||
}
|
||||
>(({ className, children, showCloseButton = true, ...props }, ref) => {
|
||||
return (
|
||||
<DialogPortal data-slot="dialog-portal">
|
||||
<DialogOverlay />
|
||||
<DialogPrimitive.Content
|
||||
ref={ref}
|
||||
data-slot="dialog-content"
|
||||
className={cn(
|
||||
"bg-background data-[state=open]:animate-in data-[state=closed]:animate-out data-[state=closed]:fade-out-0 data-[state=open]:fade-in-0 data-[state=closed]:zoom-out-95 data-[state=open]:zoom-in-95 fixed top-[50%] left-[50%] z-50 grid w-full max-w-[calc(100%-2rem)] translate-x-[-50%] translate-y-[-50%] gap-4 rounded-lg border p-6 shadow-lg duration-200 outline-none sm:max-w-lg",
|
||||
className,
|
||||
)}
|
||||
{...props}
|
||||
>
|
||||
{children}
|
||||
{showCloseButton && (
|
||||
<DialogPrimitive.Close
|
||||
data-slot="dialog-close"
|
||||
className="ring-offset-background focus:ring-ring data-[state=open]:bg-accent data-[state=open]:text-muted-foreground absolute top-4 right-4 rounded-xs opacity-70 transition-opacity hover:opacity-100 focus:ring-2 focus:ring-offset-2 focus:outline-hidden disabled:pointer-events-none [&_svg]:pointer-events-none [&_svg]:shrink-0 [&_svg:not([class*='size-'])]:size-4"
|
||||
>
|
||||
<XIcon />
|
||||
<span className="sr-only">Close</span>
|
||||
</DialogPrimitive.Close>
|
||||
)}
|
||||
</DialogPrimitive.Content>
|
||||
</DialogPortal>
|
||||
);
|
||||
});
|
||||
|
||||
DialogContent.displayName = DialogPrimitive.Content.displayName;
|
||||
|
||||
function DialogHeader({ className, ...props }: React.ComponentProps<"div">) {
|
||||
return (
|
||||
|
||||
@@ -144,24 +144,21 @@ function InputGroupInput({
|
||||
);
|
||||
}
|
||||
|
||||
const InputGroupTextarea = React.forwardRef<
|
||||
HTMLTextAreaElement,
|
||||
React.ComponentProps<"textarea">
|
||||
>(function InputGroupTextarea({ className, ...props }, ref) {
|
||||
function InputGroupTextarea({
|
||||
className,
|
||||
...props
|
||||
}: React.ComponentProps<"textarea">) {
|
||||
return (
|
||||
<Textarea
|
||||
data-slot="input-group-control"
|
||||
ref={ref}
|
||||
<Textarea
|
||||
data-slot="input-group-control"
|
||||
className={cn(
|
||||
"flex-1 resize-none rounded-none border-0 bg-transparent py-3 shadow-none focus-visible:ring-0 dark:bg-transparent",
|
||||
className,
|
||||
)}
|
||||
{...props}
|
||||
/>
|
||||
);
|
||||
});
|
||||
|
||||
InputGroupTextarea.displayName = "InputGroupTextarea";
|
||||
{...props}
|
||||
/>
|
||||
);
|
||||
}
|
||||
|
||||
export {
|
||||
InputGroup,
|
||||
|
||||
@@ -1,96 +0,0 @@
|
||||
"use client";
|
||||
|
||||
import { ChevronLeftIcon, ChevronRightIcon } from "lucide-react";
|
||||
|
||||
import { Button } from "@/components/ui/button";
|
||||
import { cn } from "@/lib/utils";
|
||||
|
||||
/** Compute the page-number window to render, inserting -1 sentinels for gaps.
|
||||
* Always shows first/last and a window around the current page. */
|
||||
function buildPageItems(current: number, totalPages: number): number[] {
|
||||
if (totalPages <= 7) {
|
||||
return Array.from({ length: totalPages }, (_, i) => i + 1);
|
||||
}
|
||||
const items = new Set<number>([1, totalPages, current]);
|
||||
for (const d of [1, 2]) {
|
||||
if (current - d >= 1) items.add(current - d);
|
||||
if (current + d <= totalPages) items.add(current + d);
|
||||
}
|
||||
const sorted = [...items].sort((a, b) => a - b);
|
||||
const out: number[] = [];
|
||||
let prev = 0;
|
||||
for (const p of sorted) {
|
||||
if (prev && p - prev > 1) out.push(-1); // gap sentinel
|
||||
out.push(p);
|
||||
prev = p;
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
export interface PaginationProps {
|
||||
page: number;
|
||||
pageSize: number;
|
||||
total: number;
|
||||
onPageChange: (page: number) => void;
|
||||
className?: string;
|
||||
}
|
||||
|
||||
/** Controlled, compact pagination control. Renders nothing when a single page. */
|
||||
export function Pagination({
|
||||
page,
|
||||
pageSize,
|
||||
total,
|
||||
onPageChange,
|
||||
className,
|
||||
}: PaginationProps) {
|
||||
const totalPages = Math.max(1, Math.ceil(total / Math.max(pageSize, 1)));
|
||||
if (totalPages <= 1) return null;
|
||||
|
||||
const items = buildPageItems(page, totalPages);
|
||||
|
||||
return (
|
||||
<nav
|
||||
className={cn("flex items-center justify-center gap-1", className)}
|
||||
aria-label="分页"
|
||||
>
|
||||
<Button
|
||||
variant="outline"
|
||||
size="icon-sm"
|
||||
disabled={page <= 1}
|
||||
onClick={() => onPageChange(page - 1)}
|
||||
aria-label="上一页"
|
||||
>
|
||||
<ChevronLeftIcon className="h-4 w-4" />
|
||||
</Button>
|
||||
{items.map((p, idx) =>
|
||||
p === -1 ? (
|
||||
<span
|
||||
key={`gap-${idx}`}
|
||||
className="text-muted-foreground px-1 text-sm select-none"
|
||||
>
|
||||
…
|
||||
</span>
|
||||
) : (
|
||||
<Button
|
||||
key={p}
|
||||
variant={p === page ? "default" : "outline"}
|
||||
size="icon-sm"
|
||||
onClick={() => onPageChange(p)}
|
||||
aria-current={p === page ? "page" : undefined}
|
||||
>
|
||||
{p}
|
||||
</Button>
|
||||
),
|
||||
)}
|
||||
<Button
|
||||
variant="outline"
|
||||
size="icon-sm"
|
||||
disabled={page >= totalPages}
|
||||
onClick={() => onPageChange(page + 1)}
|
||||
aria-label="下一页"
|
||||
>
|
||||
<ChevronRightIcon className="h-4 w-4" />
|
||||
</Button>
|
||||
</nav>
|
||||
);
|
||||
}
|
||||
@@ -22,31 +22,10 @@ function ResizablePanelGroup({
|
||||
);
|
||||
}
|
||||
|
||||
// react-resizable-panels v4 changed size semantics: a bare *number* is now
|
||||
// interpreted as PIXELS, while a unit-less *string* is a percentage. Every call
|
||||
// site in this app passes sizes as percentages (the v2/v3 convention), so a
|
||||
// value like `maxSize={60}` would otherwise pin a panel to 60px and freeze the
|
||||
// layout/drag. Normalize numbers to "<n>%" to preserve the intended percentage.
|
||||
const asPercent = (v: number | string | undefined) =>
|
||||
typeof v === "number" ? `${v}%` : v;
|
||||
|
||||
function ResizablePanel({
|
||||
defaultSize,
|
||||
minSize,
|
||||
maxSize,
|
||||
collapsedSize,
|
||||
...props
|
||||
}: React.ComponentProps<typeof ResizablePrimitive.Panel>) {
|
||||
return (
|
||||
<ResizablePrimitive.Panel
|
||||
data-slot="resizable-panel"
|
||||
defaultSize={asPercent(defaultSize)}
|
||||
minSize={asPercent(minSize)}
|
||||
maxSize={asPercent(maxSize)}
|
||||
collapsedSize={asPercent(collapsedSize)}
|
||||
{...props}
|
||||
/>
|
||||
);
|
||||
return <ResizablePrimitive.Panel data-slot="resizable-panel" {...props} />;
|
||||
}
|
||||
|
||||
function ResizableHandle({
|
||||
@@ -60,19 +39,14 @@ function ResizableHandle({
|
||||
<ResizablePrimitive.Separator
|
||||
data-slot="resizable-handle"
|
||||
className={cn(
|
||||
// Thin visible divider, but a generous invisible hit area on both sides
|
||||
// (after:w-4 → 16px) so the handle is easy to grab and drag. Hovering /
|
||||
// dragging tints the divider so it's obvious it's interactive.
|
||||
"bg-border hover:bg-primary/60 active:bg-primary relative flex w-px items-center justify-center transition-colors",
|
||||
"after:absolute after:inset-y-0 after:left-1/2 after:w-4 after:-translate-x-1/2 after:cursor-col-resize",
|
||||
"focus-visible:ring-ring focus-visible:ring-1 focus-visible:ring-offset-1 focus-visible:outline-hidden",
|
||||
"bg-border focus-visible:ring-ring relative flex w-px items-center justify-center after:absolute after:inset-y-0 after:left-1/2 after:w-1 after:-translate-x-1/2 focus-visible:ring-1 focus-visible:ring-offset-1 focus-visible:outline-hidden data-[panel-group-direction=vertical]:h-px data-[panel-group-direction=vertical]:w-full data-[panel-group-direction=vertical]:after:left-0 data-[panel-group-direction=vertical]:after:h-1 data-[panel-group-direction=vertical]:after:w-full data-[panel-group-direction=vertical]:after:translate-x-0 data-[panel-group-direction=vertical]:after:-translate-y-1/2 [&[data-panel-group-direction=vertical]>div]:rotate-90",
|
||||
className,
|
||||
)}
|
||||
{...props}
|
||||
>
|
||||
{withHandle && (
|
||||
<div className="bg-border z-10 flex h-8 w-4 items-center justify-center rounded-sm border shadow-sm">
|
||||
<GripVerticalIcon className="size-3.5" />
|
||||
<div className="bg-border z-10 flex h-4 w-3 items-center justify-center rounded-xs border">
|
||||
<GripVerticalIcon className="size-2.5" />
|
||||
</div>
|
||||
)}
|
||||
</ResizablePrimitive.Separator>
|
||||
|
||||
Loaded 100 of 1633 files, more files were not shown because too many files have changed in this diff.
Show more
Reference in new issue
Block a user