From 28098b315558667635f68e663d43d07cf3d4497d Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E5=90=89=E9=91=AB=E5=8D=87?= Date: Fri, 12 Jun 2026 11:09:32 +0800 Subject: [PATCH] =?UTF-8?q?feat:=20=E7=9F=A5=E8=AF=86=E5=BA=93=E5=AE=9E?= =?UTF-8?q?=E4=BD=93=E7=AD=9B=E9=80=89=E3=80=81=E6=8A=80=E8=83=BD=E7=99=BD?= =?UTF-8?q?=E5=90=8D=E5=8D=95=E6=94=B6=E7=B4=A7=E3=80=81=E5=B5=8C=E5=85=A5?= =?UTF-8?q?=E4=B8=BB=E9=A2=98=E4=B8=8E=E6=8F=90=E9=97=AE=E7=BB=84=E4=BB=B6?= =?UTF-8?q?=E4=BC=98=E5=8C=96?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - 知识库笔记列表支持按实体精确筛选（路由/服务/仓储层透传，关系图索引查询） - 自定义智能体技能白名单在 skill_view 与 skill_manage 的 list/view 别名上同样生效； config 未声明 skills 时视为无可见技能 - 嵌入圆桌主题解析默认深蓝，强制主题期间不再写 localStorage 覆盖主站选择 - 智能体提问建议拆出 agent-question-chips 组件，编辑器与详情面板联动优化 - 新增 qa-client 问答测试脚本 Co-Authored-By: Claude Fable 5 --- frontend-web/src/App.tsx | 26 +++- .../components/wiki/RelatedNotesDialog.tsx | 100 +++++++++++-- .../src/components/wiki/WikiHeader.tsx | 13 +- .../workspace/agents/agent-detail-panel.tsx | 41 ++++-- .../workspace/agents/agent-question-chips.tsx | 78 ++++++++++ .../agents/agent-question-editor.tsx | 33 ++++- .../settings/skill-settings-page.tsx | 13 ++ frontend-web/src/contexts/ThemeContext.tsx | 16 ++- frontend-web/src/core/embed/params.ts | 11 ++ frontend-web/src/core/knowledge/api.ts | 1 + frontend-web/src/core/knowledge/types.ts | 2 + frontend-web/src/pages/AgentChatPage.tsx | 31 +--- frontend-web/src/pages/EmbedChatPage.tsx | 15 +- .../src/pages/EmbedRoundtablePage.tsx | 26 +--- frontend-web/src/pages/NewAgentPage.tsx | 8 +- offline-backend-20260512/backend/CLAUDE.md | 3 +- .../backend/app/gateway/routers/knowledge.py | 3 +- .../harness/deerflow/knowledge/repository.py | 11 ++ .../harness/deerflow/knowledge/service.py | 2 + .../deerflow/tools/builtins/skill_tools.py | 48 +++++-- .../deerflow/tools/skill_manage_tool.py | 10 +- qa-client/qa_client.py | 135 ++++++++++++++++++ 22 files changed, 500 insertions(+), 126 deletions(-) create mode 100644 frontend-web/src/components/workspace/agents/agent-question-chips.tsx create mode 100644 qa-client/qa_client.py diff --git a/frontend-web/src/App.tsx b/frontend-web/src/App.tsx index 277f3974..23999e96 100644 --- a/frontend-web/src/App.tsx +++ b/frontend-web/src/App.tsx @@ -9,6 +9,11 @@ import { I18nProvider } from "@/core/i18n/context"; import { detectLocale } from "@/core/i18n/locale"; import { ChatProvider } from "@/contexts/ChatContext"; import { ThemeProvider } from "@/contexts/ThemeContext"; +import { + resolveEmbedRoundtableTheme, + resolveEmbedTheme, +} from "@/core/embed/params"; +import { readWujieProps } from "@/core/embed/wujie"; import { HashRouter, useLocation } from "react-router-dom"; @@ -29,7 +34,26 @@ import { NotFoundPage } from "./pages/NotFoundPage"; function AppProviders({ children }: { children: React.ReactNode }) { const locale = useMemo(() => detectLocale(), []); const location = useLocation(); - const forcedTheme = location.pathname === "/landing" ? "light" : undefined; + // 嵌入路由的主题用 forcedTheme 强制：只影响当前渲染，不写 localStorage。 + // 千万不要在嵌入页里调 ThemeContext.setTheme —— 它会持久化主题并打上 + // 「用户已显式选择」标记。无界子应用与主站同域共享 localStorage，宿主 + // 注入的 theme 会就此覆盖用户在主站自己选的主题（且每次打开宿主都覆盖）。 + const forcedTheme = useMemo(() => { + if (location.pathname === "/landing") return "light"; + 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; + } + return undefined; + }, [location.pathname, location.search]); return ( (value: T, delay = 300): T { + const [debounced, setDebounced] = useState(value); + useEffect(() => { + const id = window.setTimeout(() => setDebounced(value), delay); + return () => window.clearTimeout(id); + }, [value, delay]); + return debounced; +} + /** * Click a tag / entity chip → list every other note that shares it. - * Tags use the backend's exact `tag` filter; entities fall back to a keyword - * match (entity names appear verbatim in titles/summaries/bodies). + * Server-side keyword search + pagination so huge graphs stay usable: + * tags use the backend's exact `tag` filter, entities the exact `entity` + * filter (note→entity relations), both combinable with `keyword`. */ export function RelatedNotesDialog({ filter, @@ -34,17 +57,33 @@ export function RelatedNotesDialog({ onOpenChange: (open: boolean) => void; onSelect: (noteId: string) => void; }) { - const params = - filter?.kind === "tag" - ? { tag: filter.value, limit: 50 } - : filter?.kind === "entity" - ? { keyword: filter.value, limit: 50 } - : { limit: 0 }; + const [search, setSearch] = useState(""); + const debouncedSearch = useDebounced(search.trim()); + const [page, setPage] = useState(0); + + // New chip or new search term → back to the first page. + useEffect(() => { + setPage(0); + }, [filter?.kind, filter?.value, debouncedSearch]); + // Reset the search box whenever a different chip is opened. + useEffect(() => { + setSearch(""); + }, [filter?.kind, filter?.value]); + + const params = { + ...(filter?.kind === "tag" ? { tag: filter.value } : {}), + ...(filter?.kind === "entity" ? { entity: filter.value } : {}), + keyword: debouncedSearch || undefined, + limit: filter ? PAGE_SIZE : 0, + offset: page * PAGE_SIZE, + }; const { data, isLoading, isFetching } = useKnowledgeNotes(params); const items = (data?.items ?? []).filter( (n: KnowledgeNote) => n.id !== currentNoteId, ); + const total = data?.total ?? 0; + const pageCount = Math.max(1, Math.ceil(total / PAGE_SIZE)); return ( @@ -62,17 +101,28 @@ export function RelatedNotesDialog({ {filter?.kind === "entity" ? "提到该实体的关联知识" : "使用该标签的关联知识"} + {total > 0 ? ` · 共 ${total} 条` : ""} - +
+ + setSearch(e.target.value)} + placeholder="在关联知识中搜索标题/内容..." + className="h-9 pl-8" + /> +
+ + {isLoading || (isFetching && items.length === 0) ? (
加载中...
) : items.length === 0 ? (
- 暂无其他关联知识。 + {debouncedSearch ? "没有匹配的关联知识。" : "暂无其他关联知识。"}
) : (
    @@ -104,6 +154,32 @@ export function RelatedNotesDialog({
)} + + {pageCount > 1 ? ( +
+ + 第 {page + 1} / {pageCount} 页 + +
+ + +
+
+ ) : null} ); diff --git a/frontend-web/src/components/wiki/WikiHeader.tsx b/frontend-web/src/components/wiki/WikiHeader.tsx index a82c391d..b278ad49 100644 --- a/frontend-web/src/components/wiki/WikiHeader.tsx +++ b/frontend-web/src/components/wiki/WikiHeader.tsx @@ -1,5 +1,4 @@ import { - BookOpen, Copy, Menu, Network, @@ -41,7 +40,7 @@ function HeaderButton({ ); } -/** Light 56px wiki header: logo + product name on the left, action entries on the right. */ +/** Light 56px wiki header: plain page title on the left, action entries on the right. */ export function WikiHeader({ onToggleSidebar, onImportFile, @@ -68,14 +67,8 @@ export function WikiHeader({ > -
- -
-
-
- 问答梳理 -
-
LLM Wiki · 问答梳理中心
+
+ 问答梳理
diff --git a/frontend-web/src/components/workspace/agents/agent-detail-panel.tsx b/frontend-web/src/components/workspace/agents/agent-detail-panel.tsx index 39c9d292..c11946c0 100644 --- a/frontend-web/src/components/workspace/agents/agent-detail-panel.tsx +++ b/frontend-web/src/components/workspace/agents/agent-detail-panel.tsx @@ -422,14 +422,7 @@ export function AgentDetailPanel({ ) : (
- @@ -591,14 +584,29 @@ export function AgentDetailPanel({ )} {/* ── 常用问题 ───────────────────────────────────────────────────── */} - {(editing || questions.length > 0) && ( + {(editing || questions.length > 0 || canEdit) && (
} title="常用问题" extra={ - - {(editing && draft ? draft.questions : questions).length} 个 - +
+ + {(editing && draft ? draft.questions : questions).length} 个 + + {!editing && canEdit && ( + + + + )} +
} > {editing && draft ? ( @@ -607,6 +615,15 @@ export function AgentDetailPanel({ onChange={(next) => setDraft({ ...draft, questions: next })} disabled={isSaving} /> + ) : questions.length === 0 ? ( + ) : (
{questions.map((question, index) => { diff --git a/frontend-web/src/components/workspace/agents/agent-question-chips.tsx b/frontend-web/src/components/workspace/agents/agent-question-chips.tsx new file mode 100644 index 00000000..69bbb823 --- /dev/null +++ b/frontend-web/src/components/workspace/agents/agent-question-chips.tsx @@ -0,0 +1,78 @@ +"use client"; + +import { ArrowRightIcon, ChevronDownIcon, ChevronUpIcon } from "lucide-react"; +import { useState } from "react"; + +import type { AgentQuestion } from "@/core/agents"; +import { desensitize } from "@/lib/desensitize"; +import { cn } from "@/lib/utils"; + +/** Chips shown before the rest folds behind the 更多问题 toggle. */ +const VISIBLE_COUNT = 4; + +/** + * 常用问题 prompt chips on the agent chat landing screen — light-gray + * borderless pills stacked left-aligned between the welcome hero and the + * input box. Clicking a chip sends the question's full prompt through the + * page's existing submit flow. Lists longer than {@link VISIBLE_COUNT} fold + * behind a 更多问题 toggle; an empty list renders nothing. + */ +export function AgentQuestionChips({ + questions, + onAsk, + disabled, + className, +}: { + questions: AgentQuestion[]; + onAsk: (prompt: string) => void; + disabled?: boolean; + className?: string; +}) { + const [expanded, setExpanded] = useState(false); + + if (questions.length === 0) return null; + + const visible = expanded ? questions : questions.slice(0, VISIBLE_COUNT); + const foldedCount = questions.length - VISIBLE_COUNT; + + return ( +
+ {visible.map((question, index) => ( + + ))} + {foldedCount > 0 && ( + + )} +
+ ); +} diff --git a/frontend-web/src/components/workspace/agents/agent-question-editor.tsx b/frontend-web/src/components/workspace/agents/agent-question-editor.tsx index d0f2e223..489744b4 100644 --- a/frontend-web/src/components/workspace/agents/agent-question-editor.tsx +++ b/frontend-web/src/components/workspace/agents/agent-question-editor.tsx @@ -12,10 +12,20 @@ import { Input } from "@/components/ui/input"; import { Textarea } from "@/components/ui/textarea"; import type { AgentQuestion } from "@/core/agents"; +/** Soft cap on the chip label length — beyond this the landing chip truncates. */ +const TITLE_SOFT_LIMIT = 80; + +/** The text a question row would actually persist as (panel save falls back + * title↔prompt), used for duplicate detection. */ +function effectivePrompt(q: AgentQuestion): string { + return (q.prompt.trim() || q.title.trim()).toLowerCase(); +} + /** * Editable list of an agent's 常用问题 (recommended questions). Each row has a * short `title` (the chip label) and the full `prompt` sent on click. Supports - * add / remove / inline edit / reorder (move up & down). Pure controlled + * add / remove / inline edit / reorder (move up & down), with non-blocking + * hints for over-long titles and duplicate questions. Pure controlled * component — persistence stays with the caller (edit dialog or detail panel). */ export function AgentQuestionEditor({ @@ -52,7 +62,13 @@ export function AgentQuestionEditor({ 尚未配置常用问题，点击下方按钮添加。

)} - {value.map((item, index) => ( + {value.map((item, index) => { + const titleLength = item.title.trim().length; + const prompt = effectivePrompt(item); + const duplicateOf = prompt + ? value.findIndex((other, i) => i < index && effectivePrompt(other) === prompt) + : -1; + return (
update(index, { prompt: e.currentTarget.value })} className="min-h-16 resize-none text-sm" /> + {titleLength > TITLE_SOFT_LIMIT && ( +

+ 标题建议不超过 {TITLE_SOFT_LIMIT} 字（当前 {titleLength} 字），过长会在聊天区被截断。 +

+ )} + {duplicateOf >= 0 && ( +

+ 与第 {duplicateOf + 1} 条问题内容重复。 +

+ )}
- ))} + ); + })} +
= ({ children }) => { - const { resolvedTheme, setTheme: setNextTheme } = useNextTheme(); + // forcedTheme：嵌入路由 / landing 由 AppProviders 强制的主题。强制期间 + // 它决定渲染，且本组件不得写 localStorage（嵌入子应用与主站同域共享 + // 存储，写入会覆盖用户在主站的选择）。 + const { resolvedTheme, forcedTheme, setTheme: setNextTheme } = useNextTheme(); const [mounted, setMounted] = useState(false); // Admin-configured global default. If the user hasn't explicitly chosen // a theme yet, this value drives the rendered theme — and re-syncs on @@ -144,7 +147,7 @@ export const ThemeProvider: React.FC = ({ children }) => { // 注意：不要在这里强制写入一个 fallback —— 那会让 next-themes 视作 // 「用户已选过」，admin 默认就再也注入不进来了。 useEffect(() => { - if (!mounted) return; + if (!mounted || forcedTheme) return; const current = localStorage.getItem(STORAGE_KEY); const legacy = localStorage.getItem('theme'); @@ -162,24 +165,25 @@ export const ThemeProvider: React.FC = ({ children }) => { if (legacy) { localStorage.removeItem('theme'); } - }, [mounted, setNextTheme]); + }, [mounted, forcedTheme, setNextTheme]); // Boot-time sync: when the user has never picked a theme themselves, // apply whatever the admin configured. Re-runs whenever the admin // setting changes (so a logged-in user sees the new default after // refreshing without having to clear cookies). useEffect(() => { - if (!mounted) return; + if (!mounted || forcedTheme) return; if (hasExplicitChoice()) return; const adminDefault = appearance?.default_theme; if (!adminDefault) return; if (resolvedTheme !== adminDefault) { setNextTheme(adminDefault); } - }, [mounted, appearance?.default_theme, resolvedTheme, setNextTheme]); + }, [mounted, forcedTheme, appearance?.default_theme, resolvedTheme, setNextTheme]); + const effectiveTheme = forcedTheme ?? resolvedTheme; const theme: ThemeType = mounted - ? resolvedTheme === 'dark' + ? effectiveTheme === 'dark' ? 'dark' : 'light' : readPersistedTheme(); diff --git a/frontend-web/src/core/embed/params.ts b/frontend-web/src/core/embed/params.ts index 677ae364..be4fb8e0 100644 --- a/frontend-web/src/core/embed/params.ts +++ b/frontend-web/src/core/embed/params.ts @@ -44,3 +44,14 @@ export function resolveEmbedChatParams(search: string): EmbedChatParams { export function resolveEmbedTheme(themeParam: string): "dark" | "light" { return resolveThemeFromQuery(themeParam || null); } + +/** + * 把宿主传入的主题标识映射成 light/dark。**默认深蓝（dark）**： + * 只有明确传 light/红白 才走浅色，其余（dark / dark-blue / 深蓝 / 空）一律深蓝。 + * 不复用 `resolveEmbedTheme`（它只认严格的 "dark-blue"，传 "dark" 会被判成浅色）。 + */ +export function resolveEmbedRoundtableTheme(raw: string): "dark" | "light" { + const v = raw.trim().toLowerCase(); + if (v === "light" || v === "red" || v === "红白") return "light"; + return "dark"; +} diff --git a/frontend-web/src/core/knowledge/api.ts b/frontend-web/src/core/knowledge/api.ts index 3323bd5d..182f34f5 100644 --- a/frontend-web/src/core/knowledge/api.ts +++ b/frontend-web/src/core/knowledge/api.ts @@ -32,6 +32,7 @@ export async function listKnowledgeNotes(params: KnowledgeListParams = {}) { const qs = new URLSearchParams(); if (params.keyword) qs.set("keyword", params.keyword); if (params.tag) qs.set("tag", params.tag); + if (params.entity) qs.set("entity", params.entity); if (params.source_type) qs.set("source_type", params.source_type); if (params.status) qs.set("status", params.status); qs.set("limit", String(params.limit ?? 20)); diff --git a/frontend-web/src/core/knowledge/types.ts b/frontend-web/src/core/knowledge/types.ts index 676b7ab2..7699dae4 100644 --- a/frontend-web/src/core/knowledge/types.ts +++ b/frontend-web/src/core/knowledge/types.ts @@ -54,6 +54,8 @@ export interface KnowledgeListResponse { export interface KnowledgeListParams { keyword?: string; tag?: string; + /** Exact entity-name filter (matched via the knowledge relation graph). */ + entity?: string; source_type?: string; status?: string; limit?: number; diff --git a/frontend-web/src/pages/AgentChatPage.tsx b/frontend-web/src/pages/AgentChatPage.tsx index 8d243775..91eb96a8 100644 --- a/frontend-web/src/pages/AgentChatPage.tsx +++ b/frontend-web/src/pages/AgentChatPage.tsx @@ -8,6 +8,7 @@ import { useCallback, useEffect, useMemo, useRef, useState } from "react"; import type { PromptInputMessage } from "@/components/ai-elements/prompt-input"; import { Button } from "@/components/ui/button"; import { AgentDetailPanel } from "@/components/workspace/agents/agent-detail-panel"; +import { AgentQuestionChips } from "@/components/workspace/agents/agent-question-chips"; import { AgentWelcome } from "@/components/workspace/agent-welcome"; import { ArtifactTrigger } from "@/components/workspace/artifacts"; import { ChatBox, useThreadChat } from "@/components/workspace/chats"; @@ -28,7 +29,6 @@ import { useThreadStream } from "@/core/threads/hooks"; import { textOfMessage } from "@/core/threads/utils"; import { env } from "@/env"; import { desensitize } from "@/lib/desensitize"; -import { cn } from "@/lib/utils"; export default function AgentChatPage() { const routeBase = useRouteBase(); @@ -235,29 +235,12 @@ export default function AgentChatPage() { className="mb-7" />
- {commonQuestions.length > 0 && ( -
- {commonQuestions.map((question, index) => ( - - ))} -
- )} + handleSubmit({ text, files: [] })} + />
(); const searchParams = useSearchParams(); - const { setTheme } = useTheme(); + // 主题不在这里处理 —— ?theme= 由 AppProviders 通过 forcedTheme 强制 + // （不落 localStorage），避免污染主站用户自己选的主题。 const embedParams = resolveEmbedChatParams(searchParams.toString()); const [authReady, setAuthReady] = useState(false); const [authError, setAuthError] = useState(null); @@ -51,12 +48,6 @@ export default function EmbedChatPage() { const [settings] = useThreadSettings(threadId); const { tokenUsageEnabled } = useModels(); - useEffect(() => { - if (embedParams.theme) { - setTheme(resolveEmbedTheme(embedParams.theme)); - } - }, [embedParams.theme, setTheme]); - useEffect(() => { let cancelled = false; async function runLogin() { diff --git a/frontend-web/src/pages/EmbedRoundtablePage.tsx b/frontend-web/src/pages/EmbedRoundtablePage.tsx index 2d626389..77851c9d 100644 --- a/frontend-web/src/pages/EmbedRoundtablePage.tsx +++ b/frontend-web/src/pages/EmbedRoundtablePage.tsx @@ -3,7 +3,6 @@ import { useEffect, useState } from "react"; import { useSearchParams } from "next/navigation"; -import { useTheme } from "@/contexts/ThemeContext"; import { loginByUsername } from "@/core/auth/api"; import { setStoredAuth } from "@/core/auth"; import { readWujieProps, isWujieEnv } from "@/core/embed/wujie"; @@ -13,17 +12,6 @@ import "@/roundtable-planning/styles/roundtable-planning.css"; const DEFAULT_USERNAME = "guest"; -/** - * 把宿主传入的主题标识映射成 light/dark。**默认深蓝（dark）**： - * 只有明确传 light/红白 才走浅色，其余（dark / dark-blue / 深蓝 / 空）一律深蓝。 - * 不复用 `resolveEmbedTheme`（它只认严格的 "dark-blue"，传 "dark" 会被判成浅色）。 - */ -function resolveEmbedRoundtableTheme(raw: string): "dark" | "light" { - const v = raw.trim().toLowerCase(); - if (v === "light" || v === "red" || v === "红白") return "light"; - return "dark"; -} - /** * 圆桌规划嵌入页（独立入口，**与现有 wujie ticket 登录无关**）。 * @@ -37,9 +25,10 @@ function resolveEmbedRoundtableTheme(raw: string): "dark" | "light" { */ export default function EmbedRoundtablePage() { const searchParams = useSearchParams(); - const { setTheme } = useTheme(); - // username / password / theme：wujie props 优先，URL query 兜底。 + // username / password：wujie props 优先，URL query 兜底。 + // 主题不在这里处理 —— 嵌入路由的主题由 AppProviders 通过 forcedTheme + // 强制（不落 localStorage），避免污染主站用户自己选的主题。 const wp = readWujieProps(); const username = (typeof wp.username === "string" && wp.username.trim()) || @@ -53,19 +42,10 @@ export default function EmbedRoundtablePage() { (typeof wp.password === "string" && wp.password) || searchParams.get("password") || undefined; - // 主题色：宿主传入优先；默认「深蓝」= dark 主题（其主色为蓝，oklch hue 259）。 - const themeParam = - (typeof wp.theme === "string" && wp.theme.trim()) || - searchParams.get("theme")?.trim() || - "dark"; const [authReady, setAuthReady] = useState(false); const [authError, setAuthError] = useState(null); - useEffect(() => { - setTheme(resolveEmbedRoundtableTheme(themeParam)); - }, [themeParam, setTheme]); - useEffect(() => { if (!isWujieEnv()) return; document.documentElement.classList.add("wujie-embed"); diff --git a/frontend-web/src/pages/NewAgentPage.tsx b/frontend-web/src/pages/NewAgentPage.tsx index 29cb8756..822738e6 100644 --- a/frontend-web/src/pages/NewAgentPage.tsx +++ b/frontend-web/src/pages/NewAgentPage.tsx @@ -184,13 +184,11 @@ export default function NewAgentPage() { // Name is intentionally NOT required here so the user can start chatting // immediately. If they leave it blank, ensureAgent() generates a placeholder // and writes it back into the input so they can rename it later. - if (enabledSkills.length > 0 && selectedSkills.length === 0) { - setFormError("请至少选择一个技能"); - return null; - } + // Zero skills is a valid choice: the backend treats an empty/omitted skill + // list as "no skills" for custom agents (never "all skills"). setFormError(""); return nameInput.trim(); - }, [enabledSkills.length, nameInput, selectedSkills.length]); + }, [nameInput]); const ensureAgent = useCallback(async () => { const provided = validateForm(); diff --git a/offline-backend-20260512/backend/CLAUDE.md b/offline-backend-20260512/backend/CLAUDE.md index 595b68c1..4ee79f54 100644 --- a/offline-backend-20260512/backend/CLAUDE.md +++ b/offline-backend-20260512/backend/CLAUDE.md @@ -251,7 +251,7 @@ FastAPI application on port 8001 with health check at `GET /health`. Set `GATEWA | **Scheduled Tasks** (`/api/scheduled-tasks`) | Per-user cron-like agent tasks. `GET/POST /` - list/create owned tasks; `PATCH/DELETE /{id}` - update/delete (owner); `POST /{id}/{pause,resume,trigger}`; `GET /{id}/runs`, `GET /runs/{rid}`, run-file preview/edit/download; `GET /runs/{rid}/messages` - agent messages for the run's underlying agent run (used to inspect a stuck/running run; authorized via `_load_run`); `GET /runs/{rid}/html` - generated HTML document for an `html_page` run; `DELETE /runs/{rid}` - owner-scoped, with an admin fallback so admins can delete any user's run; publish/subscribe under `/{id}/{publish,unpublish,subscribe}` and `GET /published`. **Admin (cross-user):** `GET /admin/all` - every task; `PATCH/DELETE /admin/{id}` - edit/delete any task; `POST /admin/{id}/trigger`; `GET /admin/{id}/runs`. Admin routes require `system_role == "admin"`. | | **HTML Page Favorites** (`/api/html-page-favorites`) | Bookmarked generated HTML pages, owner-scoped. `GET /` - list; `POST /` `{page_id, title?, description?, tags?}` - bookmark a `scheduled_html_pages` row (extracts a reusable style/layout summary via GLM5 at favorite time); a page can only be favorited once per user (`DuplicateFavoriteError` → HTTP 409). `GET/DELETE /{id}` - read/delete (the create-task reference picker deletes favorites here). A favorite can be selected as a style reference when creating a new `html_page` task. | | **Public Scheduled Tasks** (`/api/public/scheduled-tasks`) | **Unauthenticated** mirror for embedding. Adds `GET /runs/{rid}/html` and `GET /{task_id}/latest-html` for the `html_page` preview; the run result keeps `result.html` (only file paths are redacted). | -| **Knowledge** (`/api/knowledge`) | Product knowledge base, **global/shared** (no user isolation; writes record `created_by`/`updated_by` for audit). `POST /ingest/thread/{thread_id}` - sediment one thread into a note (reads messages from the checkpointer, `threads:read` owner-checked); `POST /ingest/threads/batch` - batch-sediment the caller's recent threads (`skip_existing` via `.manifest.json` dedup). **Optional model selection**: both ingest endpoints accept `model_name` (overrides `knowledge.extract_model_name` for that capture; threaded through `KnowledgeService.capture_thread` → `extract_thread_llm_multi`). **Model-error surfacing**: when the chosen model produces nothing usable (timeout / API error / non-JSON) capture still **falls back to rule-based extraction and lands a draft** (`llm_fallback:true`); the single-thread (non-background) response and the batch response (`llm_fallback` count) carry this so the UI tells the user "已改用规则抽取". The "保存到知识库" chat button stays background/fire-and-forget (passes `model_name`, no per-save error toast); the batch-import dialog is synchronous and shows the fallback count. `POST /ingest/file` - **import an uploaded file** (multipart: `file` + optional `title`/`tags`/`status`/`model_name`) into wiki note(s): the file is converted to Markdown (md/txt read directly, pdf/word/ppt/excel via `convert_file_to_markdown`), distilled by the same wiki-capture LLM pipeline as thread capture (`source_type="file"`), and falls back to a verbatim **draft** note when the model returns nothing usable (`llm_fallback:true`); `POST /ingest/search` - sediment search/tool results as a reference note; `POST /notes` - manual note; `GET /notes` (keyword/tag/source_type/status + pagination); `GET/PUT/DELETE /notes/{id}` (DELETE is soft = archive, `?hard=true` removes the vault file); `POST /search` - `mode=keyword\|vector\|hybrid` (vector→keyword fallback when no embedding endpoint); `GET /notes/{id}/versions` - edit history; `GET /duplicates` + `POST /merge` - dedup; `POST /reindex` - rebuild vector index; `GET /graph` - product graph data (notes + entities, graph-blacklist filtered); `GET/POST /graph/blacklist` + `DELETE /graph/blacklist/{id}` - graph node blacklist (labels = entity names / note titles, case-insensitive match; blacklisted nodes and their edges are dropped from `/graph` and the wiki-export artifacts; table `knowledge_graph_blacklist`, auto-created); `POST /export/graph` + `GET /export/files/{graph.json\|graph.html\|graph.graphml\|cypher.txt}` - obsidian-wiki `wiki-export` artifacts (self-contained `graph.html` for iframe); `POST /resolve` `{targets:[…]}` + `GET /notes/{id}/wikilinks` - resolve `[[wikilink]]` targets to a note id / vault page / missing (powers clickable wikilinks); `GET /vault-page?path=` - read a raw vault meta/hub Markdown page (no DB note). Backed by the `deerflow.knowledge` module (see below). | +| **Knowledge** (`/api/knowledge`) | Product knowledge base, **global/shared** (no user isolation; writes record `created_by`/`updated_by` for audit). `POST /ingest/thread/{thread_id}` - sediment one thread into a note (reads messages from the checkpointer, `threads:read` owner-checked); `POST /ingest/threads/batch` - batch-sediment the caller's recent threads (`skip_existing` via `.manifest.json` dedup). **Optional model selection**: both ingest endpoints accept `model_name` (overrides `knowledge.extract_model_name` for that capture; threaded through `KnowledgeService.capture_thread` → `extract_thread_llm_multi`). **Model-error surfacing**: when the chosen model produces nothing usable (timeout / API error / non-JSON) capture still **falls back to rule-based extraction and lands a draft** (`llm_fallback:true`); the single-thread (non-background) response and the batch response (`llm_fallback` count) carry this so the UI tells the user "已改用规则抽取". The "保存到知识库" chat button stays background/fire-and-forget (passes `model_name`, no per-save error toast); the batch-import dialog is synchronous and shows the fallback count. `POST /ingest/file` - **import an uploaded file** (multipart: `file` + optional `title`/`tags`/`status`/`model_name`) into wiki note(s): the file is converted to Markdown (md/txt read directly, pdf/word/ppt/excel via `convert_file_to_markdown`), distilled by the same wiki-capture LLM pipeline as thread capture (`source_type="file"`), and falls back to a verbatim **draft** note when the model returns nothing usable (`llm_fallback:true`); `POST /ingest/search` - sediment search/tool results as a reference note; `POST /notes` - manual note; `GET /notes` (keyword/tag/entity/source_type/status + pagination; `entity` is an exact entity-name filter resolved via the note→entity relation graph, combinable with `keyword` — powers the related-notes dialog's server-side search + paging); `GET/PUT/DELETE /notes/{id}` (DELETE is soft = archive, `?hard=true` removes the vault file); `POST /search` - `mode=keyword\|vector\|hybrid` (vector→keyword fallback when no embedding endpoint); `GET /notes/{id}/versions` - edit history; `GET /duplicates` + `POST /merge` - dedup; `POST /reindex` - rebuild vector index; `GET /graph` - product graph data (notes + entities, graph-blacklist filtered); `GET/POST /graph/blacklist` + `DELETE /graph/blacklist/{id}` - graph node blacklist (labels = entity names / note titles, case-insensitive match; blacklisted nodes and their edges are dropped from `/graph` and the wiki-export artifacts; table `knowledge_graph_blacklist`, auto-created); `POST /export/graph` + `GET /export/files/{graph.json\|graph.html\|graph.graphml\|cypher.txt}` - obsidian-wiki `wiki-export` artifacts (self-contained `graph.html` for iframe); `POST /resolve` `{targets:[…]}` + `GET /notes/{id}/wikilinks` - resolve `[[wikilink]]` targets to a note id / vault page / missing (powers clickable wikilinks); `GET /vault-page?path=` - read a raw vault meta/hub Markdown page (no DB note). Backed by the `deerflow.knowledge` module (see below). | Proxied through nginx: `/api/langgraph/*` → LangGraph, all other `/api/*` → Gateway. @@ -406,6 +406,7 @@ Operators upgrading an existing DB run the whole file (`USE ;` then execute) - **Format**: Directory with `SKILL.md` (YAML frontmatter: name, description, license, allowed-tools) - **Loading**: `load_skills()` recursively scans `skills/{public,custom}` for `SKILL.md`, parses metadata, and reads enabled state from extensions_config.json - **Injection**: Enabled skills listed in agent system prompt with container paths +- **Per-agent whitelist (custom agents)**: a custom agent's `config.yaml` `skills:` field is an explicit allowlist — `None` (key omitted) and `[]` both mean **no skills**, never "all skills". Only the default lead agent (no `agent_id`) and the bootstrap flow see every enabled skill. Enforced consistently in three places: the system prompt (`get_skills_prompt_section`, empty whitelist → no `` block at all), the `skill_list` tool, and the `skill_view` tool (out-of-whitelist names answer "not found" so existence isn't leaked) — including the `skill_manage` `list`/`view` aliases, which delegate to the shared `skill_list_impl`/`skill_view_impl` with the active `agent_id`. Tests: `tests/test_skill_tools_whitelist.py` - **Installation**: `POST /api/skills/install` extracts .skill ZIP archive to custom/ directory - **Chinese name (`name_zh`)**: a Chinese display name stored in the **`skills` DB table** (`SkillRow.name_zh`, `String(128)`, migration `20260609_01`) — never written to the on-disk SKILL.md, so it works for built-in skills too. Threaded through `SkillStore` create/ensure_legacy/update/update_admin exactly like `detail`; surfaced on `SkillResponse.name_zh` from the ownership record in `_skill_to_response`. Authorization mirrors `detail`: custom skill → owner or admin; built-in/public → admin only (row materialized on demand via `ensure_legacy`). Endpoints (in `app/gateway/routers/skills.py`): `PUT /api/skills/{name}/name-zh` (set/clear), `POST /api/skills/{name}/name-zh/generate` (LLM generates one), `POST /api/skills/name-zh/generate-missing` (一键补充: **admin-only** bulk-generate for every skill that lacks one, bounded-concurrency `asyncio.gather`). **Optional model selection**: both generate endpoints accept an optional JSON body `{model_name}` — an explicit name wins, else the curator model setting, else the main model (`_resolve_skill_name_zh_model_name`). **Model-error surfacing**: `_generate_skill_name_zh` raises `SkillNameZhModelError` when the model itself fails (build/timeout/API error) — distinct from the model returning an unusable name (→ `None`). The single endpoint maps that to `502 大模型报错,无法补全技能中文名:…`; the bulk endpoint fails fast on a model-build error (502) and otherwise carries a representative `model_error` string in `SkillNameZhBulkResponse` so the UI can say "N 个因大模型报错失败". The frontend (`skill-settings-page.tsx` list + `skill-detail-dialog.tsx` editor, plus the agent skill pickers/badges in `NewAgentPage.tsx` + `agent-card.tsx`) shows `name_zh || name`, offers per-skill "AI 生成" plus an **admin-only** header "一键补充中文名" button, and a compact model dropdown (`components/workspace/model-name-select.tsx`, default = configured model) beside each action. diff --git a/offline-backend-20260512/backend/app/gateway/routers/knowledge.py b/offline-backend-20260512/backend/app/gateway/routers/knowledge.py index 2c92ffe1..8562dde9 100644 --- a/offline-backend-20260512/backend/app/gateway/routers/knowledge.py +++ b/offline-backend-20260512/backend/app/gateway/routers/knowledge.py @@ -424,6 +424,7 @@ async def list_notes( request: Request, keyword: str | None = None, tag: str | None = None, + entity: str | None = None, source_type: str | None = None, status: str | None = None, limit: int = 20, @@ -432,7 +433,7 @@ async def list_notes( service = get_knowledge_service(request) limit = max(1, min(limit, 100)) offset = max(0, offset) - return await service.list_notes(keyword=keyword, tag=tag, source_type=source_type, status=status, limit=limit, offset=offset) + return await service.list_notes(keyword=keyword, tag=tag, entity=entity, source_type=source_type, status=status, limit=limit, offset=offset) @router.get("/notes/{note_id}") diff --git a/offline-backend-20260512/backend/packages/harness/deerflow/knowledge/repository.py b/offline-backend-20260512/backend/packages/harness/deerflow/knowledge/repository.py index 436489b4..2cbe8f27 100644 --- a/offline-backend-20260512/backend/packages/harness/deerflow/knowledge/repository.py +++ b/offline-backend-20260512/backend/packages/harness/deerflow/knowledge/repository.py @@ -276,6 +276,7 @@ class KnowledgeRepository: *, keyword: str | None = None, tag: str | None = None, + entity: str | None = None, source_type: str | None = None, status: str | None = None, include_archived: bool = False, @@ -303,6 +304,16 @@ class KnowledgeRepository: .where(func.lower(KnowledgeTagRow.name) == tag.lower()) ) conditions.append(KnowledgeNoteRow.id.in_(tag_subq)) + if entity: + # Exact entity match via the note→entity relation graph — precise + # and index-backed, unlike a content LIKE scan on large vaults. + entity_subq = ( + select(KnowledgeRelationRow.from_note_id) + .join(KnowledgeEntityRow, KnowledgeRelationRow.to_entity_id == KnowledgeEntityRow.id) + .where(func.lower(KnowledgeEntityRow.name) == entity.lower()) + .where(KnowledgeRelationRow.from_note_id.is_not(None)) + ) + conditions.append(KnowledgeNoteRow.id.in_(entity_subq)) for cond in conditions: stmt = stmt.where(cond) diff --git a/offline-backend-20260512/backend/packages/harness/deerflow/knowledge/service.py b/offline-backend-20260512/backend/packages/harness/deerflow/knowledge/service.py index 38e92dad..b83ca128 100644 --- a/offline-backend-20260512/backend/packages/harness/deerflow/knowledge/service.py +++ b/offline-backend-20260512/backend/packages/harness/deerflow/knowledge/service.py @@ -523,6 +523,7 @@ class KnowledgeService: *, keyword: str | None = None, tag: str | None = None, + entity: str | None = None, source_type: str | None = None, status: str | None = None, limit: int = 20, @@ -531,6 +532,7 @@ class KnowledgeService: items, total = await self.repo.list_notes( keyword=keyword, tag=tag, + entity=entity, source_type=normalize_source_type(source_type) if source_type else None, status=status, limit=limit, diff --git a/offline-backend-20260512/backend/packages/harness/deerflow/tools/builtins/skill_tools.py b/offline-backend-20260512/backend/packages/harness/deerflow/tools/builtins/skill_tools.py index d2722103..50fae12a 100644 --- a/offline-backend-20260512/backend/packages/harness/deerflow/tools/builtins/skill_tools.py +++ b/offline-backend-20260512/backend/packages/harness/deerflow/tools/builtins/skill_tools.py @@ -39,10 +39,15 @@ def _load_agent_skill_whitelist(agent_id: str | None) -> set[str] | None: """Return the explicit skill name set declared in an agent's config.yaml. Semantics: - - ``None`` returned → no agent_id, or config missing — caller should - fall back to "all skills visible". - - ``set()`` (empty) → agent has ``skills: []`` declared — visibility - is intentionally empty. + - ``None`` returned → no agent_id, or the agent's on-disk config is + missing entirely — caller should fall back to "all skills visible" + (default lead agent behavior). + - ``set()`` (empty) → custom agent with ``skills: []`` declared **or** + with no ``skills`` key at all — visibility is intentionally empty. + This mirrors the prompt layer (``agents_config.AgentConfig.skills``: + ``None``/``[]`` both mean "no skills" for custom agents), so an agent + whose config omits ``skills`` can no longer enumerate every skill + through this tool. - Non-empty set → only these names are visible to this agent. """ if not agent_id: @@ -51,9 +56,9 @@ def _load_agent_skill_whitelist(agent_id: str | None) -> set[str] | None: from deerflow.config.agents_config import load_agent_config cfg = load_agent_config(agent_id) - if cfg is None or cfg.skills is None: + if cfg is None: return None - return set(cfg.skills) + return set(cfg.skills or []) except Exception: logger.debug("Failed to load agent config for skill_list filter", exc_info=True) return None @@ -119,16 +124,21 @@ def _load_owned_skill_names(user_id: str | None) -> set[str] | None: return None -@tool("skill_view") -def skill_view_tool(name: str) -> str: - """Read the full SKILL.md content of a skill by name. +def skill_view_impl(name: str, agent_id: str | None) -> str: + """Read a skill's SKILL.md, honoring the agent's config.yaml whitelist. - Args: - name: The skill name (hyphen-case, e.g. 'my-skill'). + Custom agents may only read skills on their whitelist — same filter as + ``skill_list``, so a restricted agent cannot bypass its allowlist by + guessing skill names. Replies "not found" rather than "forbidden" to + avoid leaking which skills exist. """ from deerflow.skills.storage import get_or_new_skill_storage from deerflow.skills.usage import bump_view + allowed = _load_agent_skill_whitelist(agent_id) + if allowed is not None and name not in allowed: + return f"Skill '{name}' not found." + storage = get_or_new_skill_storage() try: if storage.custom_skill_exists(name): @@ -156,6 +166,16 @@ def skill_view_tool(name: str) -> str: return content +@tool("skill_view") +def skill_view_tool(name: str, runtime: ToolRuntime) -> str: + """Read the full SKILL.md content of a skill by name. + + Args: + name: The skill name (hyphen-case, e.g. 'my-skill'). + """ + return skill_view_impl(name, _resolve_active_agent_id(runtime)) + + @tool("skill_list") def skill_list_tool(runtime: ToolRuntime) -> str: """List skills available to this agent, with metadata and usage stats. @@ -175,11 +195,15 @@ def skill_list_tool(runtime: ToolRuntime) -> str: state, use_count, view_count, patch_count, last_activity_at, pinned, source. """ + return skill_list_impl(_resolve_active_agent_id(runtime)) + + +def skill_list_impl(agent_id: str | None) -> str: + """List skills visible to ``agent_id`` (None = default lead agent).""" from deerflow.skills.storage import get_or_new_skill_storage from deerflow.skills.types import SkillCategory from deerflow.skills.usage import STATE_ARCHIVED, all_entries, latest_activity_at - agent_id = _resolve_active_agent_id(runtime) allowed = _load_agent_skill_whitelist(agent_id) # 默认 lead agent 没有 per-agent 白名单 —— 改为只暴露"未归档"且"内置或属于 diff --git a/offline-backend-20260512/backend/packages/harness/deerflow/tools/skill_manage_tool.py b/offline-backend-20260512/backend/packages/harness/deerflow/tools/skill_manage_tool.py index 04b6bda6..ecee81a6 100644 --- a/offline-backend-20260512/backend/packages/harness/deerflow/tools/skill_manage_tool.py +++ b/offline-backend-20260512/backend/packages/harness/deerflow/tools/skill_manage_tool.py @@ -218,14 +218,16 @@ async def _skill_manage_impl( """ # Read-only aliases — delegate to the dedicated tools so a model that # over-uses skill_manage doesn't trip on Unsupported action errors. + # Pass the active agent_id through so a custom agent's skill whitelist + # is enforced on these aliases just like on skill_list/skill_view. if action == "list": - from deerflow.tools.builtins.skill_tools import skill_list_tool + from deerflow.tools.builtins.skill_tools import _resolve_active_agent_id, skill_list_impl - return await _to_thread(skill_list_tool.invoke, {}) + return await _to_thread(skill_list_impl, _resolve_active_agent_id(runtime)) if action == "view": - from deerflow.tools.builtins.skill_tools import skill_view_tool + from deerflow.tools.builtins.skill_tools import _resolve_active_agent_id, skill_view_impl - return await _to_thread(skill_view_tool.invoke, {"name": name}) + return await _to_thread(skill_view_impl, name, _resolve_active_agent_id(runtime)) name = SkillStorage.validate_skill_name(name) lock = _get_lock(name) diff --git a/qa-client/qa_client.py b/qa-client/qa_client.py new file mode 100644 index 00000000..03a80656 --- /dev/null +++ b/qa-client/qa_client.py @@ -0,0 +1,135 @@ +# -*- coding: utf-8 -*- +"""DeerFlow 后端问答客户端（零第三方依赖，纯 Python 标准库）。 + +用法：只需修改下面的 BASE_URL（必要时改 USERNAME / LOGIN_PASSWORD）， +然后调用 ask("你的问题") 即可拿到回答文本。 + + from qa_client import ask + + answer = ask("你好，介绍一下你自己") + print(answer) + +多轮对话（共用同一个 thread_id 即可保留上下文）： + + tid = new_thread_id() + print(ask("我叫小明", thread_id=tid)) + print(ask("我叫什么名字？", thread_id=tid)) +""" + +from __future__ import annotations + +import json +import ssl +import urllib.error +import urllib.parse +import urllib.request +import uuid + +# ======================= 只需要修改这里 ======================= +BASE_URL = "http://47.94.209.59:2026" # 后端 Gateway 地址（改成你的服务器地址即可） +USERNAME = "api-client" # 登录用户名（免密登录，不存在会自动注册） +LOGIN_PASSWORD = "123ewq" # 仅当后端开启 require_password 口令门槛时才需要填（默认口令 123ewq） +MODEL_NAME = None # 可选：指定模型名；None = 使用后端默认模型 +TIMEOUT_SECONDS = 600 # 等待回答的超时时间（复杂问题可能跑几分钟） +# ============================================================= + +_token_cache = {"token": None} +_ssl_ctx = ssl.create_default_context() +_ssl_ctx.check_hostname = False +_ssl_ctx.verify_mode = ssl.CERT_NONE # 内网自签证书也能用 + + +def _post_json(url: str, payload: dict, token: str | None = None, timeout: int = 30) -> dict: + """POST JSON 并解析 JSON 响应。""" + data = json.dumps(payload).encode("utf-8") + req = urllib.request.Request(url, data=data, method="POST") + req.add_header("Content-Type", "application/json") + if token: + req.add_header("Authorization", f"Bearer {token}") + with urllib.request.urlopen(req, timeout=timeout, context=_ssl_ctx) as resp: + return json.loads(resp.read().decode("utf-8")) + + +def _login() -> str | None: + """免密用户名登录，返回 Bearer token；后端关闭鉴权时返回 None 也能正常问答。""" + if _token_cache["token"]: + return _token_cache["token"] + url = f"{BASE_URL.rstrip('/')}/api/v1/auth/login/username" + if LOGIN_PASSWORD: + url += "?" + urllib.parse.urlencode({"password": LOGIN_PASSWORD}) + try: + result = _post_json(url, {"username": USERNAME}) + _token_cache["token"] = result.get("access_token") + return _token_cache["token"] + except urllib.error.HTTPError as e: + if e.code == 401: + raise RuntimeError("登录被拒绝：后端开启了登录口令，请在 LOGIN_PASSWORD 填入正确口令（默认 123ewq）") from e + # 登录接口异常时尝试无 token 直连（后端可能处于免鉴权模式） + return None + except urllib.error.URLError: + return None + + +def _extract_answer(channel_values: dict) -> str: + """从 /api/runs/wait 返回的最终状态里取出最后一条 AI 回复文本。""" + messages = channel_values.get("messages") or [] + for msg in reversed(messages): + if not isinstance(msg, dict): + continue + if msg.get("type") not in ("ai", "AIMessageChunk", "AIMessage"): + continue + content = msg.get("content") + if isinstance(content, str) and content.strip(): + return content.strip() + if isinstance(content, list): # 结构化内容块：拼接其中的文本 + parts = [b.get("text", "") for b in content if isinstance(b, dict) and b.get("type") == "text"] + text = "".join(parts).strip() + if text: + return text + # 没有 AI 消息时，把原始返回吐出来便于排查 + return json.dumps(channel_values, ensure_ascii=False)[:2000] + + +def new_thread_id() -> str: + """生成一个新会话 id，用于多轮对话。""" + return str(uuid.uuid4()) + + +def ask(question: str, thread_id: str | None = None) -> str: + """向后端发送一个问题，阻塞直到拿到最终回答文本。 + + :param question: 要问的问题 + :param thread_id: 可选。传同一个 id 即为多轮对话（保留上下文）；不传则每次新会话 + :return: AI 的回答文本 + """ + token = _login() + configurable: dict = {} + if thread_id: + configurable["thread_id"] = thread_id + if MODEL_NAME: + configurable["model_name"] = MODEL_NAME + + payload = { + "input": {"messages": [{"role": "user", "content": question}]}, + "config": {"configurable": configurable}, + } + url = f"{BASE_URL.rstrip('/')}/api/runs/wait" + try: + result = _post_json(url, payload, token=token, timeout=TIMEOUT_SECONDS) + except urllib.error.HTTPError as e: + body = e.read().decode("utf-8", errors="replace")[:500] + raise RuntimeError(f"请求失败 HTTP {e.code}: {body}") from e + except urllib.error.URLError as e: + raise RuntimeError(f"无法连接后端 {BASE_URL}，请检查 BASE_URL 和服务是否启动: {e.reason}") from e + + return _extract_answer(result) + + +if __name__ == "__main__": + # 示例 1：单轮问答 + print(ask("你好，请简单介绍一下你自己")) + + # 示例 2：多轮对话（取消注释即可） + # tid = new_thread_id() + # print(ask("我叫小明，记住了吗？", thread_id=tid)) + # print(ask("我叫什么名字？", thread_id=tid))