代码提交
This commit is contained in:
1 parent
007ed5a7de
commit
babc5e95c6
21 files changed
+1170
-145
No files matched your search
@@ -224,6 +224,8 @@ from deerflow.config import get_app_config
|
||||
- Tools loaded via `get_available_tools()` - combines sandbox, built-in, MCP, community, and subagent tools
|
||||
- System prompt generated by `apply_prompt_template()` with skills, memory, and subagent instructions
|
||||
|
||||
**Forced-research built-in agent.** `forced-research-responder` (「强制检索输出助手」) is seeded from `app/gateway/routers/_forced_research_seed_assets/` before `_sync_legacy_agents`. `ForcedResearchMiddleware` is mounted only for this stable id. On every visible user turn it requires a successful information-gathering tool result before the model may finish; `tool_choice=required` is backed by an `after_model -> model` retry for gateways that ignore tool choice, with an eight-attempt failure cap that permits only an honest failure response. Skill discovery, reading `SKILL.md`, and writing a helper script do not satisfy the gate; web/knowledge/MCP results, non-skill uploaded-file reads, or successful Python skill-script execution do. Tool filtering removes `present_files`, `skill_manage`, task/schedule/memory/clarification surfaces. The middleware permits `write_file`/`str_replace` only for `/mnt/user-data/workspace/*.py` and permits `bash` only for `.py` execution without redirection/document paths, so skills can run scripts but the agent cannot create Markdown/document artifacts or write `outputs`. Tests: `tests/test_forced_research_agent.py`.
|
||||
|
||||
**ThreadState** (`packages/harness/deerflow/agents/thread_state.py`):
|
||||
- Extends `AgentState` with: `sandbox`, `thread_data`, `title`, `artifacts`, `todos`, `uploaded_files`, `viewed_images`
|
||||
- Uses custom reducers: `merge_artifacts` (deduplicate), `merge_viewed_images` (merge/clear)
|
||||
@@ -892,7 +894,7 @@ scheduler runs once per deployment. Per run:
|
||||
|
||||
- **可检测的模型失败**:`LLMErrorHandlingMiddleware` 在同一模型重试 3 次后**不抛异常**,而是返回兜底 `AIMessage` —— 现在给它打 `additional_kwargs[LLM_ERROR_MARKER="deerflow_llm_error"]=reason`(model_dump 保留,经 checkpointer 序列化仍在)。`is_llm_error_message(msg)` 据标记(回退按已知兜底文案正文)识别,返回 reason。
|
||||
- **候选模型链**:`app/gateway/roundtable_model_fallback.py::fallback_model_chain(current)` 按 `config.yaml models[]` 顺序给出 `[当前模型, …其余]`(去重保序)。`reason_text(reason)` 翻译成简短中文。总开关 env `ROUNDTABLE_MODEL_FALLBACK`(默认开,`0/false/off` 关 → 单模型旧行为);`ROUNDTABLE_MODEL_FALLBACK_MAX`(0=试完全部)。
|
||||
- **触发换模型的三种情形**:① 单轮 run 抛异常(start_run 失败 / idle 卡死 → `timeout`,其余 `exception`);② 兜底错误文案(`is_llm_error_message`);③ 交付为空(仅 seat/leader;leader 还含「既无正文又无派活」)。Step3 单例(report/summary/dashboard/ingest/结构化)的产物是落盘文件,**不**把「正文为空」当失败,只对真模型错误换模型。
|
||||
- **触发换模型的情形**:① 单轮 run 抛异常(start_run 失败 / idle 卡死 → `timeout`,其余 `exception`);② 兜底错误文案(`is_llm_error_message`);③ 交付无效(仅 seat/leader;leader 还含「既无正文又无派活」)。席位**只认剥掉 `<think>` 后的可见正文**——思考-only / 空串 / 断流都不算已交付(`deerflow.agents.roundtable_orchestrator.delivery`:`visible_seat_text` / `classify_seat_delivery`)。前台 `_special_run` 对此 yield `status:"invalid_delivery"` 且不广播;`GET /last-ai-message` 回 `{delivered, content, length, invalid_reason}`(思考-only 标 `thinking_only`,总控席位清单标 ⚠️)。后台引擎只把可见正文写入 `deliveries`,续轮用 `build_continue_prompt`:有无效席位时强制重派,禁止声称「已完成交付」。Step3 单例(report/summary/dashboard/ingest/结构化)的产物是落盘文件,**不**把「正文为空」当失败,只对真模型错误换模型。
|
||||
- **后台**:`InProcessRoundtableGateway._run_turn_with_fallback(...)` 包住 `_run_agent_turn`,逐候选模型试;换模型落 `model_switch` 诊断、全失败落 `model_fallback_exhausted`。`run_leader/run_seat/run_report/run_summary/run_dashboard_data` 全走它(report/summary/dashboard 的产物自愈重试保留,并记住已切到的可用模型用于后续续写轮)。各 Turn/Result 带回 `switches: list[ModelSwitch]`,引擎 `_emit_switches` → `JobProgressWriter.model_switch(...)` 往研讨时间线插一条 `role="system"` 系统提示气泡。
|
||||
- **前台**:`_special_run`/`_leader_run` stream 收尾后检测失败 → yield `{status:"model_switch", role, agent_name, failed_model, next_model, reason}` 帧并用下一个模型重新流式;终态帧带 `model_used`。结构化抽取 `flow_extract.py` 的 round1 `model.astream` 也包了同款换模型循环(emit `event:"model_switch"` 帧)。
|
||||
- **Step2 单席位隔离**:`engine._run_recommend/_run_chain/_run_dag` 的每个席位执行包 try/except —— 一个席位彻底失败(落 `seat_failed` 诊断 + 置空交付)不拖垮整批/整链/整个作业。
|
||||
|
||||
Reference in new issue
Block a user