merge: candidate subagent-process-view-impl-2026-09-28 into dev
# Conflicts: # .doc/ops/CURRENT.md
This commit is contained in:
weichen
@@ -0,0 +1,54 @@
|
||||
# 任务：小队过程流式展示与 subagent 子页面设计
|
||||
|
||||
## 状态
|
||||
|
||||
- status: done（2026-09-28）
|
||||
- owner: Cursor
|
||||
- type: 方案研究（设计文档，无代码实施）
|
||||
- created: 2026-09-28
|
||||
- base_dev_sha: 4c42d1bf74e262930bcf127e517ab6b76a0dda71
|
||||
- branch: feature/subagent-process-view-design-2026-09-28
|
||||
- worktree: .worktrees/subagent-process-view-design-2026-09-28
|
||||
- write_scope: `.doc/system-design/`、`.doc/ops/short-term/`（本文件）
|
||||
- mutex: subagent-process-view-design
|
||||
|
||||
## 用户目标与授权
|
||||
|
||||
> 是否能让助手的工作过程，像豆包、kimi、codex 等工具一样，输出思考、工具调用、过程内容输出等信息；使用流式输出，持续更新助手的工作过程。多智能体小队先设计如何使用 subagent 子页面来展示上面那样的信息，但可以先不实现。
|
||||
|
||||
（来源：2026-09-28 Cursor 会话，用户已确认计划）
|
||||
|
||||
- 授权范围：撰写系统设计文档；**不实施任何代码变更**（用户明示"可以先不实现"）
|
||||
- 交付终点：设计文档评审通过（可审产物验收）
|
||||
- 禁止项：修改 Go/TS 代码、数据库迁移、环境操作、推送远端
|
||||
|
||||
## 需求重叠探测
|
||||
|
||||
- 2026-09-28 `catalog search`（关键词：流式输出/思考/工具调用/subagent 页面、streaming agent thought tool call、小队过程可视化流式）均无命中
|
||||
- 既有关联文档为归档/演进方向（2026-09-08 三层信息架构、小队编排 §18.5），无进行中任务重叠
|
||||
|
||||
## 研究输入
|
||||
|
||||
- 三路并行只读探索（后端流式链路 / squad+taskrun 架构 / 前端聊天架构）+ 主线直接调研，交叉验证
|
||||
- 关键现状结论：协议/持久化/前端骨架已就绪，七处断点导致过程数据（思考链、工具调用、增量输出）无法到达前端；前端无 subagent 页面载体
|
||||
|
||||
## 交付物
|
||||
|
||||
- `.doc/system-design/2026-09-28-agent-process-streaming-and-subagent-view-design.md`（460 行，A–H 八章：目标参照/协议扩展/七断点打通/subagent 子页面/分期路线/风险边界/断点对照表/参考文件）
|
||||
- 已登记 `.doc/system-design/README.md` 报告清单（0c 条目）
|
||||
|
||||
## 验证结果（2026-09-28 实测）
|
||||
|
||||
- 行号核对：设计文档引用的 14 处 文件:行 全部与真实代码一致（sed 逐处核对：stream/types.go:68、pinative.go:102、manager.go:364、loop.go:178-183、backend.go:38、runner.mjs:600/24、sse.go:476、workflow.go:543、useTaskRunAttempts.ts:474、taskRuns.ts:113、session.go:624、phase_snapshot.go:225 等）
|
||||
- 内链核对：文档内 21 个代码文件相对链接、6 个 .doc 文档引用全部存在
|
||||
- 术语修正：stripThinkBlocks 剥离对象修正为 `<think>…</think>`（原文误写为特殊符号）
|
||||
- 无硬编码模型名（grep MiniMax/M2.7 零命中于方案正文，仅协议描述提及 provider 家族作为 reasoning_content 语义来源）
|
||||
- 产品 PI 测试 NOT REQUIRED（依据：验证契约「范围判定」——纯设计文档，无运行模板/提示词/业务变更；本任务不产生产品代码）
|
||||
|
||||
## 五态收口
|
||||
|
||||
- review_state: done（设计文档自检完成；待用户评审确认）
|
||||
- integration_state: not_applicable（不合入代码；文档随本任务分支提交）
|
||||
- deployment: not_applicable
|
||||
- cleanup: pending（worktree 待独立维护者复查后清理）
|
||||
- remote_sync: not_applicable（未授权 push）
|
||||
@@ -0,0 +1,460 @@
|
||||
# 智能体工作过程流式展示与 subagent 子页面设计
|
||||
|
||||
> 版本：v1.0（设计稿，未实施）
|
||||
> 日期：2026-09-28
|
||||
> 任务：subagent-process-view-design-2026-09-28
|
||||
> 作者：Cursor（owner），三路并行代码探索 + 主线调研交叉验证
|
||||
> 交付物性质：系统设计文档，**不含任何代码实施**（用户明示"可以先不实现"）
|
||||
|
||||
---
|
||||
|
||||
## A. 目标与参照
|
||||
|
||||
### A.1 用户需求
|
||||
|
||||
1. 让助手的工作过程像豆包、kimi、codex 等工具一样对外输出：**思考（thinking）、工具调用（tool call）、过程内容输出**三类信息；
|
||||
2. 使用**流式输出**持续更新助手的工作过程，而非终态一次性呈现；
|
||||
3. 多智能体小队设计 **subagent 子页面**承载上述展示（先设计，不实现）。
|
||||
|
||||
### A.2 参照产品的展示形态拆解
|
||||
|
||||
| 产品 | 思考展示 | 工具调用展示 | 增量输出 | 交互模式 |
|
||||
|---|---|---|---|---|
|
||||
| 豆包 | "深度思考"折叠块：思考中呼吸态滚动，完成后折叠为一行摘要，可点击展开全文 | 工具卡片：图标+工具名+入参摘要+结果摘要+耗时 | 正文逐 token 流式 | 主流内嵌折叠，不打断阅读 |
|
||||
| kimi | 推理过程折叠区，默认收起，进行中有进度感 | 搜索/工具步骤条 | 逐段流式 | 同上 |
|
||||
| codex（CLI/Web） | 思考行内灰字流式 | 命令执行块（命令+输出+退出码） | 流式 | 工程向，密度更高 |
|
||||
|
||||
共同设计语言：**默认折叠（低认知负荷）→ 状态可感知（进行中动效）→ 按需展开（全文/详情）→ 归档后可回放**。
|
||||
|
||||
### A.3 本仓已有基座与约束
|
||||
|
||||
本设计不是从零发明，而是延伸既有三层信息架构（2026-09-08 streaming-events 设计，已归档但语义仍有效）：
|
||||
|
||||
- **L1 phase/status**（"现在在干嘛"）：已实现并持久化（`task_run_events`）；
|
||||
- **L2 summary/delta/briefing**（"结论是什么"）：已实现（delta 为整段伪流式）；
|
||||
- **L3 detail**（"推演依据"）：**协议已定义（`DetailPayload.category: agent_thought/tool_call/step_output/log`），但无生产者、双重丢弃**——本设计的主战场。
|
||||
|
||||
既有 UX 约束必须继承：
|
||||
|
||||
1. **不打扰原则**：detail 默认折叠，不默认展开思考链/工具日志（2026-09-09 chat-wait-error-retry 设计 P1 明确）；
|
||||
2. **透明可追溯**：详细思考与草稿不丢失，按需可查（工作区文件留存兜底）；
|
||||
3. **治理红线**：过程展示只是可视化，**不改变验收证据边界**——PI 叙事、模型输出不能冒充 PASS/测试证据（AGENTS.md 安全与证据）。
|
||||
|
||||
### A.4 概念区分（重要）
|
||||
|
||||
本设计的 "subagent" 指**产品内小队成员**（`internal/squad` 的 AgentSpec：organizer/regulatory/writer/leader 等 8 标准专员 + ad_hoc 派生专员），**不是**本仓开发流程的工程工具 PI 子任务派发（`scripts/pi_subagents.py` + `.doc/ops/2026-09-24-pi-subagent-executor.md`）。两者名字相近但完全不同域，实施时不得混用数据源。
|
||||
|
||||
---
|
||||
|
||||
## B. 过程事件协议扩展
|
||||
|
||||
### B.1 现状协议回顾
|
||||
|
||||
`internal/stream/types.go` 的 `StreamEvent` 已定义 12 种事件类型；其中 `DetailPayload`：
|
||||
|
||||
```go
|
||||
// internal/stream/types.go:68
|
||||
type DetailPayload struct {
|
||||
Category string `json:"category"` // agent_thought, step_output, tool_call, log
|
||||
StepID string `json:"step_id,omitempty"` // 步骤标识 (如 tasks/01_regulatory)
|
||||
Role string `json:"role,omitempty"` // 对应角色
|
||||
Title string `json:"title"`
|
||||
Content string `json:"content"` // 完整推演正文或技术草稿
|
||||
Collapsible bool `json:"collapsible"`
|
||||
}
|
||||
```
|
||||
|
||||
现状问题：
|
||||
|
||||
1. 只有 `Content` 全文单帧，无增量模式（思考链逐 token 需要数百帧，全文帧会放大 N 倍流量）；
|
||||
2. 无 `RunID` 冗余（事件重放时按 run 已隔离，可省；但 live 合并需要前端从通道上下文取）；
|
||||
3. 并行步骤下 `Role`/`StepID` 已有，但前端折叠键没有用它们（见 B.4）。
|
||||
|
||||
### B.2 DetailPayload 扩展（增量模式）
|
||||
|
||||
```go
|
||||
// 设计草案（不实施）
|
||||
type DetailPayload struct {
|
||||
Category string `json:"category"` // agent_thought | tool_call | step_output | log（沿用四类）
|
||||
StepID string `json:"step_id,omitempty"` // 步骤标识（如 tasks/03_regulatory）
|
||||
Role string `json:"role,omitempty"` // 成员角色 key（如 regulatory）
|
||||
RunID string `json:"run_id,omitempty"` // 冗余 run 标识（重放通道按 run 已隔离，可选）
|
||||
Seq int64 `json:"seq,omitempty"` // 同一 (category, step_id, role, item_id) 流内的序号
|
||||
ItemID string `json:"item_id,omitempty"` // 聚合项 ID：一个思考块/一次工具调用/一份产出
|
||||
// 增量模式三选一：
|
||||
Delta string `json:"delta,omitempty"` // 增量帧：追加到 ItemID 项
|
||||
Content string `json:"content,omitempty"` // 终帧/快照帧：ItemID 项的完整内容（或无增量能力的发送方）
|
||||
Done bool `json:"done,omitempty"` // 终帧标记（ItemID 项结束）
|
||||
Title string `json:"title,omitempty"`
|
||||
Collapsible bool `json:"collapsible"`
|
||||
// tool_call 特化字段（Category == "tool_call" 时）：
|
||||
Tool *ToolCallInfo `json:"tool,omitempty"`
|
||||
}
|
||||
|
||||
type ToolCallInfo struct {
|
||||
Name string `json:"name"` // write_file / read_file / grep ...
|
||||
Status string `json:"status"` // calling | succeeded | failed
|
||||
Input string `json:"input,omitempty"` // 入参摘要（JSON，可能分帧）
|
||||
Output string `json:"output,omitempty"` // 结果摘要（可能分帧）
|
||||
DurationMs int64 `json:"duration_ms,omitempty"`
|
||||
}
|
||||
```
|
||||
|
||||
**增量语义**（关键设计决策）：
|
||||
|
||||
- 发送方优先逐帧 `Delta`；`Content` 用于两类场景：(a) 无流能力的发送方（如既有 `step_output` 整段产出），(b) 周期性**快照对账帧**（防止增量丢帧后内容漂移）；
|
||||
- 客户端按 `(StepID, Role, Category, ItemID)` 聚合：`Delta` 追加、`Content` 覆盖、`Done` 定稿；
|
||||
- 节流（发送方义务，见 F.1）：逐 token 事件合并为 ~100–300ms 批帧，单帧上限（如 8KB）超限分帧。
|
||||
|
||||
### B.3 llm 层事件扩展（reasoning 链）
|
||||
|
||||
`internal/agentx/llm` 现有事件只有 `text_delta/tool_call_delta/message_end/error`。OpenAI 兼容国产模型（MiniMax/DeepSeek 系）在 delta 中返回 `reasoning_content` 字段（思考链）。扩展：
|
||||
|
||||
```go
|
||||
// internal/agentx/llm/types.go 设计草案
|
||||
const (
|
||||
KindTextDelta EventKind = "text_delta"
|
||||
KindReasoningDelta EventKind = "reasoning_delta" // 新增
|
||||
KindToolCallDelta EventKind = "tool_call_delta"
|
||||
...
|
||||
)
|
||||
|
||||
type Event struct {
|
||||
Kind EventKind
|
||||
Text string
|
||||
ReasoningDelta string // 新增：reasoning_content 增量
|
||||
...
|
||||
}
|
||||
```
|
||||
|
||||
`ParseSSE`（internal/agentx/llm/sse.go）补充解析 `delta.reasoning_content` → `KindReasoningDelta`。
|
||||
|
||||
**容器侧对应**：`deploy/pi-agent/runner.mjs:24` 的 `stripThinkBlocks` 目前把 `<think>…</think>` 块从产物中剥离（丢弃）。设计上不改变"正文不含 think"的既有约定，而是在**剥离前**把 think 块作为 reasoning 事件分流转发（见 C.5）。
|
||||
|
||||
### B.4 并行步骤归属路由（修复既有缺陷）
|
||||
|
||||
现状 `PhasePayload` 折叠键为 `phase_id#step_index`（[taskrun/phase_snapshot.go:225](../../../internal/taskrun/phase_snapshot.go) 与 [web/src/engine/phaseEventMapper.ts:65](../../../web/src/engine/phaseEventMapper.ts)），**并行组成员**（多个 role 同时 `step_executing`）若无 step_index 会互相覆盖/误折叠。
|
||||
|
||||
设计：折叠键升级为 `phase_id#step_index@role`（role 存在时必选参与键）。前端 `applyPhaseEvent` 与后端 `FoldPhaseEvents` 同步修改，语义保持对齐（两处折叠逻辑本就要求一致，见 phase_snapshot.go 注释）。detail 事件天然带 `role/step_id`，聚合不受影响。
|
||||
|
||||
### B.5 通道选择：沿用 stream/StreamEvent，不新增域事件
|
||||
|
||||
理由：
|
||||
|
||||
1. `events.StaticContracts`（internal/events/registry.go）是静态 allowlist（13 种业务事件 + YAML parity 测试），加 agent 过程事件需改 Go+YAML+测试三处，且域事件总线面向**业务投影**（通知/阶段推进），不面向用户实时流；
|
||||
2. `task_run_events`（JSONB payload）+ `after_seq` 重放已为过程事件准备好持久化与断线恢复；
|
||||
3. chat SSE 通道（链路 A）无持久化，但 B 轨参谋对话的过程展示可接受"live-only + 终态落库"（chat_turns 记录终态），与现状一致。
|
||||
|
||||
---
|
||||
|
||||
## C. 后端链路打通设计（七断点逐条）
|
||||
|
||||
### C.0 总览
|
||||
|
||||
```
|
||||
【链路 B：长任务（本设计主通道，先做）】
|
||||
worker goroutine (taskrun.Manager.runWorker)
|
||||
└ bizdev.ExecuteTask → squad.WorkflowDriver.runOneStep
|
||||
└ agent.Backend.Execute(OnEvent 回调) ← 断点 4
|
||||
├ pi-native: agentx.Run(OnEvent) ← 断点 2/3（直通）
|
||||
│ └ llm.Stream: reasoning/text/tool delta ← 断点 1
|
||||
└ pi-agent: POST runner /v1/run/stream ← 断点 5（NDJSON 流）
|
||||
└ Streamer.EmitDetail(category, stepID, role, ...) ← 断点 6 之前：detail 已能出 Streamer
|
||||
└ Manager.appendEvent → task_run_events (seq) ← 持久化（已就绪）
|
||||
└ SSE: GET /tasks/runs/{id}/events ← 透传（已就绪）
|
||||
└ 前端 detail 渲染分支 ← 断点 7
|
||||
|
||||
【链路 A：聊天 SSE（P4 再做）】
|
||||
chat.Session → coordinator.generateDialogueReply
|
||||
└ dialogueBackend.Execute(OnEvent) → token 级 EmitDelta（替换一次性 EmitDelta 伪流式）
|
||||
```
|
||||
|
||||
### C.1 断点 1：reasoning_content 解析
|
||||
|
||||
- 文件：[internal/agentx/llm/sse.go](../../../internal/agentx/llm/sse.go)（`sseChoice.Delta` 增 `ReasoningContent string \`json:"reasoning_content"\``）、[internal/agentx/llm/types.go](../../../internal/agentx/llm/types.go)（`Event` 增 `ReasoningDelta` + `KindReasoningDelta`）
|
||||
- 语义：reasoning 与 text 分开累计；`message_end` 的 `Message.Content` 仍只含 text（与 runner `stripThinkBlocks` 语义对齐：正文无 think）。
|
||||
- 兼容性：无 reasoning 的模型零行为变化（字段空即无事件）。
|
||||
|
||||
### C.2 断点 2：agentx 循环转发 tool_call_delta
|
||||
|
||||
- 文件：[internal/agentx/loop.go:178-183](../../../internal/agentx/loop.go) `consume()`：`case llm.KindToolCallDelta` 增加转发（emit 一个 `tool_call_delta` Event，带 ToolName/ArgumentsDelta）；`KindReasoningDelta` 同理转发 `reasoning_delta`。
|
||||
- `agentx.Event`（loop.go:19）增字段：`Kind string`（llm 层事件细分）或复用 `Type` 前缀约定（`text_delta`/`reasoning_delta`/`tool_call_delta`/`tool_execution_start`/`tool_execution_end`…已够用，推荐复用 Type）。
|
||||
|
||||
### C.3 断点 3：Backend 接口过程回调
|
||||
|
||||
**设计取舍**：不改 `Backend.Execute` 签名（改签名会波及 stub/fallback/limiter 全部实现与调用方），改为 **Request 增可选回调**：
|
||||
|
||||
```go
|
||||
// internal/agent/backend.go 设计草案
|
||||
type Request struct {
|
||||
...
|
||||
// OnEvent 过程事件回调（nil = 不需要过程，行为与现状完全一致）
|
||||
// 实现方必须保证：回调在 Execute 返回前完成（同步或等待收尾），无泄漏 goroutine。
|
||||
OnEvent func(ev ProcessEvent)
|
||||
}
|
||||
|
||||
// ProcessEvent 对 agentx.Event 与 runner NDJSON 事件的统一投影
|
||||
type ProcessEvent struct {
|
||||
Type string // text_delta | reasoning_delta | tool_call_delta | tool_execution_start | tool_execution_end | message_end | ...
|
||||
Text string // delta 文本（Type 为 *delta 时）
|
||||
Tool string // 工具名
|
||||
Detail string // 工具参数/结果摘要（tool_execution_* 时）
|
||||
}
|
||||
```
|
||||
|
||||
- pi-native（[internal/agent/pinative.go](../../../internal/agent/pinative.go)）：`agentx.Run` 的 `OnEvent` 双写——既拼 `execution.log`（现状保留），又投影到 `req.OnEvent`；
|
||||
- `WrapWithConcurrencyLimit` 透传 Request 即透传回调，无需改；
|
||||
- fallback/stub：fallback 转发 primary 的回调事件即可；stub 无事件。
|
||||
|
||||
### C.4 断点 4：squad 层桥接 Streamer
|
||||
|
||||
[internal/squad/workflow.go](../../../internal/squad/workflow.go) `runOneStep`（构造 `agent.Request` 处，约 :543）：
|
||||
|
||||
```go
|
||||
// 设计草案
|
||||
if req.OnEvent == nil {
|
||||
req.OnEvent = func(ev agent.ProcessEvent) { /* 仅拼日志（现状） */ }
|
||||
}
|
||||
// 注入流式桥（streamer 可为 nil，安全）
|
||||
req.OnEvent = chainEvents(req.OnEvent, squadDetailBridge(streamer, step))
|
||||
```
|
||||
|
||||
`squadDetailBridge`（新函数，~40 行）把 `ProcessEvent` 翻译为 `EmitDetail`：
|
||||
|
||||
| agentx/runner 事件 | EmitDetail 映射 |
|
||||
|---|---|
|
||||
| `reasoning_delta` | category=agent_thought, delta 帧, ItemID=`step.ID#thought<N>` |
|
||||
| `text_delta` | category=agent_thought 的草稿子项（或 P2 可选：不改，任务正文走终态 step_output，避免与最终 deliverable 重复）→ **推荐 P0 不发 text_delta，仅 reasoning + tool** |
|
||||
| `tool_call_delta` | category=tool_call, delta 帧（入参增量）, ItemID=`step.ID#tool<N>` |
|
||||
| `tool_execution_start` | tool_call 项 status=calling 帧 |
|
||||
| `tool_execution_end` | tool_call 项 status=succeeded/failed + Output 摘要 + duration 帧 |
|
||||
| `message_end` | 对应 ItemID 的 done 帧 |
|
||||
|
||||
节流在 bridge 内做（100–300ms 合并、8KB 分帧、工具输出截断 N KB——工具结果可能巨大，只发首尾摘要）。
|
||||
|
||||
**B 轨参谋（链路 A，P4）**：[internal/coordinator/dialogue_llm.go](../../../internal/coordinator/dialogue_llm.go) `generateDialogueReply` 同样注入 `OnEvent`，把 `text_delta` 逐帧转 `streamer.EmitDelta`（替换 [internal/chat/session.go:624](../../../internal/chat/session.go) 一次性 `EmitDelta(NaturalReply)`），`reasoning_delta` 转 `EmitDetail(agent_thought)`。注意 chat SSE 无持久化——断线后靠 `chat_turns` 终态查询（现状机制），过程帧不可恢复（可接受，与豆包/kimi 行为一致）。
|
||||
|
||||
### C.5 断点 5：远程 runner 流式端点
|
||||
|
||||
[deploy/pi-agent/runner.mjs](../../../deploy/pi-agent/runner.mjs) 增加端点（不动 `/v1/run` 既有契约——`ResultV1` typed 校验、wire 预算 8MB 等全部保留）：
|
||||
|
||||
```
|
||||
POST /v1/run/stream Content-Type: application/x-ndjson（chunked）
|
||||
请求体 = /v1/run 同构 + stream: true
|
||||
响应 = 逐行 NDJSON：
|
||||
{"type":"proc","ev":{...pi RPC stdout 事件行...}} // 过程事件（原样转发已解析的行）
|
||||
...（执行中持续 flush）
|
||||
{"type":"result","body":{...与 /v1/run 完全相同的 typed 结果...}} // 终帧
|
||||
```
|
||||
|
||||
- runner 侧改动集中在 `handleRun`：stdout `on("data")` 里既有的逐行 JSON 解析循环（runner.mjs:365-388）旁路一个转发器；`stripThinkBlocks` 之前先把 think 块作为 `{"type":"reasoning_delta"}` proc 帧发出；
|
||||
- Go 侧 [internal/agent/piagent.go](../../../internal/agent/piagent.go) `executeRemote`：`req.OnEvent != nil` 时改走流式端点——`http.Client` 无总超时的流式变体（`Timeout` 由 ctx 控制），bufio 逐行读：`proc` 帧 → `req.OnEvent` 投影；`result` 帧 → 走既有 `DecodeRemoteRun` 校验路径（**终态校验逻辑零改动**，保证治理边界不放松）；
|
||||
- 降级：流式端点 404（旧 runner）→ 回落 `/v1/run` 一次性调用 + 无过程事件（`capabilities` 探测亦可，但 404 回落更简单）。
|
||||
|
||||
### C.6 断点 6：includeDetails 参数化
|
||||
|
||||
两处硬编码/丢弃：
|
||||
|
||||
1. [internal/taskrun/manager.go:364](../../../internal/taskrun/manager.go) `stream.WithListener(ctx, false, listener)`——worker 侧 Streamer 是否发 detail 由**创建时**决定。设计：`CreateAndStart`/`StartClaimedRun` 透传 `IncludeDetails`（`RunRecord` 或 `CreateOptions` 增字段；默认 false 保持现状）。**注意**：detail 是否持久化与是否推送解耦——建议 **detail 一律持久化**（`task_run_events` JSONB 足够），`includeDetails` 只控制 live 推送与 chat 通道；重放端点用 query `include_details=0` 过滤，历史 subagent 页面仍可全量回放。取舍理由：持久化膨胀可控（F.1 节流后估算），而"刷新后过程消失"违背透明可追溯原则；
|
||||
2. [internal/server/sse.go:476](../../../internal/server/sse.go) `_ = includeDetails`——`StreamChatRequest.IncludeDetails` 已存在（chat 通道）但 task 通道读取后丢弃；接线到 worker 创建。
|
||||
|
||||
### C.7 断点 7：前端协议与分支
|
||||
|
||||
- [web/src/api/chatStream.ts](../../../web/src/api/chatStream.ts) `StreamEvent` 增 `detail?: { category?: string; step_id?: string; role?: string; item_id?: string; seq?: number; delta?: string; content?: string; done?: boolean; title?: string; collapsible?: boolean; tool?: { name?: string; status?: string; input?: string; output?: string; duration_ms?: number } }`；
|
||||
- [web/src/hooks/useTaskRunAttempts.ts](../../../web/src/hooks/useTaskRunAttempts.ts) `onEvent` switch 增 `case 'detail'`：聚合进消息的 `processItems`（见 D.4），不进正文。
|
||||
|
||||
---
|
||||
|
||||
## D. subagent 子页面设计（重点）
|
||||
|
||||
### D.1 入口与路由
|
||||
|
||||
**决策：扩展 hash 路由，不引入 react-router。**
|
||||
|
||||
理由：
|
||||
|
||||
| 方案 | 优点 | 代价 |
|
||||
|---|---|---|
|
||||
| A. hash 扩展（推荐） | 复用 [web/src/engine/anchorNavigator.ts](../../../web/src/engine/anchorNavigator.ts) 既有解析/监听先例（`#/s/{sessionId}/m/{messageId}`）；零新依赖；与"工作台为唯一应用视图"的产品形态匹配 | 顶层 view 状态机要加一态；深链参数解析手写 |
|
||||
| B. react-router | 标准 | 新依赖 + 全仓路由心智迁移；App.tsx 1758 行重构风险；本需求只有一个子页面，收益不成比例 |
|
||||
|
||||
路由形态：
|
||||
|
||||
```
|
||||
#/s/{sessionId}/run/{runId}/agents → 子页面（默认选中当前 running/最近完成成员）
|
||||
#/s/{sessionId}/run/{runId}/agents/{roleId} → 子页面（选中指定成员）
|
||||
```
|
||||
|
||||
实现骨架（设计草案）：
|
||||
|
||||
- App.tsx 顶层 view 状态机扩展：`'workbench' | 'admin' | 'subagents'`；`parseAnchorHash` 扩展识别 `/run/{runId}/agents` 段；
|
||||
- `handleSubagentsView` 由 TaskRunCard 的"查看小队过程"入口触发（写 hash 触发 hashchange，与消息深链同一机制）；
|
||||
- 子页面经 `OnDemand`（[web/src/components/lazy/OnDemand.tsx](../../../web/src/components/lazy/OnDemand.tsx)）懒加载，**props 最小化**：仅 `runId`、`sessionId`、`onExit`（返回 workbench）——规避 App.tsx props 钻透：子页面自持 SSE 订阅与状态，不依赖父容器消息流。
|
||||
|
||||
### D.2 页面信息架构
|
||||
|
||||
```
|
||||
┌──────────────────────────────────────────────────────────────┐
|
||||
│ 小队过程 · {任务标题} · run:{runId 截断} [刷新] [返回对话] │
|
||||
├──────────────┬───────────────────────────────────────────────┤
|
||||
│ 成员卡列表 │ 选中成员（role: regulatory 法规专员） │
|
||||
│ │ ┌ 状态条: ⚙️ 正在执行 · 已用 2m13s · 第2/5步 ┐ │
|
||||
│ 👤 organizer │ └──────────────────────────────────────────┘ │
|
||||
│ ✅ 完成 1m2s│ ┌ ThinkingBlock（默认折叠）──────────────┐ │
|
||||
│ 👤 searcher │ │ 💭 思考中…（流式增量，呼吸态） │ │
|
||||
│ ✅ 完成 3m4s│ └───────────────────────────────────────┘ │
|
||||
│ 👤 regulatory│ ┌ ToolCallCard ─────────────────────────┐ │
|
||||
│ ⚙️ 执行中 │ │ 🔧 read_file tasks/01.../output.md │ │
|
||||
│ 👤 writer │ │ ✅ 213ms · 12KB │ │
|
||||
│ ⏳ 等待 │ └──────────────────────────────────────┘ │
|
||||
│ 👤 leader │ ┌ ThinkingBlock（第二段）──────────────┐ │
|
||||
│ ⏳ 等待 │ ... │ │
|
||||
│ │ ┌ 步骤产出 output.md（终态 step_output）┐ │
|
||||
│ [成员详情] │ │ 📄 阶段性产出（可展开全文） │ │
|
||||
│ │ └──────────────────────────────────────┘ │
|
||||
├──────────────┴───────────────────────────────────────────────┤
|
||||
│ 事件时间线（跨成员 phase 事件折叠，复用 applyPhaseEvent 语义） │
|
||||
└──────────────────────────────────────────────────────────────┘
|
||||
```
|
||||
|
||||
- **左栏数据源**：`squad.json`（成员编制：role/Name/Responsibilities——工作区文件，经新只读端点或既有 workspace API 投影）+ phase/detail 事件聚合出状态与耗时（`step_executing`/`step_completed`/`step_failed` by role）；
|
||||
- **右栏数据源**：detail 事件流按 `role == selected` 过滤，聚合为 `processItems`（见 D.4）；
|
||||
- **底部时间线**：全量 phase 事件（复用 `TaskRunCard` 折叠逻辑与 `PhaseEvent` 类型）。
|
||||
|
||||
### D.3 三类过程组件
|
||||
|
||||
| 组件 | 状态 | 内容来源 | 交互 |
|
||||
|---|---|---|---|
|
||||
| `ThinkingBlock` | thinking（呼吸态+增量滚动）/ done（折叠为一行摘要：前 40 字 + 字数）/ error | `agent_thought` 增量帧聚合 | 默认折叠；点击展开全文（Markdown 渲染）；进行中可"实时展开" |
|
||||
| `ToolCallCard` | calling（转轮）/ succeeded（绿勾+耗时）/ failed（红叹+错误摘要） | `tool_call` 帧聚合（name/input/output/status/duration） | 入参/结果默认摘要（N KB 截断），点击展开 JSON/代码（**需补块级渲染**，见 D.6） |
|
||||
| `ProcessStepOutput` | done | 既有 `step_output` 整段 content | 默认折叠，展开渲染 Markdown |
|
||||
|
||||
组件均为纯展示组件（props 驱动），放置 `web/src/components/subagents/`，配 `node:test` 静态断言测试（与仓内组件测试形态一致）。
|
||||
|
||||
### D.4 前端数据流与状态
|
||||
|
||||
```typescript
|
||||
// 设计草案：web/src/hooks/useRunProcessStream.ts
|
||||
type ProcessItem = {
|
||||
key: string; // `${stepId}/${itemId}`
|
||||
category: 'agent_thought' | 'tool_call' | 'step_output' | 'log';
|
||||
role: string;
|
||||
stepId: string;
|
||||
title?: string;
|
||||
text: string; // delta 累计或 content 快照
|
||||
status: 'streaming' | 'done' | 'error';
|
||||
tool?: { name: string; status: string; input?: string; output?: string; durationMs?: number };
|
||||
updatedAt: number;
|
||||
};
|
||||
|
||||
function applyDetailEvent(items: Record<string, ProcessItem>, ev: DetailEvent): Record<string, ProcessItem>
|
||||
// 纯函数（与 applyPhaseEvent 同风格，node:test 可测）：
|
||||
// delta → 追加 text、status=streaming
|
||||
// content → 覆盖 text（快照对账）
|
||||
// done → status=done
|
||||
// tool 帧 → 合并 tool 字段
|
||||
```
|
||||
|
||||
订阅策略（复用既有骨架）：
|
||||
|
||||
- **live**：`subscribeTaskRunEvents(runId, { onEvent, afterSeq })`（[web/src/api/taskRuns.ts:113](../../../web/src/api/taskRuns.ts)），事件分发：`phase` → 左栏状态 + 底部时间线；`detail` → `applyDetailEvent`；
|
||||
- **刷新/进入已结束 run**：先 `readTaskRunSSEOnce(runId, 0, …, { replayOnly: true })` 全量重放（事件已持久化，见 C.6），再决定是否续订 live；
|
||||
- **超量回放保护**：`after_seq` 分页重放（既有能力），detail 大 run 可先只重放 phase + 按需拉成员 detail（`GET /runs/{id}/events?role=xxx&include_details=1` 过滤参数为可选增强，P2.5）。
|
||||
|
||||
### D.5 与主对话气泡的联动
|
||||
|
||||
- 气泡内 `TaskRunCard` 保持现状（L1 时间线），新增两处：
|
||||
1. 运行中：卡片底部一行"查看小队过程 →"（写 hash 深链进子页面）；
|
||||
2. 完成后：L3 摘要条（如"5 名专员 · 12 次工具调用 · 3 段推演"），点击同样进子页面；
|
||||
- **不把全量 detail 渲染进气泡**——认知负荷原则：气泡只承载 L1/L2，L3 全文在子页面。`MessageBubble` 现有 `message.phaseEvents` 通道不动。
|
||||
|
||||
### D.6 前端基础设施缺口补齐（实施前置）
|
||||
|
||||
1. **块级 Markdown + 代码高亮**：现状 `renderInlineMarkdown` 仅行内语法（[web/src/engine/renderMarkdown.ts](../../../web/src/engine/renderMarkdown.ts)）。工具调用入参/结果需要 fenced code block + JSON 高亮。方案：`renderMarkdown` 增块级分支（heading/fence/list——解析器自研或引入 `marked`/`react-markdown`；倾向自研轻量扩展，避免新依赖与 XSS 面扩大，沿用现有 escape 纪律）；
|
||||
2. **高频增量渲染性能**：processItems 更新频率受后端节流（100–300ms 批帧）保护；子页面右栏**不用虚拟列表**（成员过程项数量级 ~10²），但聚合函数必须纯函数 + `React.memo` 行项；
|
||||
3. **通用折叠组件**：自研 `CollapsibleBlock`（受控 `open` + `Summary` 插槽），三组件复用。
|
||||
|
||||
### D.7 后端配套只读端点（成员编制）
|
||||
|
||||
子页面左栏需要 `squad.json`。设计（P2 内最小）：
|
||||
|
||||
```
|
||||
GET /api/v1/tasks/runs/{run_id}/squad
|
||||
→ { members: [{ role, name, responsibilities, source }], task_key, title }
|
||||
```
|
||||
|
||||
实现：taskrun Manager 读 `record.WorkspaceID` → 工作区 `squad.json` 反序列化（只读投影，权限沿用 run 事件的 `RequireProjectAccess`）。
|
||||
|
||||
---
|
||||
|
||||
## E. 分期实施路线
|
||||
|
||||
| 期 | 内容 | 验证形态（按验证契约） | 依赖 |
|
||||
|---|---|---|---|
|
||||
| **P0 协议+直通** | B.2/B.3 协议扩展；C.1–C.4（pi-native 全链路）；C.6 includeDetails 参数化 + detail 持久化；前端 chatStream 类型 + detail 分支（仅聚合不渲染） | Go 单测（llm sse_test 增 reasoning case、squad bridge 节流测试）；stub/pi-native E2E；真实 PI Runner（dev 环境）冒烟 | 无 |
|
||||
| **P1 气泡过程块** | D.5 气泡内 L3 摘要 + 折叠 ThinkingBlock（最小渲染，块级 markdown 只做 fence+JSON） | 前端 node:test + 真实浏览器验收（交互变更） | P0 |
|
||||
| **P2 subagent 子页面** | D.1–D.7 全量（hash 路由、成员卡、过程流、squad 端点） | 前端单测 + Playwright spec（新增 e2e/subagents-view.spec.ts）+ 真实 PI 全链路验收 | P1 |
|
||||
| **P2.5（可选）** | 事件过滤参数（role/include_details query）；大 run 分页回放 | 接口测试 | P2 |
|
||||
| **P3 runner 流式** | C.5 `/v1/run/stream` + piagent 流式消费 + 404 回落 | runner 单测（mjs 侧 settle-policy 风格）+ 真实容器验收（dev） | P0（可并行） |
|
||||
| **P4 B 轨参谋** | C.4 尾段：dialogue_llm OnEvent → token 级 EmitDelta + reasoning detail | 真实 PI 对话链路验收（chat SSE） | P0 |
|
||||
| **P5（演进）** | 历史回放端点（transcript.jsonl/execution.log 投影 API）；per-member 持久化表（如 `task_run_step_events`，仅当 JSONB 容量实测超限） | 视实测 | P2/P3 |
|
||||
|
||||
每期独立合入（Conventional Commits：`feat(stream)/feat(web)` 前后端分原子提交），P0 即可演示 pi-native 路径的"思考流+工具卡"。
|
||||
|
||||
---
|
||||
|
||||
## F. 风险与边界
|
||||
|
||||
### F.1 事件量与持久化膨胀
|
||||
|
||||
- **节流义务在 bridge**（C.4）：reasoning/text delta 合并 100–300ms 批帧；单帧 ≤8KB；工具输出截断（如 4KB 首部 + 省略标记）；`step_output` 保持现状单帧（终态产物已有 `deliverable` 通道与文档存储，不重复全文持久化时可评估只存摘要——P0 先全文，P2.5 实测再收紧）；
|
||||
- **容量估算**：单 run 5 成员 × (1 思考块 2–8KB + 6 工具调用 × 1KB 摘要) ≈ 50KB/run 量级（节流后），`task_run_events` JSONB 单行无硬限，可控；
|
||||
- **兜底**：`execution.log`/`transcript.jsonl` 仍全量落盘工作区（现状不变），P5 投影 API 作历史回放。
|
||||
|
||||
### F.2 认知负荷与安全
|
||||
|
||||
- detail 默认折叠贯穿三处（气泡摘要、子页面 ThinkingBlock、ToolCallCard 入参/结果摘要）——继承 2026-09-09 设计 P1 约束；
|
||||
- **脱敏**：bridge 层过滤——不透传 prompt/AGENTS.md 全文（这些是任务输入不是过程）；工具入参只发 workspace 相对路径相关字段；配置/密钥字段（provider/key）永不进事件（`ProcessEvent` 设计上就不含这些来源）；
|
||||
- **治理红线**：过程事件不写 `events` 域总线、不参与验收判定；测试/审查结论仍以报告路径与 receipt 为准（pi-subagents 规则），本设计纯可视化。
|
||||
|
||||
### F.3 兼容与降级
|
||||
|
||||
- 旧 runner（无 `/v1/run/stream`）→ 404 回落一次性调用，功能退化为 L1 时间线（现状）；
|
||||
- `OnEvent == nil` → 所有后端行为与现状完全一致（回调是纯增量）；
|
||||
- 前端不识别 detail（旧前端 + 新后端）→ JSON 字段忽略，无破坏；
|
||||
- `includeDetails` 默认 false → chat 通道现状不变。
|
||||
|
||||
### F.4 并行正确性
|
||||
|
||||
- 并行组多成员同时发 detail：`Streamer.Emit` 已有 RWMutex；bridge 无共享状态（每 step 独立闭包）；
|
||||
- 前端聚合键含 role/stepID（B.4），并行不串扰；
|
||||
- `step_output` 终态帧在并行完成回调内发出（现状 workflow.go:746 位置不变）。
|
||||
|
||||
### F.5 明确不做
|
||||
|
||||
- 不引入 WebSocket（SSE + after_seq 重放已覆盖需求）；
|
||||
- 不做过程事件的域总线登记（B.5 理由）；
|
||||
- 不做 B 轨 chat 过程帧的持久化与断线重放（P4 live-only，终态落库现状）；
|
||||
- 本期不实现任何代码（用户授权边界）。
|
||||
|
||||
---
|
||||
|
||||
## G. 附录：断点-文件-方案对照表
|
||||
|
||||
| # | 断点 | 文件:行 | 方案章节 |
|
||||
|---|---|---|---|
|
||||
| 1 | reasoning_content 未解析 | internal/agentx/llm/sse.go、types.go | C.1 / B.3 |
|
||||
| 2 | agentx 丢弃 tool_call_delta | internal/agentx/loop.go:178-183 | C.2 |
|
||||
| 3 | Backend 黑盒无回调 | internal/agent/backend.go:38、pinative.go:102 | C.3 |
|
||||
| 4 | squad 层无成员过程流 | internal/squad/workflow.go:543(runOneStep) | C.4 |
|
||||
| 5 | runner /v1/run 一次性 | deploy/pi-agent/runner.mjs:600 | C.5 |
|
||||
| 6 | includeDetails 硬编码 false / 丢弃 | internal/taskrun/manager.go:364、server/sse.go:476 | C.6 |
|
||||
| 7 | 前端无 detail 协议与分支 | web/src/api/chatStream.ts、hooks/useTaskRunAttempts.ts:474 | C.7 |
|
||||
|
||||
## H. 参考文件
|
||||
|
||||
- 三层信息架构基座：`.doc/archive/design/superpowers/specs/2026-09-08-streaming-events-and-digestible-ux-design.md`
|
||||
- 认知负荷约束：`.doc/archive/design/superpowers/specs/2026-09-09-chat-wait-error-retry-ux-design.md`
|
||||
- 小队编排（含 §18.5 演进方向）：`.doc/system-design/agent-squad/02-智能体小队编排-设计-思路-架构-实现-难点-解决方案-亮点与优势.md`
|
||||
- 长任务 resume：`.doc/archive/design/superpowers/specs/2026-09-09-longtask-phase-live-resume-ux-design.md`
|
||||
- 任务记录：`.doc/ops/short-term/2026-09-28-subagent-process-view-design.md`
|
||||
@@ -7,6 +7,14 @@
|
||||
|
||||
## 📋 报告清单
|
||||
|
||||
### 0c. 小队过程流式展示与 subagent 子页面（2026-09-28）
|
||||
|
||||
| 项 | 链接 |
|
||||
|-----|------|
|
||||
| 主设计 | [2026-09-28-agent-process-streaming-and-subagent-view-design.md](2026-09-28-agent-process-streaming-and-subagent-view-design.md) |
|
||||
| 任务 | [../ops/short-term/2026-09-28-subagent-process-view-design.md](../ops/short-term/2026-09-28-subagent-process-view-design.md) |
|
||||
| 一句话总结 | 思考/工具调用/过程内容七断点打通（reasoning 解析→Backend 回调→runner 流式→detail 持久化）+ subagent 子页面（hash 路由 + 成员卡 + 过程流组件）；设计稿未实施 |
|
||||
|
||||
### 0b. 自由协作与认领提醒（2026-09-24）
|
||||
|
||||
| 项 | 链接 |
|
||||
|
||||
@@ -51,6 +51,33 @@ function extractAssistantText(ev) {
|
||||
return chunks.join("");
|
||||
}
|
||||
|
||||
// toProcFrame 把 pi RPC stdout 事件行投影为 Go 侧 agent.ProcessEvent 同构帧
|
||||
// （2026-09-28 过程流式设计 C.5）。只保留过程可视化需要的类型，其他类型原样
|
||||
// 透传 type 供前端扩展。think 块在 stripThinkBlocks 剥离前分流为 reasoning_delta。
|
||||
function toProcFrame(ev) {
|
||||
const t = ev.type || "";
|
||||
const msg = ev.message && typeof ev.message === "object" ? ev.message : null;
|
||||
if (t === "message_end" || t === "turn_end") {
|
||||
return { type: "message_end" };
|
||||
}
|
||||
if (msg && msg.role === "assistant") {
|
||||
const parts = Array.isArray(msg.content) ? msg.content : [];
|
||||
for (const part of parts) {
|
||||
if (part && typeof part === "object" && part.type === "reasoning" && typeof part.text === "string") {
|
||||
return { type: "reasoning_delta", text: part.text };
|
||||
}
|
||||
}
|
||||
}
|
||||
if (t === "tool_execution_start" || t === "tool_execution_end") {
|
||||
return {
|
||||
type: t,
|
||||
tool: typeof ev.tool === "string" ? ev.tool : (msg && typeof msg.name === "string" ? msg.name : ""),
|
||||
detail: typeof ev.detail === "string" ? ev.detail : "",
|
||||
};
|
||||
}
|
||||
return { type: t };
|
||||
}
|
||||
|
||||
function materializeReasoningTargets({
|
||||
workspace,
|
||||
targets,
|
||||
@@ -254,7 +281,7 @@ function authorizeRunner(req, res) {
|
||||
return false;
|
||||
}
|
||||
|
||||
async function handleRun(body) {
|
||||
async function handleRun(body, onProcEvent) {
|
||||
const rawWs = body.workspace || "";
|
||||
const workspace = resolveWorkspace(rawWs);
|
||||
const model = body.model || runtimeModel || DEFAULT_MODEL;
|
||||
@@ -382,6 +409,11 @@ async function handleRun(body) {
|
||||
lastAssistantText = assistantChunk;
|
||||
}
|
||||
applyAgentEvent(settle, ev);
|
||||
if (onProcEvent) {
|
||||
try { onProcEvent(toProcFrame(ev)); } catch (e) {
|
||||
console.warn("[runner] proc-event forward failed:", e.message);
|
||||
}
|
||||
}
|
||||
if (t === "agent_settled") {
|
||||
child.stdin.end();
|
||||
}
|
||||
@@ -596,6 +628,35 @@ const server = http.createServer(async (req, res) => {
|
||||
return;
|
||||
}
|
||||
|
||||
// 4a. Streaming Agent Task Run Endpoint (NDJSON chunked, 2026-09-28 process streaming design C.5)
|
||||
if (method === "POST" && url.pathname === "/v1/run/stream") {
|
||||
if (!authorizeRunner(req, res)) {
|
||||
return;
|
||||
}
|
||||
res.writeHead(200, {
|
||||
"Content-Type": "application/x-ndjson",
|
||||
"Cache-Control": "no-cache",
|
||||
"X-Accel-Buffering": "no",
|
||||
});
|
||||
const writeFrame = (obj) => {
|
||||
try {
|
||||
res.write(JSON.stringify(obj) + "\n");
|
||||
} catch {
|
||||
// client disconnect: keep running so the run still settles; write errors ignored
|
||||
}
|
||||
};
|
||||
const onProcEvent = (frame) => writeFrame({ type: "proc", ev: frame });
|
||||
try {
|
||||
const body = await parseJsonBody(req);
|
||||
const result = await handleRun(body, onProcEvent);
|
||||
writeFrame({ type: "result", body: withResultV1(result) });
|
||||
} catch (err) {
|
||||
writeFrame({ type: "result", body: withResultV1({ ok: false, error: "runner_failed", patches: [], run_id: "unknown" }) });
|
||||
}
|
||||
res.end();
|
||||
return;
|
||||
}
|
||||
|
||||
// 4. Agent Task Run Endpoint
|
||||
if (method === "POST" && url.pathname === "/v1/run") {
|
||||
if (!authorizeRunner(req, res)) {
|
||||
|
||||
@@ -15,6 +15,19 @@ type Request struct {
|
||||
Timeout time.Duration `json:"timeout"`
|
||||
// ToolProfile is workspace (default, document tools) or reasoning (no tools).
|
||||
ToolProfile string `json:"tool_profile,omitempty"`
|
||||
// OnEvent 可选过程事件回调（nil = 调用方不需要过程，后端行为与现状完全一致）。
|
||||
// 实现方必须保证：回调在 Execute 返回前完成（同步调用或等待收尾），无泄漏 goroutine；
|
||||
// 回调内不得阻塞（消费方自行节流/异步入队）。
|
||||
OnEvent func(ev ProcessEvent)
|
||||
}
|
||||
|
||||
// ProcessEvent 对 agentx 事件与远程 runner NDJSON 过程帧的统一投影，
|
||||
// 供小队过程流式展示消费（见 .doc/system-design/2026-09-28-agent-process-streaming-and-subagent-view-design.md）。
|
||||
type ProcessEvent struct {
|
||||
Type string // text_delta | reasoning_delta | tool_call_delta | tool_execution_start | tool_execution_end | message_end | ...
|
||||
Text string // delta 文本（Type 为 *_delta 时）
|
||||
Tool string // 工具名（tool_* 时）
|
||||
Detail string // 工具参数/结果摘要（tool_execution_* 时）
|
||||
}
|
||||
|
||||
const (
|
||||
|
||||
@@ -1,6 +1,7 @@
|
||||
package agent
|
||||
|
||||
import (
|
||||
"bufio"
|
||||
"bytes"
|
||||
"context"
|
||||
"encoding/json"
|
||||
@@ -217,6 +218,14 @@ func (p *PiAgentBackend) executeRemote(ctx context.Context, runnerURL string, re
|
||||
}
|
||||
reqBody, _ := json.Marshal(p.runnerPayload(req, workDir, true, timeout))
|
||||
|
||||
// 过程流式路径（2026-09-28 设计 C.5）：调用方带 OnEvent 时优先走 NDJSON 流式
|
||||
// 端点，实时转发 proc 帧；旧 runner 404 时回落一次性 /v1/run（无过程事件）。
|
||||
if req.OnEvent != nil {
|
||||
if res, attempted, err := p.executeRemoteStream(ctx, runnerURL, reqBody, req, expectedRunID, timeout); attempted {
|
||||
return res, err
|
||||
}
|
||||
}
|
||||
|
||||
httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, runnerURL+"/v1/run", bytes.NewReader(reqBody))
|
||||
if err != nil {
|
||||
return &Result{
|
||||
@@ -244,8 +253,85 @@ func (p *PiAgentBackend) executeRemote(ctx context.Context, runnerURL string, re
|
||||
if len(limited) > MaxResultBodyBytes {
|
||||
return &Result{Success: false, ErrorReason: "runner result body too large"}, nil
|
||||
}
|
||||
return decodeRemoteResult(limited, expectedRunID, workDir, req)
|
||||
}
|
||||
|
||||
decoded, decErr := DecodeRemoteRun(limited, expectedRunID, true)
|
||||
// executeRemoteStream 调用 runner /v1/run/stream：proc 帧实时转发 OnEvent，
|
||||
// result 终帧走与一次性路径完全相同的 DecodeRemoteRun 校验。
|
||||
// attempted=false 表示端点不可用（404/405），调用方应回落 /v1/run。
|
||||
func (p *PiAgentBackend) executeRemoteStream(ctx context.Context, runnerURL string, reqBody []byte, req Request, expectedRunID string, timeout time.Duration) (*Result, bool, error) {
|
||||
httpReq, err := http.NewRequestWithContext(ctx, http.MethodPost, runnerURL+"/v1/run/stream", bytes.NewReader(reqBody))
|
||||
if err != nil {
|
||||
return nil, false, nil
|
||||
}
|
||||
httpReq.Header.Set("Content-Type", "application/json")
|
||||
httpReq.Header.Set("Accept", "application/x-ndjson")
|
||||
ApplyRunnerAuth(httpReq, p.cfg.PiAgentRunnerAuthToken)
|
||||
|
||||
// 无总超时的 client：超时由 ctx 控制，逐行流式读取期间保持连接。
|
||||
client := &http.Client{}
|
||||
resp, err := client.Do(httpReq)
|
||||
if err != nil {
|
||||
return nil, false, nil // 网络错误也回落一次性路径
|
||||
}
|
||||
defer resp.Body.Close()
|
||||
if resp.StatusCode != http.StatusOK {
|
||||
// 任何非 200（含旧 runner 404 与瞬态 5xx）都回落一次性路径，
|
||||
// 保证错误分类与既有 /v1/run 语义完全一致（不改变治理边界）。
|
||||
resp.Body.Close()
|
||||
return nil, false, nil
|
||||
}
|
||||
|
||||
var resultBody []byte
|
||||
var sawResult bool
|
||||
sc := bufio.NewScanner(resp.Body)
|
||||
sc.Buffer(make([]byte, 0, 64*1024), MaxResultBodyBytes)
|
||||
for sc.Scan() {
|
||||
if ctx.Err() != nil {
|
||||
return &Result{Success: false, ErrorReason: "stream context cancelled", Retryable: true}, true, nil
|
||||
}
|
||||
line := strings.TrimSpace(sc.Text())
|
||||
if line == "" {
|
||||
continue
|
||||
}
|
||||
var frame struct {
|
||||
Type string `json:"type"`
|
||||
Ev *ProcessEvent `json:"ev,omitempty"`
|
||||
Body json.RawMessage `json:"body,omitempty"`
|
||||
}
|
||||
if err := json.Unmarshal([]byte(line), &frame); err != nil {
|
||||
continue
|
||||
}
|
||||
switch frame.Type {
|
||||
case "proc":
|
||||
if frame.Ev != nil && req.OnEvent != nil {
|
||||
req.OnEvent(*frame.Ev)
|
||||
}
|
||||
case "result":
|
||||
resultBody = frame.Body
|
||||
sawResult = true
|
||||
}
|
||||
if sawResult {
|
||||
break // result 是终帧
|
||||
}
|
||||
}
|
||||
if !sawResult {
|
||||
return &Result{
|
||||
Success: false,
|
||||
ErrorReason: "pi runner stream ended without result frame",
|
||||
Retryable: true,
|
||||
}, true, nil
|
||||
}
|
||||
res, decErr := decodeRemoteResult(resultBody, expectedRunID, req.WorkDir, req)
|
||||
if decErr != nil {
|
||||
return res, true, decErr
|
||||
}
|
||||
return res, true, nil
|
||||
}
|
||||
|
||||
// decodeRemoteResult 一次性与流式路径共用的终态解码与落盘。
|
||||
func decodeRemoteResult(body []byte, expectedRunID, workDir string, req Request) (*Result, error) {
|
||||
decoded, decErr := DecodeRemoteRun(body, expectedRunID, true)
|
||||
if decErr != nil {
|
||||
reason := "typed result rejected"
|
||||
msg := decErr.Error()
|
||||
|
||||
@@ -93,6 +93,7 @@ func (p *PiNativeBackend) Execute(ctx context.Context, req Request) (*Result, er
|
||||
defer cancel()
|
||||
|
||||
var logs strings.Builder
|
||||
onProcess := req.OnEvent
|
||||
runRes, err := agentx.Run(runCtx, agentx.Config{
|
||||
Model: model,
|
||||
SystemPrompt: nativeSystemPrompt(req, profile),
|
||||
@@ -110,6 +111,14 @@ func (p *PiNativeBackend) Execute(ctx context.Context, req Request) (*Result, er
|
||||
logs.WriteString(ev.Text)
|
||||
}
|
||||
logs.WriteByte('\n')
|
||||
if onProcess != nil {
|
||||
onProcess(ProcessEvent{
|
||||
Type: ev.Type,
|
||||
Text: ev.Text,
|
||||
Tool: ev.Tool,
|
||||
Detail: ev.Detail,
|
||||
})
|
||||
}
|
||||
},
|
||||
}, llm.Message{Role: llm.RoleUser, Content: req.Prompt})
|
||||
if err != nil {
|
||||
|
||||
@@ -11,6 +11,7 @@ import (
|
||||
type sseChoice struct {
|
||||
Delta struct {
|
||||
Content string `json:"content"`
|
||||
Reasoning string `json:"reasoning_content"`
|
||||
ToolCalls []struct {
|
||||
Index int `json:"index"`
|
||||
ID string `json:"id"`
|
||||
@@ -94,6 +95,17 @@ func emitParsed(ctx context.Context, r io.Reader, out chan<- Event) {
|
||||
case out <- Event{Kind: KindTextDelta, Text: choice.Delta.Content}:
|
||||
}
|
||||
}
|
||||
// reasoning_content（国产 OpenAI 兼容模型的思考链）与正文分流：
|
||||
// 只作为 reasoning_delta 事件外发，不计入 Message.Content（与 runner
|
||||
// stripThinkBlocks"正文无 think"语义对齐）。
|
||||
if choice.Delta.Reasoning != "" {
|
||||
select {
|
||||
case <-ctx.Done():
|
||||
emitAbort(out, content.String(), ctx.Err())
|
||||
return
|
||||
case out <- Event{Kind: KindReasoningDelta, ReasoningDelta: choice.Delta.Reasoning}:
|
||||
}
|
||||
}
|
||||
for _, tc := range choice.Delta.ToolCalls {
|
||||
cur, ok := calls[tc.Index]
|
||||
if !ok {
|
||||
|
||||
@@ -189,3 +189,36 @@ func TestStreamErrorString(t *testing.T) {
|
||||
t.Fatal("error string")
|
||||
}
|
||||
}
|
||||
|
||||
func TestParseSSEReasoningDelta(t *testing.T) {
|
||||
body := strings.Join([]string{
|
||||
`data: {"choices":[{"delta":{"reasoning_content":"先分析需求"},"index":0}]}`,
|
||||
`data: {"choices":[{"delta":{"reasoning_content":"再检索标准"},"index":0}]}`,
|
||||
`data: {"choices":[{"delta":{"content":"结论"},"finish_reason":"stop"}],"usage":{"prompt_tokens":1,"completion_tokens":2,"total_tokens":3}}`,
|
||||
`data: [DONE]`,
|
||||
"",
|
||||
}, "\n")
|
||||
var reasoning []string
|
||||
var text []string
|
||||
var end Event
|
||||
for ev := range ParseSSE(context.Background(), strings.NewReader(body)) {
|
||||
switch ev.Kind {
|
||||
case KindReasoningDelta:
|
||||
reasoning = append(reasoning, ev.ReasoningDelta)
|
||||
case KindTextDelta:
|
||||
text = append(text, ev.Text)
|
||||
case KindMessageEnd:
|
||||
end = ev
|
||||
}
|
||||
}
|
||||
if strings.Join(reasoning, "") != "先分析需求再检索标准" {
|
||||
t.Fatalf("reasoning deltas %v", reasoning)
|
||||
}
|
||||
// 正文与思考链分流：Content 只含 text，不含 reasoning
|
||||
if strings.Join(text, "") != "结论" {
|
||||
t.Fatalf("text deltas %v", text)
|
||||
}
|
||||
if end.Message.Content != "结论" {
|
||||
t.Fatalf("end content %q must not include reasoning", end.Message.Content)
|
||||
}
|
||||
}
|
||||
|
||||
@@ -30,6 +30,7 @@ type EventKind string
|
||||
|
||||
const (
|
||||
KindTextDelta EventKind = "text_delta"
|
||||
KindReasoningDelta EventKind = "reasoning_delta"
|
||||
KindToolCallDelta EventKind = "tool_call_delta"
|
||||
KindMessageEnd EventKind = "message_end"
|
||||
KindError EventKind = "error"
|
||||
@@ -72,6 +73,7 @@ type Usage struct {
|
||||
type Event struct {
|
||||
Kind EventKind
|
||||
Text string
|
||||
ReasoningDelta string
|
||||
ToolIndex int
|
||||
ToolID string
|
||||
ToolName string
|
||||
|
||||
@@ -22,6 +22,9 @@ type Event struct {
|
||||
Message llm.Message
|
||||
Tool string
|
||||
Text string
|
||||
// Detail carries tool arguments/results or reasoning deltas for
|
||||
// process-visualization consumers; log writers may ignore it.
|
||||
Detail string
|
||||
}
|
||||
|
||||
// Config drives one agent session.
|
||||
@@ -178,6 +181,10 @@ func consume(ctx context.Context, cfg Config, messages []llm.Message) (llm.Event
|
||||
switch ev.Kind {
|
||||
case llm.KindTextDelta:
|
||||
emit(cfg, Event{Type: "text_delta", Text: ev.Text})
|
||||
case llm.KindReasoningDelta:
|
||||
emit(cfg, Event{Type: "reasoning_delta", Text: ev.ReasoningDelta})
|
||||
case llm.KindToolCallDelta:
|
||||
emit(cfg, Event{Type: "tool_call_delta", Tool: ev.ToolName, Detail: ev.ArgumentsDelta})
|
||||
case llm.KindError:
|
||||
return ev, ev.Err
|
||||
case llm.KindMessageEnd:
|
||||
|
||||
@@ -233,6 +233,7 @@ func (s *Server) Routes() http.Handler {
|
||||
mux.HandleFunc("GET /api/v1/tasks/runs", s.handleListTaskRuns)
|
||||
mux.HandleFunc("GET /api/v1/tasks/runs/{run_id}", s.handleGetTaskRun)
|
||||
mux.HandleFunc("GET /api/v1/tasks/runs/{run_id}/events", s.handleTaskRunEvents)
|
||||
mux.HandleFunc("GET /api/v1/tasks/runs/{run_id}/squad", s.handleGetTaskRunSquad)
|
||||
mux.HandleFunc("POST /api/v1/tasks/runs/{run_id}/cancel", s.handleCancelTaskRun)
|
||||
mux.HandleFunc("POST /api/v1/tasks/runs/{run_id}/resume", s.handleResumeTaskRun)
|
||||
mux.HandleFunc("POST /api/v1/admin/"+workspace.WorkspacesDirName+"/reconcile", s.handleReconcileWorkspace)
|
||||
|
||||
@@ -473,6 +473,9 @@ func (s *Server) handleRunTaskStream(w http.ResponseWriter, r *http.Request) {
|
||||
return
|
||||
}
|
||||
}
|
||||
// includeDetails 参数保留解析兼容，但 worker 侧已一律产出 detail 事件
|
||||
// （2026-09-28 过程流式设计 C.6：detail 全量持久化与推送，认知负荷由前端
|
||||
// 默认折叠保证；该参数不再控制任务通道的事件产出）。
|
||||
_ = includeDetails
|
||||
_ = requestID
|
||||
|
||||
|
||||
@@ -672,6 +672,48 @@ func (s *Server) handleGetTaskRun(w http.ResponseWriter, r *http.Request) {
|
||||
writeJSONResponse(w, http.StatusOK, s.buildTaskRunResponse(rec))
|
||||
}
|
||||
|
||||
// handleGetTaskRunSquad 只读投影工作区 squad.json，供小队过程子页面左栏。
|
||||
func (s *Server) handleGetTaskRunSquad(w http.ResponseWriter, r *http.Request) {
|
||||
user, ok := s.requirePrincipal(w, r)
|
||||
if !ok {
|
||||
return
|
||||
}
|
||||
runID := r.PathValue("run_id")
|
||||
if runID == "" {
|
||||
writeSafeError(w, r, http.StatusBadRequest, "run_id required")
|
||||
return
|
||||
}
|
||||
if !s.assertTaskRunOwnedByUser(w, r.Context(), runID, user) {
|
||||
return
|
||||
}
|
||||
rec, err := s.taskRuns().GetRun(runID)
|
||||
if err != nil {
|
||||
writeDomainError(w, r, http.StatusNotFound, err)
|
||||
return
|
||||
}
|
||||
if rec == nil || strings.TrimSpace(rec.WorkspaceID) == "" || s.resolver() == nil {
|
||||
writeSafeError(w, r, http.StatusNotFound, "workspace not found")
|
||||
return
|
||||
}
|
||||
dir, err := s.resolver().ResolveStoredWorkDir(rec.WorkspaceID)
|
||||
if err != nil {
|
||||
writeSafeError(w, r, http.StatusNotFound, "workspace not found")
|
||||
return
|
||||
}
|
||||
proj, err := squad.ReadSquadProjection(dir)
|
||||
if err != nil {
|
||||
writeSafeError(w, r, http.StatusNotFound, "squad not found")
|
||||
return
|
||||
}
|
||||
if proj.TaskKey == "" {
|
||||
proj.TaskKey = rec.TaskKey
|
||||
}
|
||||
if proj.Title == "" {
|
||||
proj.Title = rec.TaskKey
|
||||
}
|
||||
writeJSONResponse(w, http.StatusOK, proj)
|
||||
}
|
||||
|
||||
func (s *Server) handleResumeTaskRun(w http.ResponseWriter, r *http.Request) {
|
||||
user, ok := s.requirePrincipal(w, r)
|
||||
if !ok {
|
||||
|
||||
@@ -0,0 +1,406 @@
|
||||
package squad
|
||||
|
||||
import (
|
||||
"fmt"
|
||||
"strings"
|
||||
"sync"
|
||||
"time"
|
||||
|
||||
"chat-agent-system/internal/agent"
|
||||
"chat-agent-system/internal/stream"
|
||||
)
|
||||
|
||||
// 过程流常量：节流窗口 / 单帧上限 / 工具文本截断（设计文档 F.1）。
|
||||
const (
|
||||
processFlushInterval = 200 * time.Millisecond
|
||||
processMaxFrameBytes = 8 * 1024
|
||||
processToolTrimBytes = 4 * 1024
|
||||
)
|
||||
|
||||
// detailCategory 过程事件类别（与 stream.DetailPayload.Category 枚举对齐）。
|
||||
const (
|
||||
detailCategoryAgentThought = "agent_thought"
|
||||
detailCategoryToolCall = "tool_call"
|
||||
)
|
||||
|
||||
// deltaStream 一个聚合项的增量缓冲（节流冲刷）。
|
||||
// 只被 bridge 主路径使用；timer 回调仅触发 flush 任务。
|
||||
type deltaStream struct {
|
||||
mu sync.Mutex
|
||||
itemID string
|
||||
buf strings.Builder
|
||||
timer *time.Timer
|
||||
done bool
|
||||
}
|
||||
|
||||
func (d *deltaStream) append(text string) {
|
||||
d.mu.Lock()
|
||||
defer d.mu.Unlock()
|
||||
if d.done || text == "" {
|
||||
return
|
||||
}
|
||||
d.buf.WriteString(text)
|
||||
}
|
||||
|
||||
// takeIfReady 返回并清空缓冲；由外部节拍调用（非定时器自转）。
|
||||
func (d *deltaStream) take() string {
|
||||
d.mu.Lock()
|
||||
defer d.mu.Unlock()
|
||||
out := d.buf.String()
|
||||
d.buf.Reset()
|
||||
return out
|
||||
}
|
||||
|
||||
func (d *deltaStream) finish() {
|
||||
d.mu.Lock()
|
||||
defer d.mu.Unlock()
|
||||
if d.timer != nil {
|
||||
d.timer.Stop()
|
||||
d.timer = nil
|
||||
}
|
||||
d.done = true
|
||||
}
|
||||
|
||||
// processBridge 把 agent.ProcessEvent 翻译为 stream.EventDetail 增量帧。
|
||||
// 每个 step（小队成员的一次执行）独立持有一个 bridge；无跨 step 共享状态，并行安全。
|
||||
type processBridge struct {
|
||||
streamer *stream.Streamer
|
||||
step StepSpec
|
||||
friendly string // 角色显示名
|
||||
flushTick *time.Ticker
|
||||
stopCh chan struct{}
|
||||
stopOnce sync.Once
|
||||
wg sync.WaitGroup
|
||||
|
||||
mu sync.Mutex
|
||||
thoughtSeq int
|
||||
toolSeq int
|
||||
thought *thoughtBlock // 当前思考块（nil = 尚未开始）
|
||||
currentTool *toolBlock
|
||||
emitted bool // 是否已发出过事件（Stop 只在发过时才有意义）
|
||||
}
|
||||
|
||||
// thoughtBlock 一个 message 周期内的思考聚合项。
|
||||
type thoughtBlock struct {
|
||||
itemID string
|
||||
stream *deltaStream
|
||||
title string
|
||||
}
|
||||
|
||||
// toolBlock 一次工具调用的聚合项。
|
||||
type toolBlock struct {
|
||||
itemID string
|
||||
name string
|
||||
started time.Time
|
||||
input strings.Builder
|
||||
}
|
||||
|
||||
// newProcessBridge 创建过程桥。返回 nil 表示该 step 无流式需求（streamer 缺失），
|
||||
// 调用方对 nil bridge 调用任何方法都是安全 no-op。
|
||||
func newProcessBridge(streamer *stream.Streamer, step StepSpec, friendlyRole string) *processBridge {
|
||||
if streamer == nil {
|
||||
return nil
|
||||
}
|
||||
b := &processBridge{
|
||||
streamer: streamer,
|
||||
step: step,
|
||||
friendly: friendlyRole,
|
||||
stopCh: make(chan struct{}),
|
||||
}
|
||||
// 定时冲刷循环：所有 delta 缓冲统一按 flushInterval 节拍冲刷。
|
||||
b.flushTick = time.NewTicker(processFlushInterval)
|
||||
b.wg.Add(1)
|
||||
go func() {
|
||||
defer b.wg.Done()
|
||||
for {
|
||||
select {
|
||||
case <-b.stopCh:
|
||||
return
|
||||
case <-b.flushTick.C:
|
||||
b.flushPending()
|
||||
}
|
||||
}
|
||||
}()
|
||||
return b
|
||||
}
|
||||
|
||||
// Stop 结束 bridge：冲刷残余、定稿当前项、停止节拍循环。幂等（可安全多次调用）。
|
||||
func (b *processBridge) Stop() {
|
||||
if b == nil {
|
||||
return
|
||||
}
|
||||
b.stopOnce.Do(func() {
|
||||
b.flushPending()
|
||||
b.mu.Lock()
|
||||
th := b.thought
|
||||
tl := b.currentTool
|
||||
b.mu.Unlock()
|
||||
if th != nil {
|
||||
b.finishThought(th)
|
||||
}
|
||||
if tl != nil {
|
||||
b.finishTool(tl, "succeeded", "")
|
||||
}
|
||||
close(b.stopCh)
|
||||
b.flushTick.Stop()
|
||||
b.wg.Wait()
|
||||
})
|
||||
}
|
||||
|
||||
// Handle 把一个 agent.ProcessEvent 翻译为 detail 事件流。
|
||||
// 调用方保证事件到达顺序与 agentx 循环一致（同 step 内串行）。
|
||||
func (b *processBridge) Handle(ev agent.ProcessEvent) {
|
||||
if b == nil {
|
||||
return
|
||||
}
|
||||
select {
|
||||
case <-b.stopCh:
|
||||
return // Stop 后忽略（防御：Execute 返回后的残余事件）
|
||||
default:
|
||||
}
|
||||
switch ev.Type {
|
||||
case "reasoning_delta":
|
||||
b.onReasoningDelta(ev.Text)
|
||||
case "text_delta":
|
||||
// P0 决策（设计文档 C.4）：任务正文不发 detail——终态走既有 step_output
|
||||
// 通道，避免与最终 deliverable 重复推送大文本。
|
||||
case "tool_call_delta":
|
||||
b.onToolCallDelta(ev)
|
||||
case "tool_execution_start":
|
||||
b.onToolStart(ev.Tool)
|
||||
case "tool_execution_end":
|
||||
b.onToolEnd(ev.Tool, ev.Detail)
|
||||
case "message_end":
|
||||
b.onMessageEnd()
|
||||
}
|
||||
}
|
||||
|
||||
func (b *processBridge) onReasoningDelta(text string) {
|
||||
b.mu.Lock()
|
||||
if b.thought == nil {
|
||||
b.thoughtSeq++
|
||||
b.thought = &thoughtBlock{
|
||||
itemID: fmt.Sprintf("thought#%s#%d", b.step.ID, b.thoughtSeq),
|
||||
stream: &deltaStream{},
|
||||
title: fmt.Sprintf("%s思考过程 %d", b.friendly, b.thoughtSeq),
|
||||
}
|
||||
b.emitted = true
|
||||
}
|
||||
th := b.thought
|
||||
b.mu.Unlock()
|
||||
th.stream.append(text)
|
||||
}
|
||||
|
||||
func (b *processBridge) onToolCallDelta(ev agent.ProcessEvent) {
|
||||
b.mu.Lock()
|
||||
t := b.currentTool
|
||||
if t == nil {
|
||||
// 模型在 tool_execution_start 前就吐参数增量：只预建内部块（不发 calling 帧，
|
||||
// 正式的 calling 帧由随后的 tool_execution_start 发出）。
|
||||
b.toolSeq++
|
||||
t = &toolBlock{
|
||||
itemID: fmt.Sprintf("tool#%s#%d", b.step.ID, b.toolSeq),
|
||||
name: ev.Tool,
|
||||
started: time.Now(),
|
||||
}
|
||||
b.currentTool = t
|
||||
b.emitted = true
|
||||
}
|
||||
b.mu.Unlock()
|
||||
t.input.WriteString(ev.Detail)
|
||||
}
|
||||
|
||||
func (b *processBridge) onToolStart(name string) {
|
||||
b.mu.Lock()
|
||||
var cur *toolBlock
|
||||
if b.currentTool != nil {
|
||||
// 已由 tool_call_delta 预建（罕见时序）：复用同一聚合项，只补名与时间。
|
||||
cur = b.currentTool
|
||||
if cur.name == "" {
|
||||
cur.name = name
|
||||
}
|
||||
cur.started = time.Now()
|
||||
} else {
|
||||
b.toolSeq++
|
||||
cur = &toolBlock{
|
||||
itemID: fmt.Sprintf("tool#%s#%d", b.step.ID, b.toolSeq),
|
||||
name: name,
|
||||
started: time.Now(),
|
||||
}
|
||||
b.currentTool = cur
|
||||
}
|
||||
b.emitted = true
|
||||
itemID, toolName := cur.itemID, cur.name
|
||||
b.mu.Unlock()
|
||||
// tool_execution_start 是正式的"开始调用"信号：无论是否预建，都要发 calling 帧。
|
||||
b.emitDetail(&stream.DetailPayload{
|
||||
Category: detailCategoryToolCall,
|
||||
StepID: b.step.ID,
|
||||
Role: b.step.Role,
|
||||
ItemID: itemID,
|
||||
Tool: &stream.ToolCallInfo{Name: toolName, Status: "calling"},
|
||||
Title: fmt.Sprintf("%s调用工具 %s", b.friendly, toolName),
|
||||
Collapsible: true,
|
||||
})
|
||||
}
|
||||
|
||||
func (b *processBridge) onToolEnd(name, result string) {
|
||||
b.mu.Lock()
|
||||
t := b.currentTool
|
||||
b.currentTool = nil
|
||||
b.mu.Unlock()
|
||||
if t == nil {
|
||||
return
|
||||
}
|
||||
status := "succeeded"
|
||||
if strings.Contains(strings.ToLower(result), "error") || strings.Contains(strings.ToLower(result), "failed") {
|
||||
status = "failed"
|
||||
}
|
||||
duration := time.Since(t.started).Milliseconds()
|
||||
info := &stream.ToolCallInfo{
|
||||
Name: name,
|
||||
Status: status,
|
||||
Input: trimProcessText(t.input.String()),
|
||||
Output: trimProcessText(result),
|
||||
DurationMs: duration,
|
||||
}
|
||||
b.emitDetail(&stream.DetailPayload{
|
||||
Category: detailCategoryToolCall,
|
||||
StepID: b.step.ID,
|
||||
Role: b.step.Role,
|
||||
ItemID: t.itemID,
|
||||
Tool: info,
|
||||
Collapsible: true,
|
||||
})
|
||||
// 工具项即时定稿（状态帧 + done 标记合并语义：终帧即最后状态）
|
||||
b.emitDetail(&stream.DetailPayload{
|
||||
Category: detailCategoryToolCall,
|
||||
StepID: b.step.ID,
|
||||
Role: b.step.Role,
|
||||
ItemID: t.itemID,
|
||||
Tool: info,
|
||||
Done: true,
|
||||
Collapsible: true,
|
||||
})
|
||||
}
|
||||
|
||||
func (b *processBridge) onMessageEnd() {
|
||||
b.mu.Lock()
|
||||
th := b.thought
|
||||
b.thought = nil
|
||||
b.mu.Unlock()
|
||||
if th != nil {
|
||||
b.finishThought(th)
|
||||
}
|
||||
}
|
||||
|
||||
// flushPending 冲刷当前思考块缓冲（节拍或 Stop 调用）。
|
||||
func (b *processBridge) flushPending() {
|
||||
if b == nil {
|
||||
return
|
||||
}
|
||||
b.mu.Lock()
|
||||
th := b.thought
|
||||
b.mu.Unlock()
|
||||
if th == nil {
|
||||
return
|
||||
}
|
||||
text := th.stream.take()
|
||||
if text == "" {
|
||||
return
|
||||
}
|
||||
if len(text) > processMaxFrameBytes {
|
||||
text = text[:processMaxFrameBytes] + "\n…(截断)"
|
||||
}
|
||||
b.emitDetail(&stream.DetailPayload{
|
||||
Category: detailCategoryAgentThought,
|
||||
StepID: b.step.ID,
|
||||
Role: b.step.Role,
|
||||
ItemID: th.itemID,
|
||||
Delta: text,
|
||||
Collapsible: true,
|
||||
})
|
||||
}
|
||||
|
||||
func (b *processBridge) finishThought(th *thoughtBlock) {
|
||||
th.stream.finish()
|
||||
// 冲残余
|
||||
if text := th.stream.take(); text != "" {
|
||||
b.emitDetail(&stream.DetailPayload{
|
||||
Category: detailCategoryAgentThought,
|
||||
StepID: b.step.ID,
|
||||
Role: b.step.Role,
|
||||
ItemID: th.itemID,
|
||||
Delta: text,
|
||||
Collapsible: true,
|
||||
})
|
||||
}
|
||||
b.emitDetail(&stream.DetailPayload{
|
||||
Category: detailCategoryAgentThought,
|
||||
StepID: b.step.ID,
|
||||
Role: b.step.Role,
|
||||
ItemID: th.itemID,
|
||||
Title: th.title,
|
||||
Done: true,
|
||||
Collapsible: true,
|
||||
})
|
||||
}
|
||||
|
||||
func (b *processBridge) finishTool(t *toolBlock, status, result string) {
|
||||
b.emitDetail(&stream.DetailPayload{
|
||||
Category: detailCategoryToolCall,
|
||||
StepID: b.step.ID,
|
||||
Role: b.step.Role,
|
||||
ItemID: t.itemID,
|
||||
Tool: &stream.ToolCallInfo{
|
||||
Name: t.name,
|
||||
Status: status,
|
||||
Input: trimProcessText(t.input.String()),
|
||||
Output: trimProcessText(result),
|
||||
DurationMs: time.Since(t.started).Milliseconds(),
|
||||
},
|
||||
Done: true,
|
||||
Collapsible: true,
|
||||
})
|
||||
}
|
||||
|
||||
func (b *processBridge) emitDetail(p *stream.DetailPayload) {
|
||||
if b == nil || b.streamer == nil {
|
||||
return
|
||||
}
|
||||
b.streamer.Emit(stream.StreamEvent{Type: stream.EventDetail, Detail: p})
|
||||
}
|
||||
|
||||
// trimProcessText 工具入参/结果截断（防超大帧；保留首尾）。
|
||||
func trimProcessText(s string) string {
|
||||
if len(s) <= processToolTrimBytes {
|
||||
return s
|
||||
}
|
||||
headN := processToolTrimBytes / 2
|
||||
tailN := processToolTrimBytes / 4
|
||||
head := s[:headN]
|
||||
tail := s[len(s)-tailN:]
|
||||
return head + "\n…(中间省略 " + fmt.Sprint(len(s)-headN-tailN) + " 字节)…\n" + tail
|
||||
}
|
||||
|
||||
// AttachProcessBridge 给 agent.Request 装配过程桥（若 streamer 缺失或无流式需求则原样返回）。
|
||||
// 返回的 stop 函数在 Execute 返回后必须调用一次（定稿与冲刷残余）。
|
||||
func AttachProcessBridge(streamer *stream.Streamer, step StepSpec, req *agent.Request) (stop func()) {
|
||||
if streamer == nil || req == nil {
|
||||
return func() {}
|
||||
}
|
||||
bridge := newProcessBridge(streamer, step, FriendlyRoleName(step.Role))
|
||||
if bridge == nil {
|
||||
return func() {}
|
||||
}
|
||||
prev := req.OnEvent
|
||||
if prev == nil {
|
||||
req.OnEvent = bridge.Handle
|
||||
} else {
|
||||
req.OnEvent = func(ev agent.ProcessEvent) {
|
||||
prev(ev)
|
||||
bridge.Handle(ev)
|
||||
}
|
||||
}
|
||||
return bridge.Stop
|
||||
}
|
||||
@@ -0,0 +1,215 @@
|
||||
package squad
|
||||
|
||||
import (
|
||||
"strings"
|
||||
"sync"
|
||||
"testing"
|
||||
"time"
|
||||
|
||||
"chat-agent-system/internal/agent"
|
||||
"chat-agent-system/internal/stream"
|
||||
)
|
||||
|
||||
func TestProcessBridgeNilStreamerIsNoop(t *testing.T) {
|
||||
step := StepSpec{ID: "step_01", Role: "regulatory"}
|
||||
req := &agent.Request{}
|
||||
stop := AttachProcessBridge(nil, step, req)
|
||||
if req.OnEvent != nil {
|
||||
t.Fatal("nil streamer must not attach OnEvent")
|
||||
}
|
||||
req.OnEvent = func(agent.ProcessEvent) {} // 直接赋值不应有 panic
|
||||
stop() // no-op stop 不 panic
|
||||
}
|
||||
|
||||
func TestProcessBridgeReasoningFlow(t *testing.T) {
|
||||
var mu sync.Mutex
|
||||
var events []stream.StreamEvent
|
||||
streamer := stream.NewStreamer(true, func(ev stream.StreamEvent) {
|
||||
mu.Lock()
|
||||
events = append(events, ev)
|
||||
mu.Unlock()
|
||||
})
|
||||
step := StepSpec{ID: "step_02", Role: "engineer"}
|
||||
req := &agent.Request{}
|
||||
stop := AttachProcessBridge(streamer, step, req)
|
||||
defer stop()
|
||||
|
||||
req.OnEvent(agent.ProcessEvent{Type: "reasoning_delta", Text: "第一段思考"})
|
||||
req.OnEvent(agent.ProcessEvent{Type: "reasoning_delta", Text: "继续推理"})
|
||||
req.OnEvent(agent.ProcessEvent{Type: "message_end"})
|
||||
stop()
|
||||
// 再发事件（stop 后）不应 panic 且不应再发帧
|
||||
req.OnEvent(agent.ProcessEvent{Type: "reasoning_delta", Text: "ignored"})
|
||||
|
||||
mu.Lock()
|
||||
defer mu.Unlock()
|
||||
var deltas, dones int
|
||||
var joined strings.Builder
|
||||
for _, ev := range events {
|
||||
if ev.Type != stream.EventDetail || ev.Detail == nil {
|
||||
continue
|
||||
}
|
||||
if ev.Detail.Category != detailCategoryAgentThought {
|
||||
t.Fatalf("unexpected category %s", ev.Detail.Category)
|
||||
}
|
||||
if ev.Detail.Delta != "" {
|
||||
deltas++
|
||||
joined.WriteString(ev.Detail.Delta)
|
||||
}
|
||||
if ev.Detail.Done {
|
||||
dones++
|
||||
}
|
||||
if ev.Detail.StepID != step.ID || ev.Detail.Role != step.Role {
|
||||
t.Fatalf("step identity missing: %+v", ev.Detail)
|
||||
}
|
||||
}
|
||||
if joined.String() != "第一段思考继续推理" {
|
||||
t.Fatalf("joined delta %q", joined.String())
|
||||
}
|
||||
if dones != 1 {
|
||||
t.Fatalf("expected exactly 1 done frame, got %d", dones)
|
||||
}
|
||||
if deltas < 1 {
|
||||
t.Fatal("expected at least one delta frame")
|
||||
}
|
||||
}
|
||||
|
||||
func TestProcessBridgeToolFlow(t *testing.T) {
|
||||
var mu sync.Mutex
|
||||
var events []stream.StreamEvent
|
||||
streamer := stream.NewStreamer(true, func(ev stream.StreamEvent) {
|
||||
mu.Lock()
|
||||
events = append(events, ev)
|
||||
mu.Unlock()
|
||||
})
|
||||
step := StepSpec{ID: "step_03", Role: "writer"}
|
||||
b := newProcessBridge(streamer, step, "撰写专员")
|
||||
if b == nil {
|
||||
t.Fatal("bridge expected")
|
||||
}
|
||||
|
||||
b.Handle(agent.ProcessEvent{Type: "tool_call_delta", Tool: "read_file", Detail: `{"path":"a.md"`})
|
||||
b.Handle(agent.ProcessEvent{Type: "tool_execution_start", Tool: "read_file"})
|
||||
b.Handle(agent.ProcessEvent{Type: "tool_call_delta", Tool: "read_file", Detail: `}`})
|
||||
b.Handle(agent.ProcessEvent{Type: "tool_execution_end", Tool: "read_file", Detail: "file content ..."})
|
||||
b.Stop()
|
||||
|
||||
mu.Lock()
|
||||
defer mu.Unlock()
|
||||
var calling, done, withInput int
|
||||
for _, ev := range events {
|
||||
if ev.Type != stream.EventDetail || ev.Detail == nil || ev.Detail.Category != detailCategoryToolCall {
|
||||
continue
|
||||
}
|
||||
if ev.Detail.Tool == nil {
|
||||
t.Fatal("tool_call frame must carry Tool info")
|
||||
}
|
||||
switch {
|
||||
case ev.Detail.Tool.Status == "calling":
|
||||
calling++
|
||||
case ev.Detail.Done:
|
||||
done++
|
||||
}
|
||||
if strings.Contains(ev.Detail.Tool.Input, "a.md") {
|
||||
withInput++
|
||||
}
|
||||
}
|
||||
if calling != 1 {
|
||||
t.Fatalf("calling frames %d", calling)
|
||||
}
|
||||
if done < 1 {
|
||||
t.Fatal("expected done frame for tool item")
|
||||
}
|
||||
if withInput == 0 {
|
||||
t.Fatal("tool input aggregate lost")
|
||||
}
|
||||
}
|
||||
|
||||
func TestProcessBridgeToolFailureStatus(t *testing.T) {
|
||||
var mu sync.Mutex
|
||||
var events []stream.StreamEvent
|
||||
streamer := stream.NewStreamer(true, func(ev stream.StreamEvent) {
|
||||
mu.Lock()
|
||||
events = append(events, ev)
|
||||
mu.Unlock()
|
||||
})
|
||||
b := newProcessBridge(streamer, StepSpec{ID: "s", Role: "r"}, "测试")
|
||||
b.Handle(agent.ProcessEvent{Type: "tool_execution_start", Tool: "write_file"})
|
||||
b.Handle(agent.ProcessEvent{Type: "tool_execution_end", Tool: "write_file", Detail: "error: disk full"})
|
||||
b.Stop()
|
||||
|
||||
mu.Lock()
|
||||
defer mu.Unlock()
|
||||
sawFailed := false
|
||||
for _, ev := range events {
|
||||
if ev.Detail != nil && ev.Detail.Tool != nil && ev.Detail.Tool.Status == "failed" {
|
||||
sawFailed = true
|
||||
}
|
||||
}
|
||||
if !sawFailed {
|
||||
t.Fatal("error result must map to failed status")
|
||||
}
|
||||
}
|
||||
|
||||
func TestTrimProcessText(t *testing.T) {
|
||||
small := strings.Repeat("a", 100)
|
||||
if got := trimProcessText(small); got != small {
|
||||
t.Fatal("small text must pass through")
|
||||
}
|
||||
big := strings.Repeat("b", processToolTrimBytes*4)
|
||||
got := trimProcessText(big)
|
||||
if len(got) >= len(big) {
|
||||
t.Fatal("big text must be trimmed")
|
||||
}
|
||||
if !strings.Contains(got, "…(中间省略") {
|
||||
t.Fatal("trim marker missing")
|
||||
}
|
||||
}
|
||||
|
||||
func TestProcessBridgeConcurrentSteps(t *testing.T) {
|
||||
// 并行组成员各自持有独立 bridge，事件互不串扰（设计 F.4）。
|
||||
var mu sync.Mutex
|
||||
seen := map[string]string{} // itemID -> role
|
||||
streamer := stream.NewStreamer(true, func(ev stream.StreamEvent) {
|
||||
if ev.Type != stream.EventDetail || ev.Detail == nil {
|
||||
return
|
||||
}
|
||||
mu.Lock()
|
||||
seen[ev.Detail.ItemID] = ev.Detail.Role
|
||||
mu.Unlock()
|
||||
})
|
||||
b1 := newProcessBridge(streamer, StepSpec{ID: "p1", Role: "regulatory"}, "法规专员")
|
||||
b2 := newProcessBridge(streamer, StepSpec{ID: "p2", Role: "market"}, "市场专员")
|
||||
var wg sync.WaitGroup
|
||||
wg.Add(2)
|
||||
go func() {
|
||||
defer wg.Done()
|
||||
for i := 0; i < 20; i++ {
|
||||
b1.Handle(agent.ProcessEvent{Type: "reasoning_delta", Text: "r1"})
|
||||
}
|
||||
b1.Stop()
|
||||
}()
|
||||
go func() {
|
||||
defer wg.Done()
|
||||
for i := 0; i < 20; i++ {
|
||||
b2.Handle(agent.ProcessEvent{Type: "reasoning_delta", Text: "r2"})
|
||||
}
|
||||
b2.Stop()
|
||||
}()
|
||||
wg.Wait()
|
||||
time.Sleep(50 * time.Millisecond) // 等冲刷
|
||||
mu.Lock()
|
||||
defer mu.Unlock()
|
||||
for itemID, role := range seen {
|
||||
if !strings.Contains(itemID, "p1") && !strings.Contains(itemID, "p2") {
|
||||
t.Fatalf("unexpected item %s", itemID)
|
||||
}
|
||||
want := "regulatory"
|
||||
if strings.Contains(itemID, "p2") {
|
||||
want = "market"
|
||||
}
|
||||
if role != want {
|
||||
t.Fatalf("item %s mapped to role %s, want %s", itemID, role, want)
|
||||
}
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,60 @@
|
||||
package squad
|
||||
|
||||
import (
|
||||
"encoding/json"
|
||||
"os"
|
||||
"path/filepath"
|
||||
)
|
||||
|
||||
// SquadMemberView 是子页面左栏的只读编制投影（不含模型名与工具权限）。
|
||||
type SquadMemberView struct {
|
||||
Role string `json:"role"`
|
||||
Name string `json:"name"`
|
||||
Responsibilities []string `json:"responsibilities"`
|
||||
Source string `json:"source,omitempty"`
|
||||
}
|
||||
|
||||
// SquadProjection GET /tasks/runs/{id}/squad 的响应体。
|
||||
type SquadProjection struct {
|
||||
Members []SquadMemberView `json:"members"`
|
||||
TaskKey string `json:"task_key"`
|
||||
Title string `json:"title"`
|
||||
}
|
||||
|
||||
// ReadSquadProjection 读取工作区 squad.json（及可选 workspace.json 标题）。
|
||||
// 缺少 squad.json 时返回错误，由调用方决定 404。
|
||||
func ReadSquadProjection(workDir string) (SquadProjection, error) {
|
||||
data, err := os.ReadFile(filepath.Join(workDir, "squad.json"))
|
||||
if err != nil {
|
||||
return SquadProjection{}, err
|
||||
}
|
||||
var cfg SquadConfig
|
||||
if err := json.Unmarshal(data, &cfg); err != nil {
|
||||
return SquadProjection{}, err
|
||||
}
|
||||
members := make([]SquadMemberView, 0, len(cfg.Agents))
|
||||
for _, agent := range cfg.Agents {
|
||||
resp := agent.Responsibilities
|
||||
if resp == nil {
|
||||
resp = []string{}
|
||||
}
|
||||
members = append(members, SquadMemberView{
|
||||
Role: agent.Role,
|
||||
Name: agent.Name,
|
||||
Responsibilities: resp,
|
||||
Source: string(agent.Source),
|
||||
})
|
||||
}
|
||||
title := cfg.SquadName
|
||||
taskKey := ""
|
||||
if metaRaw, err := os.ReadFile(filepath.Join(workDir, "workspace.json")); err == nil {
|
||||
var meta WorkspaceMetadata
|
||||
if json.Unmarshal(metaRaw, &meta) == nil {
|
||||
if meta.StageTitle != "" {
|
||||
title = meta.StageTitle
|
||||
}
|
||||
taskKey = meta.TaskKey
|
||||
}
|
||||
}
|
||||
return SquadProjection{Members: members, TaskKey: taskKey, Title: title}, nil
|
||||
}
|
||||
@@ -0,0 +1,38 @@
|
||||
package squad
|
||||
|
||||
import (
|
||||
"os"
|
||||
"path/filepath"
|
||||
"testing"
|
||||
)
|
||||
|
||||
func TestReadSquadProjection(t *testing.T) {
|
||||
dir := t.TempDir()
|
||||
squadJSON := `{
|
||||
"squad_name": "立项小队",
|
||||
"version": "1",
|
||||
"agents": [
|
||||
{"role": "regulatory", "name": "法规专员", "source": "standard", "responsibilities": ["筛法规"]}
|
||||
]
|
||||
}`
|
||||
if err := os.WriteFile(filepath.Join(dir, "squad.json"), []byte(squadJSON), 0o644); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
ws := `{"task_key":"intake","stage_title":"用户需求收集"}`
|
||||
if err := os.WriteFile(filepath.Join(dir, "workspace.json"), []byte(ws), 0o644); err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
got, err := ReadSquadProjection(dir)
|
||||
if err != nil {
|
||||
t.Fatal(err)
|
||||
}
|
||||
if got.TaskKey != "intake" || got.Title != "用户需求收集" {
|
||||
t.Fatalf("meta: %+v", got)
|
||||
}
|
||||
if len(got.Members) != 1 || got.Members[0].Role != "regulatory" || got.Members[0].Name != "法规专员" {
|
||||
t.Fatalf("members: %+v", got.Members)
|
||||
}
|
||||
if _, err := ReadSquadProjection(t.TempDir()); err == nil {
|
||||
t.Fatal("missing squad.json must error")
|
||||
}
|
||||
}
|
||||
@@ -551,6 +551,9 @@ func (d *WorkflowDriver) runOneStep(
|
||||
Targets: targets,
|
||||
Timeout: stepTimeout,
|
||||
}
|
||||
// 过程流桥：把成员执行的思考/工具调用增量翻译为 detail 事件（streamer 缺失时为 no-op）。
|
||||
// 注意 retry 循环内多次 Execute 共用一个 bridge 语义（同 step 聚合），Stop 只在首次调用。
|
||||
stopProcessBridge := AttachProcessBridge(streamer, step, &req)
|
||||
|
||||
var (
|
||||
res *agent.Result
|
||||
@@ -561,6 +564,7 @@ func (d *WorkflowDriver) runOneStep(
|
||||
lastAssess agent.FailureAssessment
|
||||
basePrompt = string(taskData)
|
||||
)
|
||||
defer stopProcessBridge()
|
||||
for attempt := 1; attempt <= agent.MaxTransientStepInvocations; attempt++ {
|
||||
onDisk, _ := os.ReadFile(prevOutPath)
|
||||
wroteThisInvocation := deliverableWroteThisInvocation(prevContent, onDisk)
|
||||
|
||||
@@ -69,9 +69,25 @@ type DetailPayload struct {
|
||||
Category string `json:"category"` // agent_thought, step_output, tool_call, log
|
||||
StepID string `json:"step_id,omitempty"` // 步骤标识 (如 tasks/01_regulatory)
|
||||
Role string `json:"role,omitempty"` // 对应角色
|
||||
Title string `json:"title"` // 详情标题
|
||||
Title string `json:"title"`
|
||||
Content string `json:"content"` // 完整推演正文或技术草稿
|
||||
Collapsible bool `json:"collapsible"` // 客户端是否应当默认折叠显示
|
||||
// ---- 过程流扩展（2026-09-28 subagent 过程展示设计 B.2）----
|
||||
ItemID string `json:"item_id,omitempty"` // 聚合项 ID：一个思考块/一次工具调用
|
||||
// 增量模式：Delta 追加到 ItemID 项；Content 为快照对账帧；Done 定稿。
|
||||
Delta string `json:"delta,omitempty"`
|
||||
Done bool `json:"done,omitempty"`
|
||||
// ToolCallInfo 工具调用状态帧（Category == "tool_call" 时）。
|
||||
Tool *ToolCallInfo `json:"tool,omitempty"`
|
||||
}
|
||||
|
||||
// ToolCallInfo 一次工具调用的状态投影（thinking 流式展示用）。
|
||||
type ToolCallInfo struct {
|
||||
Name string `json:"name"` // write_file / read_file / ...
|
||||
Status string `json:"status"` // calling | succeeded | failed
|
||||
Input string `json:"input,omitempty"` // 入参摘要（截断）
|
||||
Output string `json:"output,omitempty"` // 结果摘要（截断）
|
||||
DurationMs int64 `json:"duration_ms,omitempty"`
|
||||
}
|
||||
|
||||
// DeliverablePayload A轨正式合规文档交付载荷
|
||||
|
||||
@@ -476,7 +476,10 @@ func (m *Manager) runWorker(ctx context.Context, ar *activeRun, taskReq bizdev.T
|
||||
}
|
||||
m.appendEvent(ar, ev)
|
||||
}
|
||||
ctxWithS, streamer := stream.WithListener(ctx, false, listener)
|
||||
// includeDetails=true：任务通道的 detail（思考链/工具调用过程）全量持久化并推送
|
||||
// （2026-09-28 过程流式设计 C.6——"刷新后过程消失"违背透明可追溯原则；
|
||||
// 认知负荷由前端默认折叠保证，旧前端忽略 detail 字段无破坏）。
|
||||
ctxWithS, streamer := stream.WithListener(ctx, true, listener)
|
||||
|
||||
var taskResult *bizdev.TaskRunResult
|
||||
var execErr error
|
||||
|
||||
+42
-2
@@ -112,7 +112,7 @@ import { RequestEpoch, shareInflight } from './data/initFanout';
|
||||
import {
|
||||
resolveSessionProjectId,
|
||||
} from './lib/sessionAdapter';
|
||||
import { navigateToAnchor, parseAnchorHash } from './engine/anchorNavigator';
|
||||
import { navigateToAnchor, parseAnchorHash, parseSubagentsHash, type SubagentsTarget } from './engine/anchorNavigator';
|
||||
import { useUnreadMarks } from './hooks/useUnreadMarks';
|
||||
import { fetchChatCategories, fetchChatDates } from './api/chatWindow';
|
||||
import { ChatCategoryStat, ChatDateStat, QuotedMessage } from './types/workbench';
|
||||
@@ -130,6 +130,8 @@ export {
|
||||
|
||||
const loadAdminConsoleView = () =>
|
||||
import('./components/admin/console/AdminConsoleView').then((m) => ({ default: m.AdminConsoleView }));
|
||||
const loadSubagentsView = () =>
|
||||
import('./components/subagents/SubagentsView').then((m) => ({ default: m.SubagentsView }));
|
||||
export default function App() {
|
||||
// Wave 2：启动时幂等剔除 localStorage 中的 demo seed 残留（须先于 load*）
|
||||
purgeDemoSeedsFromLocalStorage();
|
||||
@@ -324,7 +326,8 @@ export default function App() {
|
||||
const [rightsModalOpen, setRightsModalOpen] = useState(false);
|
||||
const [userSettingsOpen, setUserSettingsOpen] = useState(false);
|
||||
// 顶层视图切换：单屏工作台 / 全屏系统管理台（仅管理员可进入）
|
||||
const [view, setView] = useState<'workbench' | 'admin'>('workbench');
|
||||
const [view, setView] = useState<'workbench' | 'admin' | 'subagents'>('workbench');
|
||||
const [subagentsTarget, setSubagentsTarget] = useState<SubagentsTarget | null>(null);
|
||||
const [docVisibilityModalOpen, setDocVisibilityModalOpen] = useState(false);
|
||||
const [targetDocForVisibility, setTargetDocForVisibility] = useState<DeliverableItem | null>(null);
|
||||
|
||||
@@ -1134,6 +1137,16 @@ export default function App() {
|
||||
useEffect(() => {
|
||||
const handleHashChange = () => {
|
||||
if (sessionValidatedRef.current !== true) return;
|
||||
const sub = parseSubagentsHash(window.location.hash);
|
||||
if (sub) {
|
||||
setSubagentsTarget(sub);
|
||||
setView('subagents');
|
||||
if (sub.sessionId && sub.sessionId !== activeSessionIdRef.current) {
|
||||
setActiveSessionId(sub.sessionId);
|
||||
}
|
||||
return;
|
||||
}
|
||||
setView((current) => (current === 'subagents' ? 'workbench' : current));
|
||||
const parsed = parseAnchorHash(window.location.hash);
|
||||
if (!parsed) return;
|
||||
const currentActive = activeSessionIdRef.current;
|
||||
@@ -1161,6 +1174,14 @@ export default function App() {
|
||||
return () => window.removeEventListener('hashchange', handleHashChange);
|
||||
}, []); // 仅在挂载时监听 hashchange，避免 activeSessionId 变更时因残留 hash 产生回弹
|
||||
|
||||
useEffect(() => {
|
||||
if (sessionValidated !== true) return;
|
||||
const sub = parseSubagentsHash(window.location.hash);
|
||||
if (!sub) return;
|
||||
setSubagentsTarget(sub);
|
||||
setView('subagents');
|
||||
}, [sessionValidated]);
|
||||
|
||||
// 引用消息交互
|
||||
const handleQuoteMessage = (message: ChatMessage) => {
|
||||
let msgId = message.serverId;
|
||||
@@ -1591,6 +1612,25 @@ export default function App() {
|
||||
);
|
||||
}
|
||||
|
||||
if (view === 'subagents' && subagentsTarget) {
|
||||
return (
|
||||
<OnDemand
|
||||
label="小队过程"
|
||||
load={loadSubagentsView}
|
||||
componentProps={{
|
||||
runId: subagentsTarget.runId,
|
||||
sessionId: subagentsTarget.sessionId,
|
||||
roleId: subagentsTarget.roleId,
|
||||
onExit: () => {
|
||||
setView('workbench');
|
||||
const next = `${window.location.pathname}${window.location.search}`;
|
||||
window.history.replaceState(null, '', next);
|
||||
},
|
||||
}}
|
||||
/>
|
||||
);
|
||||
}
|
||||
|
||||
// 全屏系统管理台：与工作台并列的顶层视图，不经过 AppShell
|
||||
if (view === 'admin') {
|
||||
return (
|
||||
|
||||
@@ -18,6 +18,24 @@ export type StreamEvent = {
|
||||
status?: { message?: string; intent?: string };
|
||||
delta?: string;
|
||||
error?: string;
|
||||
detail?: {
|
||||
category?: string;
|
||||
step_id?: string;
|
||||
role?: string;
|
||||
title?: string;
|
||||
content?: string;
|
||||
collapsible?: boolean;
|
||||
item_id?: string;
|
||||
delta?: string;
|
||||
done?: boolean;
|
||||
tool?: {
|
||||
name?: string;
|
||||
status?: string;
|
||||
input?: string;
|
||||
output?: string;
|
||||
duration_ms?: number;
|
||||
};
|
||||
};
|
||||
deliverable?: {
|
||||
task_key?: string;
|
||||
title?: string;
|
||||
|
||||
@@ -215,3 +215,26 @@ export async function getTaskRun(runId: string) {
|
||||
}
|
||||
return res.json();
|
||||
}
|
||||
|
||||
export type SquadMemberDTO = {
|
||||
role: string;
|
||||
name: string;
|
||||
responsibilities: string[];
|
||||
source?: string;
|
||||
};
|
||||
|
||||
export type SquadProjection = {
|
||||
members: SquadMemberDTO[];
|
||||
task_key: string;
|
||||
title: string;
|
||||
};
|
||||
|
||||
export async function getTaskRunSquad(runId: string): Promise<SquadProjection> {
|
||||
const res = await apiFetch(`/api/v1/tasks/runs/${encodeURIComponent(runId)}/squad`, {
|
||||
headers: getBearerAuthHeaders(),
|
||||
});
|
||||
if (!res.ok) {
|
||||
throw new Error(`get task run squad failed: HTTP ${res.status}`);
|
||||
}
|
||||
return res.json();
|
||||
}
|
||||
|
||||
@@ -9,6 +9,8 @@ import {
|
||||
runningPhaseDetail,
|
||||
type TaskRunCardBadgeTone,
|
||||
} from '../../lib/taskRunCard';
|
||||
import { subagentsHash } from '../../engine/anchorNavigator';
|
||||
import { summarizeProcess } from '../../engine/processEvent';
|
||||
|
||||
export interface TaskRunCardProps {
|
||||
message: ChatMessage;
|
||||
@@ -274,6 +276,33 @@ export const TaskRunCard: React.FC<TaskRunCardProps> = ({ message }) => {
|
||||
<span style={{ wordBreak: 'break-word' }}>{artifactBits.join(' · ')}</span>
|
||||
</div>
|
||||
)}
|
||||
|
||||
{message.taskRunId && message.sessionId && (
|
||||
<button
|
||||
type="button"
|
||||
data-testid="task-run-card-subagents"
|
||||
onClick={() => {
|
||||
window.location.hash = subagentsHash({
|
||||
sessionId: message.sessionId,
|
||||
runId: message.taskRunId || '',
|
||||
});
|
||||
}}
|
||||
style={{
|
||||
display: 'block',
|
||||
width: '100%',
|
||||
textAlign: 'left',
|
||||
border: 0,
|
||||
borderTop: '1px solid var(--border)',
|
||||
background: 'transparent',
|
||||
padding: '8px 12px',
|
||||
fontSize: '12px',
|
||||
color: 'var(--brand-500)',
|
||||
cursor: 'pointer',
|
||||
}}
|
||||
>
|
||||
{isLive ? '查看小队过程 →' : summarizeProcess(message.processItems) || '查看小队过程 →'}
|
||||
</button>
|
||||
)}
|
||||
</div>
|
||||
);
|
||||
};
|
||||
|
||||
@@ -43,7 +43,7 @@ export interface WorkbenchViewProps {
|
||||
currentDocRole: UserRole;
|
||||
setRightsModalOpen: Dispatch<SetStateAction<boolean>>;
|
||||
setUserSettingsOpen: Dispatch<SetStateAction<boolean>>;
|
||||
setView: (view: 'workbench' | 'admin') => void;
|
||||
setView: (view: 'workbench' | 'admin' | 'subagents') => void;
|
||||
setProjectMembersModalOpen: Dispatch<SetStateAction<boolean>>;
|
||||
currentWfState: WorkbenchWorkflowState;
|
||||
workflowGovernanceEnabled: boolean;
|
||||
|
||||
@@ -0,0 +1,32 @@
|
||||
import { useState, type ReactNode } from 'react';
|
||||
|
||||
export function CollapsibleBlock(props: {
|
||||
summary: string;
|
||||
children: ReactNode;
|
||||
defaultOpen?: boolean;
|
||||
testId?: string;
|
||||
}) {
|
||||
const [open, setOpen] = useState(props.defaultOpen ?? false);
|
||||
return (
|
||||
<div data-testid={props.testId} style={{ border: '1px solid var(--border)', borderRadius: 8, marginBottom: 8 }}>
|
||||
<button
|
||||
type="button"
|
||||
aria-expanded={open}
|
||||
onClick={() => setOpen((value) => !value)}
|
||||
style={{
|
||||
width: '100%',
|
||||
textAlign: 'left',
|
||||
background: 'transparent',
|
||||
border: 0,
|
||||
padding: '8px 10px',
|
||||
cursor: 'pointer',
|
||||
fontSize: 13,
|
||||
color: 'var(--text)',
|
||||
}}
|
||||
>
|
||||
{props.summary}
|
||||
</button>
|
||||
{open ? <div style={{ padding: '0 10px 10px', fontSize: 13 }}>{props.children}</div> : null}
|
||||
</div>
|
||||
);
|
||||
}
|
||||
@@ -0,0 +1,12 @@
|
||||
import type { ProcessItem } from '../../engine/processEvent';
|
||||
import { renderBlockMarkdown } from '../../engine/renderBlockMarkdown';
|
||||
import { CollapsibleBlock } from './CollapsibleBlock';
|
||||
|
||||
export function ProcessStepOutput({ item }: { item: ProcessItem }) {
|
||||
const title = item.title || '步骤产出';
|
||||
return (
|
||||
<CollapsibleBlock testId="process-step-output" summary={`产出 · ${title}`}>
|
||||
<div dangerouslySetInnerHTML={{ __html: renderBlockMarkdown(item.text) }} />
|
||||
</CollapsibleBlock>
|
||||
);
|
||||
}
|
||||
@@ -0,0 +1,49 @@
|
||||
import assert from 'node:assert/strict';
|
||||
import test from 'node:test';
|
||||
import React from 'react';
|
||||
import { renderToStaticMarkup } from 'react-dom/server';
|
||||
import { SubagentsLayout } from './SubagentsLayout';
|
||||
|
||||
test('SubagentsLayout shows the running member and folded thought summary', () => {
|
||||
const html = renderToStaticMarkup(
|
||||
React.createElement(SubagentsLayout, {
|
||||
title: '用户需求收集',
|
||||
runId: 'run-abcdef123456',
|
||||
members: [
|
||||
{ role: 'regulatory', name: '法规专员', responsibilities: ['筛法规'] },
|
||||
{ role: 'writer', name: '撰写专员', responsibilities: ['成稿'] },
|
||||
],
|
||||
phases: [
|
||||
{ step: 'step_executing#2', icon: '⚙️', label: '撰写中', status: 'running', role: 'writer' },
|
||||
],
|
||||
items: [
|
||||
{
|
||||
key: 'tasks/02/thought-1',
|
||||
category: 'agent_thought',
|
||||
role: 'writer',
|
||||
stepId: 'tasks/02',
|
||||
text: '先核对素材再写用户需求表的适用范围说明',
|
||||
status: 'done',
|
||||
updatedAt: 1,
|
||||
},
|
||||
{
|
||||
key: 'tasks/01/thought-9',
|
||||
category: 'agent_thought',
|
||||
role: 'regulatory',
|
||||
stepId: 'tasks/01',
|
||||
text: '法规侧不应出现在撰写栏',
|
||||
status: 'done',
|
||||
updatedAt: 2,
|
||||
},
|
||||
],
|
||||
onSelectRole: () => undefined,
|
||||
onExit: () => undefined,
|
||||
}),
|
||||
);
|
||||
assert.match(html, /data-testid="subagents-view"/);
|
||||
assert.match(html, /撰写专员/);
|
||||
assert.match(html, /先核对素材/);
|
||||
assert.doesNotMatch(html, /法规侧不应出现/);
|
||||
assert.match(html, /返回对话/);
|
||||
assert.match(html, /思考 · /);
|
||||
});
|
||||
@@ -0,0 +1,104 @@
|
||||
import type { SquadMemberDTO } from '../../api/taskRuns';
|
||||
import { processItemsOf, type ProcessItem } from '../../engine/processEvent';
|
||||
import { defaultMemberRole, memberRunStatus } from '../../lib/squadMembers';
|
||||
import type { PhaseEvent } from '../../types/workbench';
|
||||
import { ProcessStepOutput } from './ProcessStepOutput';
|
||||
import { ThinkingBlock } from './ThinkingBlock';
|
||||
import { ToolCallCard } from './ToolCallCard';
|
||||
|
||||
const STATUS_LABEL = {
|
||||
waiting: '等待',
|
||||
running: '执行中',
|
||||
completed: '完成',
|
||||
failed: '异常',
|
||||
} as const;
|
||||
|
||||
export function SubagentsLayout(props: {
|
||||
title: string;
|
||||
runId: string;
|
||||
members: SquadMemberDTO[];
|
||||
items: ProcessItem[];
|
||||
phases: PhaseEvent[];
|
||||
selectedRole?: string;
|
||||
onSelectRole: (role: string) => void;
|
||||
onExit: () => void;
|
||||
notice?: string;
|
||||
}) {
|
||||
const selected = props.selectedRole && props.members.some((member) => member.role === props.selectedRole)
|
||||
? props.selectedRole
|
||||
: defaultMemberRole(props.members, props.phases);
|
||||
const member = props.members.find((item) => item.role === selected);
|
||||
const visible = props.items.filter((item) => item.role === selected || (!item.role && !selected));
|
||||
const runLabel = props.runId.length > 12 ? `${props.runId.slice(0, 8)}…` : props.runId;
|
||||
|
||||
return (
|
||||
<div data-testid="subagents-view" style={{ minHeight: '100vh', background: 'var(--bg, #f8fafc)', color: 'var(--text, #0f172a)' }}>
|
||||
<header style={{ display: 'flex', justifyContent: 'space-between', gap: 12, padding: '12px 16px', borderBottom: '1px solid var(--border, #e2e8f0)' }}>
|
||||
<div>
|
||||
<div style={{ fontWeight: 700 }}>小队过程 · {props.title || '任务'}</div>
|
||||
<div style={{ fontSize: 12, color: 'var(--text-faint, #64748b)' }}>run:{runLabel}</div>
|
||||
</div>
|
||||
<button type="button" onClick={props.onExit}>返回对话</button>
|
||||
</header>
|
||||
{props.notice ? <p role="status" style={{ margin: '8px 16px' }}>{props.notice}</p> : null}
|
||||
<div style={{ display: 'grid', gridTemplateColumns: '240px 1fr', minHeight: 'calc(100vh - 120px)' }}>
|
||||
<aside data-testid="subagents-members" style={{ borderRight: '1px solid var(--border, #e2e8f0)', padding: 8 }}>
|
||||
{props.members.length === 0 ? <p>还没有小队编制</p> : null}
|
||||
{props.members.map((item) => {
|
||||
const status = memberRunStatus(item.role, props.phases);
|
||||
const active = item.role === selected;
|
||||
return (
|
||||
<button
|
||||
key={item.role}
|
||||
type="button"
|
||||
data-testid="subagents-member"
|
||||
data-role={item.role}
|
||||
aria-pressed={active}
|
||||
onClick={() => props.onSelectRole(item.role)}
|
||||
style={{
|
||||
display: 'block',
|
||||
width: '100%',
|
||||
textAlign: 'left',
|
||||
marginBottom: 6,
|
||||
padding: 8,
|
||||
borderRadius: 8,
|
||||
border: active ? '1px solid var(--brand-500, #2563eb)' : '1px solid transparent',
|
||||
background: active ? 'rgba(37, 99, 235, 0.08)' : 'transparent',
|
||||
}}
|
||||
>
|
||||
<div style={{ fontWeight: 650 }}>{item.name || item.role}</div>
|
||||
<div style={{ fontSize: 12 }}>{STATUS_LABEL[status]}</div>
|
||||
</button>
|
||||
);
|
||||
})}
|
||||
</aside>
|
||||
<section data-testid="subagents-detail" style={{ padding: 12 }}>
|
||||
{member ? (
|
||||
<p style={{ marginTop: 0 }}>
|
||||
{member.name} · {member.role}
|
||||
{member.responsibilities?.length ? ` · ${member.responsibilities.join('、')}` : ''}
|
||||
</p>
|
||||
) : (
|
||||
<p>选择一名专员查看过程</p>
|
||||
)}
|
||||
{visible.length === 0 ? <p>还没有过程记录</p> : null}
|
||||
{visible.map((item) => {
|
||||
if (item.category === 'tool_call') return <ToolCallCard key={item.key} item={item} />;
|
||||
if (item.category === 'step_output') return <ProcessStepOutput key={item.key} item={item} />;
|
||||
if (item.category === 'agent_thought') return <ThinkingBlock key={item.key} item={item} />;
|
||||
return <ProcessStepOutput key={item.key} item={item} />;
|
||||
})}
|
||||
</section>
|
||||
</div>
|
||||
<ol data-testid="subagents-timeline" style={{ listStyle: 'none', margin: 0, padding: '8px 16px', borderTop: '1px solid var(--border, #e2e8f0)', fontSize: 12 }}>
|
||||
{props.phases.map((phase) => (
|
||||
<li key={phase.step}>{phase.icon} {phase.label}</li>
|
||||
))}
|
||||
</ol>
|
||||
</div>
|
||||
);
|
||||
}
|
||||
|
||||
export function itemsForLayout(items: Record<string, ProcessItem>): ProcessItem[] {
|
||||
return processItemsOf(items);
|
||||
}
|
||||
@@ -0,0 +1,55 @@
|
||||
import { useEffect, useState } from 'react';
|
||||
import { getTaskRunSquad, type SquadMemberDTO } from '../../api/taskRuns';
|
||||
import { subagentsHash } from '../../engine/anchorNavigator';
|
||||
import { useRunProcessStream } from '../../hooks/useRunProcessStream';
|
||||
import { itemsForLayout, SubagentsLayout } from './SubagentsLayout';
|
||||
|
||||
export function SubagentsView(props: {
|
||||
runId: string;
|
||||
sessionId: string;
|
||||
roleId?: string;
|
||||
onExit: () => void;
|
||||
}) {
|
||||
const stream = useRunProcessStream(props.runId);
|
||||
const [members, setMembers] = useState<SquadMemberDTO[]>([]);
|
||||
const [title, setTitle] = useState('');
|
||||
const [notice, setNotice] = useState<string | undefined>();
|
||||
|
||||
useEffect(() => {
|
||||
let cancelled = false;
|
||||
void getTaskRunSquad(props.runId)
|
||||
.then((squad) => {
|
||||
if (cancelled) return;
|
||||
setMembers(squad.members || []);
|
||||
setTitle(squad.title || squad.task_key || '');
|
||||
setNotice(undefined);
|
||||
})
|
||||
.catch(() => {
|
||||
if (!cancelled) setNotice('暂时读不到小队编制，过程事件仍会显示。');
|
||||
});
|
||||
return () => {
|
||||
cancelled = true;
|
||||
};
|
||||
}, [props.runId]);
|
||||
|
||||
const selectRole = (role: string) => {
|
||||
const next = subagentsHash({ sessionId: props.sessionId, runId: props.runId, roleId: role });
|
||||
if (window.location.hash !== next) window.location.hash = next;
|
||||
};
|
||||
|
||||
const streamNotice = stream.status === 'error' ? '过程流中断，可返回对话后重新进入。' : notice;
|
||||
|
||||
return (
|
||||
<SubagentsLayout
|
||||
title={title}
|
||||
runId={props.runId}
|
||||
members={members}
|
||||
items={itemsForLayout(stream.items)}
|
||||
phases={stream.phases}
|
||||
selectedRole={props.roleId}
|
||||
onSelectRole={selectRole}
|
||||
onExit={props.onExit}
|
||||
notice={streamNotice}
|
||||
/>
|
||||
);
|
||||
}
|
||||
@@ -0,0 +1,22 @@
|
||||
import type { ProcessItem } from '../../engine/processEvent';
|
||||
import { renderBlockMarkdown } from '../../engine/renderBlockMarkdown';
|
||||
import { CollapsibleBlock } from './CollapsibleBlock';
|
||||
|
||||
function summaryOf(item: ProcessItem): string {
|
||||
const head = item.text.replace(/\s+/g, ' ').trim().slice(0, 40);
|
||||
const extra = item.text.length > 40 ? '…' : '';
|
||||
const state = item.status === 'streaming' ? '思考中' : item.status === 'error' ? '思考中断' : '思考';
|
||||
return `${state} · ${head}${extra}（${item.text.length} 字）`;
|
||||
}
|
||||
|
||||
export function ThinkingBlock({ item }: { item: ProcessItem }) {
|
||||
return (
|
||||
<CollapsibleBlock
|
||||
testId="thinking-block"
|
||||
defaultOpen={item.status === 'streaming'}
|
||||
summary={summaryOf(item)}
|
||||
>
|
||||
<div dangerouslySetInnerHTML={{ __html: renderBlockMarkdown(item.text) }} />
|
||||
</CollapsibleBlock>
|
||||
);
|
||||
}
|
||||
@@ -0,0 +1,24 @@
|
||||
import type { ProcessItem } from '../../engine/processEvent';
|
||||
import { renderBlockMarkdown } from '../../engine/renderBlockMarkdown';
|
||||
import { CollapsibleBlock } from './CollapsibleBlock';
|
||||
|
||||
function statusLabel(status: string): string {
|
||||
if (status === 'succeeded') return '完成';
|
||||
if (status === 'failed') return '失败';
|
||||
return '调用中';
|
||||
}
|
||||
|
||||
export function ToolCallCard({ item }: { item: ProcessItem }) {
|
||||
const tool = item.tool;
|
||||
const name = tool?.name || item.title || '工具';
|
||||
const status = tool?.status || (item.status === 'error' ? 'failed' : 'calling');
|
||||
const duration = tool?.durationMs != null ? ` · ${tool.durationMs}ms` : '';
|
||||
const body = [tool?.input ? `入参\n\`\`\`json\n${tool.input}\n\`\`\`` : '', tool?.output ? `结果\n\`\`\`\n${tool.output}\n\`\`\`` : '']
|
||||
.filter(Boolean)
|
||||
.join('\n\n');
|
||||
return (
|
||||
<CollapsibleBlock testId="tool-call-card" summary={`工具 ${name} · ${statusLabel(status)}${duration}`}>
|
||||
{body ? <div dangerouslySetInnerHTML={{ __html: renderBlockMarkdown(body) }} /> : <p>无入参或结果</p>}
|
||||
</CollapsibleBlock>
|
||||
);
|
||||
}
|
||||
@@ -4,7 +4,9 @@ import {
|
||||
parseAnchorHash,
|
||||
formatAnchorHash,
|
||||
navigateToAnchor,
|
||||
parseSubagentsHash,
|
||||
resolveTurnAnchorId,
|
||||
subagentsHash,
|
||||
} from './anchorNavigator';
|
||||
|
||||
test('parseAnchorHash: 正确解析标准深链 hash', () => {
|
||||
@@ -153,3 +155,19 @@ test('resolveTurnAnchorId: 方案 A 轮次归一化测试', () => {
|
||||
assert.equal(resolveTurnAnchorId(102, []), 102, 'empty messages returns raw id');
|
||||
});
|
||||
|
||||
test('parseSubagentsHash 解析小队过程深链且不误伤消息锚点', () => {
|
||||
assert.equal(parseAnchorHash('#/s/sess/run/run-1/agents'), null);
|
||||
assert.deepEqual(parseSubagentsHash('#/s/sess/run/run-1/agents'), {
|
||||
sessionId: 'sess',
|
||||
runId: 'run-1',
|
||||
roleId: undefined,
|
||||
});
|
||||
assert.deepEqual(parseSubagentsHash('#/s/sess/run/run-1/agents/regulatory'), {
|
||||
sessionId: 'sess',
|
||||
runId: 'run-1',
|
||||
roleId: 'regulatory',
|
||||
});
|
||||
const hashed = subagentsHash({ sessionId: 'a/b', runId: 'run 1', roleId: '法规' });
|
||||
assert.deepEqual(parseSubagentsHash(hashed), { sessionId: 'a/b', runId: 'run 1', roleId: '法规' });
|
||||
});
|
||||
|
||||
|
||||
@@ -110,6 +110,33 @@ export function parseAnchorHash(hash: string): AnchorTarget | null {
|
||||
return { sessionId, messageId };
|
||||
}
|
||||
|
||||
export type SubagentsTarget = {
|
||||
sessionId: string;
|
||||
runId: string;
|
||||
roleId?: string;
|
||||
};
|
||||
|
||||
/**
|
||||
* 小队过程子页面深链：#/s/{sessionId}/run/{runId}/agents[/{roleId}]
|
||||
* 与消息锚点 #/s/{sessionId}/m/{messageId} 互斥。
|
||||
*/
|
||||
export function parseSubagentsHash(hash: string): SubagentsTarget | null {
|
||||
if (!hash) return null;
|
||||
const cleaned = hash.startsWith('#') ? hash.substring(1) : hash;
|
||||
const match = cleaned.match(/^\/s\/([^/]+)\/run\/([^/]+)\/agents(?:\/([^/]+))?$/);
|
||||
if (!match) return null;
|
||||
const sessionId = decodeURIComponent(match[1]);
|
||||
const runId = decodeURIComponent(match[2]);
|
||||
if (!sessionId || !runId) return null;
|
||||
const roleId = match[3] ? decodeURIComponent(match[3]) : undefined;
|
||||
return { sessionId, runId, roleId };
|
||||
}
|
||||
|
||||
export function subagentsHash(target: SubagentsTarget): string {
|
||||
const base = `#/s/${encodeURIComponent(target.sessionId)}/run/${encodeURIComponent(target.runId)}/agents`;
|
||||
return target.roleId ? `${base}/${encodeURIComponent(target.roleId)}` : base;
|
||||
}
|
||||
|
||||
/**
|
||||
* 触发 DOM 元素的锚点高亮脉冲
|
||||
*/
|
||||
|
||||
@@ -114,6 +114,10 @@ export function mergeLiveOverHistory(
|
||||
clientTurnId: live.clientTurnId ?? history.clientTurnId,
|
||||
clientMessageId: live.clientMessageId ?? history.clientMessageId,
|
||||
phaseEvents: live.phaseEvents?.length ? live.phaseEvents : history.phaseEvents,
|
||||
processItems:
|
||||
live.processItems && Object.keys(live.processItems).length > 0
|
||||
? live.processItems
|
||||
: history.processItems,
|
||||
livePhaseLabel: preferHistoryTerminal
|
||||
? undefined
|
||||
: live.livePhaseLabel ?? history.livePhaseLabel,
|
||||
|
||||
@@ -83,17 +83,19 @@ export function applyPhaseEvent(strip: PhaseEvent[], phase?: PhaseEventPayload):
|
||||
const base = detail || meta.label;
|
||||
const label = decorateLabel(base, phase);
|
||||
const key = stepKeyOf(phase);
|
||||
const role = phase.role?.trim() || undefined;
|
||||
|
||||
// 已存在的条目：刷新文案；非失败事件再次到达时，failed 条目恢复 running（重试/续跑重新执行该步骤）
|
||||
const existingIdx = strip.findIndex((e) => e.step === key);
|
||||
if (existingIdx >= 0) {
|
||||
return strip.map((e, i) => {
|
||||
if (i !== existingIdx) return e;
|
||||
const nextRole = role || e.role;
|
||||
if (e.status === 'failed' && phase.phase_id !== 'step_failed') {
|
||||
return { ...e, label, detail, status: 'running' as const };
|
||||
return { ...e, label, detail, role: nextRole, status: 'running' as const };
|
||||
}
|
||||
// 已收口条目被重复事件刷新文案时，保持 label 与 status 语义一致。
|
||||
return settleStep({ ...e, label, detail });
|
||||
return settleStep({ ...e, label, detail, role: nextRole });
|
||||
});
|
||||
}
|
||||
|
||||
@@ -101,10 +103,12 @@ export function applyPhaseEvent(strip: PhaseEvent[], phase?: PhaseEventPayload):
|
||||
if (phase.phase_id === 'step_failed') {
|
||||
if (strip.some((e) => e.status === 'running')) {
|
||||
return strip.map((e) =>
|
||||
e.status === 'running' ? { ...e, status: 'failed' as const, detail: detail ?? e.detail } : e,
|
||||
e.status === 'running'
|
||||
? { ...e, status: 'failed' as const, detail: detail ?? e.detail, role: role || e.role }
|
||||
: e,
|
||||
);
|
||||
}
|
||||
return [...strip, { step: key, icon: meta.icon, label, detail, status: 'failed' as const }];
|
||||
return [...strip, { step: key, icon: meta.icon, label, detail, role, status: 'failed' as const }];
|
||||
}
|
||||
|
||||
// 收口型 phase：running 条目置 completed，不追加新节点
|
||||
@@ -118,6 +122,6 @@ export function applyPhaseEvent(strip: PhaseEvent[], phase?: PhaseEventPayload):
|
||||
);
|
||||
return [
|
||||
...settled,
|
||||
{ step: key, icon: meta.icon, label, detail, status: 'running' as const },
|
||||
{ step: key, icon: meta.icon, label, detail, role, status: 'running' as const },
|
||||
];
|
||||
}
|
||||
@@ -0,0 +1,86 @@
|
||||
import assert from 'node:assert/strict';
|
||||
import test from 'node:test';
|
||||
import { applyDetailEvent, processStamp, summarizeProcess } from './processEvent';
|
||||
|
||||
test('applyDetailEvent appends reasoning deltas then closes on done', () => {
|
||||
let items = applyDetailEvent({}, {
|
||||
category: 'agent_thought',
|
||||
step_id: 'tasks/01',
|
||||
role: 'regulatory',
|
||||
item_id: 'thought-1',
|
||||
delta: '先看',
|
||||
}, 1);
|
||||
items = applyDetailEvent(items, {
|
||||
category: 'agent_thought',
|
||||
step_id: 'tasks/01',
|
||||
item_id: 'thought-1',
|
||||
delta: '法规',
|
||||
}, 2);
|
||||
assert.equal(items['tasks/01/thought-1'].text, '先看法规');
|
||||
assert.equal(items['tasks/01/thought-1'].status, 'streaming');
|
||||
items = applyDetailEvent(items, {
|
||||
category: 'agent_thought',
|
||||
step_id: 'tasks/01',
|
||||
item_id: 'thought-1',
|
||||
done: true,
|
||||
}, 3);
|
||||
assert.equal(items['tasks/01/thought-1'].status, 'done');
|
||||
assert.equal(items['tasks/01/thought-1'].text, '先看法规');
|
||||
});
|
||||
|
||||
test('content snapshot replaces accumulated delta', () => {
|
||||
let items = applyDetailEvent({}, {
|
||||
category: 'agent_thought',
|
||||
item_id: 't',
|
||||
delta: '旧',
|
||||
}, 1);
|
||||
items = applyDetailEvent(items, {
|
||||
category: 'agent_thought',
|
||||
item_id: 't',
|
||||
content: '对账后的全文',
|
||||
}, 2);
|
||||
assert.equal(items['/t'].text, '对账后的全文');
|
||||
});
|
||||
|
||||
test('tool frames merge status without dropping the name', () => {
|
||||
let items = applyDetailEvent({}, {
|
||||
category: 'tool_call',
|
||||
role: 'writer',
|
||||
step_id: 'tasks/02',
|
||||
item_id: 'call-1',
|
||||
tool: { name: 'read_file', status: 'calling', input: 'output.md' },
|
||||
}, 1);
|
||||
items = applyDetailEvent(items, {
|
||||
category: 'tool_call',
|
||||
item_id: 'call-1',
|
||||
step_id: 'tasks/02',
|
||||
tool: { status: 'failed', output: 'missing' },
|
||||
done: true,
|
||||
}, 2);
|
||||
const tool = items['tasks/02/call-1'].tool;
|
||||
assert.equal(tool?.name, 'read_file');
|
||||
assert.equal(tool?.status, 'failed');
|
||||
assert.equal(tool?.input, 'output.md');
|
||||
assert.equal(tool?.output, 'missing');
|
||||
assert.equal(items['tasks/02/call-1'].status, 'error');
|
||||
});
|
||||
|
||||
test('summarizeProcess counts roles tools and thoughts', () => {
|
||||
const items = applyDetailEvent(
|
||||
applyDetailEvent({}, {
|
||||
category: 'agent_thought',
|
||||
role: 'regulatory',
|
||||
item_id: 'a',
|
||||
delta: '想',
|
||||
}, 1),
|
||||
{
|
||||
category: 'tool_call',
|
||||
role: 'writer',
|
||||
item_id: 'b',
|
||||
tool: { name: 'write_file', status: 'succeeded' },
|
||||
},
|
||||
2,
|
||||
);
|
||||
assert.equal(summarizeProcess(items), '2 名专员 · 1 次工具调用 · 1 段推演');
|
||||
assert.notEqual(processStamp(items), processStamp({}));
|
||||
});
|
||||
@@ -0,0 +1,133 @@
|
||||
/** 小队过程 detail 帧的纯函数聚合。气泡与子页面共用，不写进消息正文。 */
|
||||
|
||||
export type ProcessCategory = 'agent_thought' | 'tool_call' | 'step_output' | 'log';
|
||||
|
||||
export type ProcessItemStatus = 'streaming' | 'done' | 'error';
|
||||
|
||||
export type ProcessTool = {
|
||||
name: string;
|
||||
status: string;
|
||||
input?: string;
|
||||
output?: string;
|
||||
durationMs?: number;
|
||||
};
|
||||
|
||||
export type ProcessItem = {
|
||||
key: string;
|
||||
category: ProcessCategory;
|
||||
role: string;
|
||||
stepId: string;
|
||||
title?: string;
|
||||
text: string;
|
||||
status: ProcessItemStatus;
|
||||
tool?: ProcessTool;
|
||||
updatedAt: number;
|
||||
};
|
||||
|
||||
export type DetailEvent = {
|
||||
category?: string;
|
||||
step_id?: string;
|
||||
role?: string;
|
||||
title?: string;
|
||||
content?: string;
|
||||
collapsible?: boolean;
|
||||
item_id?: string;
|
||||
delta?: string;
|
||||
done?: boolean;
|
||||
tool?: {
|
||||
name?: string;
|
||||
status?: string;
|
||||
input?: string;
|
||||
output?: string;
|
||||
duration_ms?: number;
|
||||
};
|
||||
};
|
||||
|
||||
const CATEGORIES = new Set<ProcessCategory>(['agent_thought', 'tool_call', 'step_output', 'log']);
|
||||
|
||||
function asCategory(raw?: string): ProcessCategory {
|
||||
if (raw && CATEGORIES.has(raw as ProcessCategory)) return raw as ProcessCategory;
|
||||
return 'log';
|
||||
}
|
||||
|
||||
function mergeTool(prev: ProcessTool | undefined, incoming: DetailEvent['tool']): ProcessTool | undefined {
|
||||
if (!incoming && !prev) return undefined;
|
||||
return {
|
||||
name: incoming?.name || prev?.name || '',
|
||||
status: incoming?.status || prev?.status || 'calling',
|
||||
input: incoming?.input ?? prev?.input,
|
||||
output: incoming?.output ?? prev?.output,
|
||||
durationMs: incoming?.duration_ms ?? prev?.durationMs,
|
||||
};
|
||||
}
|
||||
|
||||
function nextStatus(prev: ProcessItemStatus | undefined, ev: DetailEvent, tool?: ProcessTool): ProcessItemStatus {
|
||||
if (tool?.status === 'failed') return 'error';
|
||||
if (ev.done || tool?.status === 'succeeded') return 'done';
|
||||
if (ev.delta || ev.content) return prev === 'done' || prev === 'error' ? prev : 'streaming';
|
||||
return prev ?? 'streaming';
|
||||
}
|
||||
|
||||
/**
|
||||
* 把一条 detail 帧折叠进过程项表。
|
||||
* delta 追加；content 是快照对账（覆盖）；done / tool 成功收口。
|
||||
*/
|
||||
export function applyDetailEvent(
|
||||
items: Record<string, ProcessItem>,
|
||||
ev?: DetailEvent | null,
|
||||
now = Date.now(),
|
||||
): Record<string, ProcessItem> {
|
||||
if (!ev?.category && !ev?.item_id && !ev?.delta && !ev?.content) return items;
|
||||
const stepId = ev.step_id || '';
|
||||
const itemId = ev.item_id || ev.title || ev.category || 'item';
|
||||
const key = `${stepId}/${itemId}`;
|
||||
const prev = items[key];
|
||||
let text = prev?.text ?? '';
|
||||
if (typeof ev.content === 'string' && ev.content.length > 0) {
|
||||
text = ev.content;
|
||||
} else if (ev.delta) {
|
||||
text += ev.delta;
|
||||
}
|
||||
const tool = mergeTool(prev?.tool, ev.tool);
|
||||
const status = nextStatus(prev?.status, ev, tool);
|
||||
return {
|
||||
...items,
|
||||
[key]: {
|
||||
key,
|
||||
category: asCategory(ev.category || prev?.category),
|
||||
role: ev.role || prev?.role || '',
|
||||
stepId,
|
||||
title: ev.title || prev?.title,
|
||||
text,
|
||||
status,
|
||||
tool,
|
||||
updatedAt: now,
|
||||
},
|
||||
};
|
||||
}
|
||||
|
||||
export function processItemsOf(items?: Record<string, ProcessItem> | null): ProcessItem[] {
|
||||
if (!items) return [];
|
||||
return Object.values(items).sort((a, b) => a.updatedAt - b.updatedAt || a.key.localeCompare(b.key));
|
||||
}
|
||||
|
||||
/** 气泡 L3 摘要：专员数 · 工具次数 · 推演段数。空则返回空串。 */
|
||||
export function summarizeProcess(items?: Record<string, ProcessItem> | null): string {
|
||||
const list = processItemsOf(items);
|
||||
if (list.length === 0) return '';
|
||||
const roles = new Set(list.map((item) => item.role).filter(Boolean));
|
||||
const tools = list.filter((item) => item.category === 'tool_call').length;
|
||||
const thoughts = list.filter((item) => item.category === 'agent_thought').length;
|
||||
const parts: string[] = [];
|
||||
if (roles.size > 0) parts.push(`${roles.size} 名专员`);
|
||||
if (tools > 0) parts.push(`${tools} 次工具调用`);
|
||||
if (thoughts > 0) parts.push(`${thoughts} 段推演`);
|
||||
return parts.join(' · ');
|
||||
}
|
||||
|
||||
/** 供消息合并判断过程项是否变化（不计时间戳）。 */
|
||||
export function processStamp(items?: Record<string, ProcessItem> | null): string {
|
||||
return processItemsOf(items)
|
||||
.map((item) => `${item.key}:${item.status}:${item.text.length}:${item.tool?.status ?? ''}`)
|
||||
.join('|');
|
||||
}
|
||||
@@ -0,0 +1,11 @@
|
||||
import assert from 'node:assert/strict';
|
||||
import test from 'node:test';
|
||||
import { renderBlockMarkdown } from './renderBlockMarkdown';
|
||||
|
||||
test('renderBlockMarkdown escapes html and pretty-prints json fences', () => {
|
||||
const html = renderBlockMarkdown('see <script>\n```json\n{"a":1}\n```');
|
||||
assert.doesNotMatch(html, /<script>/);
|
||||
assert.match(html, /<script>/);
|
||||
assert.match(html, /<pre><code class="lang-json">/);
|
||||
assert.match(html, /"a": 1/);
|
||||
});
|
||||
@@ -0,0 +1,39 @@
|
||||
/** 工具入参/结果用的轻量块级渲染：转义优先，只识别 fenced code。不引入新依赖。 */
|
||||
|
||||
function escapeHtml(value: string): string {
|
||||
return value
|
||||
.replace(/&/g, '&')
|
||||
.replace(/</g, '<')
|
||||
.replace(/>/g, '>')
|
||||
.replace(/"/g, '"');
|
||||
}
|
||||
|
||||
function fenceBody(lang: string, body: string): string {
|
||||
let text = body.replace(/\n$/, '');
|
||||
if (lang === 'json') {
|
||||
try {
|
||||
text = JSON.stringify(JSON.parse(text), null, 2);
|
||||
} catch {
|
||||
// 保持原文，仍然转义。
|
||||
}
|
||||
}
|
||||
const klass = lang ? ` class="lang-${escapeHtml(lang)}"` : '';
|
||||
return `<pre><code${klass}>${escapeHtml(text)}</code></pre>`;
|
||||
}
|
||||
|
||||
export function renderBlockMarkdown(source: string): string {
|
||||
const text = source ?? '';
|
||||
const parts: string[] = [];
|
||||
const re = /```([a-zA-Z0-9_-]*)\n([\s\S]*?)```/g;
|
||||
let last = 0;
|
||||
for (const match of text.matchAll(re)) {
|
||||
const index = match.index ?? 0;
|
||||
if (index > last) parts.push(`<p>${escapeHtml(text.slice(last, index)).replace(/\n/g, '<br>')}</p>`);
|
||||
parts.push(fenceBody(match[1] || '', match[2] || ''));
|
||||
last = index + match[0].length;
|
||||
}
|
||||
if (last < text.length) {
|
||||
parts.push(`<p>${escapeHtml(text.slice(last)).replace(/\n/g, '<br>')}</p>`);
|
||||
}
|
||||
return parts.join('');
|
||||
}
|
||||
@@ -12,6 +12,7 @@ import {
|
||||
saveSessionDraft,
|
||||
} from '../data/storage';
|
||||
import { enrichMessagesWithCategories } from '../engine/categoryHelper';
|
||||
import { processStamp } from '../engine/processEvent';
|
||||
import {
|
||||
filterPendingLiveMessages,
|
||||
isLiveMessageInHistory,
|
||||
@@ -529,6 +530,7 @@ export function useChatWindow({
|
||||
liveM.status !== m.status ||
|
||||
liveM.livePhaseLabel !== m.livePhaseLabel ||
|
||||
(liveM.phaseEvents?.length || 0) !== (m.phaseEvents?.length || 0) ||
|
||||
processStamp(liveM.processItems) !== processStamp(m.processItems) ||
|
||||
liveM.errorInfo !== m.errorInfo ||
|
||||
liveM.serverId !== m.serverId)
|
||||
) {
|
||||
|
||||
@@ -0,0 +1,59 @@
|
||||
import { useEffect, useState } from 'react';
|
||||
import type { StreamEvent } from '../api/chatStream';
|
||||
import { readTaskRunSSEOnce, subscribeTaskRunEvents } from '../api/taskRuns';
|
||||
import { applyPhaseEvent } from '../engine/phaseEventMapper';
|
||||
import { applyDetailEvent, type ProcessItem } from '../engine/processEvent';
|
||||
import type { PhaseEvent } from '../types/workbench';
|
||||
|
||||
export type RunProcessStatus = 'loading' | 'live' | 'done' | 'error';
|
||||
|
||||
/** 子页面自持订阅：先重放已落库事件，未终态再续订 live。 */
|
||||
export function useRunProcessStream(runId: string): {
|
||||
items: Record<string, ProcessItem>;
|
||||
phases: PhaseEvent[];
|
||||
status: RunProcessStatus;
|
||||
} {
|
||||
const [items, setItems] = useState<Record<string, ProcessItem>>({});
|
||||
const [phases, setPhases] = useState<PhaseEvent[]>([]);
|
||||
const [status, setStatus] = useState<RunProcessStatus>('loading');
|
||||
|
||||
useEffect(() => {
|
||||
const controller = new AbortController();
|
||||
let itemState: Record<string, ProcessItem> = {};
|
||||
let phaseState: PhaseEvent[] = [];
|
||||
let terminal = false;
|
||||
const onEvent = (ev: StreamEvent) => {
|
||||
if (ev.type === 'detail') {
|
||||
itemState = applyDetailEvent(itemState, ev.detail);
|
||||
setItems(itemState);
|
||||
} else if (ev.type === 'phase') {
|
||||
phaseState = applyPhaseEvent(phaseState, ev.phase);
|
||||
setPhases(phaseState);
|
||||
} else if (ev.type === 'done' || ev.type === 'error') {
|
||||
terminal = true;
|
||||
}
|
||||
};
|
||||
void (async () => {
|
||||
try {
|
||||
const replay = await readTaskRunSSEOnce(runId, 0, onEvent, controller.signal, { replayOnly: true });
|
||||
if (controller.signal.aborted) return;
|
||||
if (replay.terminal || terminal) {
|
||||
setStatus('done');
|
||||
return;
|
||||
}
|
||||
setStatus('live');
|
||||
await subscribeTaskRunEvents(
|
||||
runId,
|
||||
{ onEvent, afterSeq: replay.lastSeq },
|
||||
controller.signal,
|
||||
);
|
||||
if (!controller.signal.aborted) setStatus('done');
|
||||
} catch {
|
||||
if (!controller.signal.aborted) setStatus('error');
|
||||
}
|
||||
})();
|
||||
return () => controller.abort();
|
||||
}, [runId]);
|
||||
|
||||
return { items, phases, status };
|
||||
}
|
||||
@@ -12,6 +12,7 @@ import { createTaskRun, getTaskRun, resumeTaskRun, subscribeTaskRunEvents } from
|
||||
import { readActiveTasks, removeActiveTask, touchActiveTaskPhase, upsertActiveTask } from '../lib/activeTaskCache';
|
||||
import { buildTaskRequest, streamTask } from '../api/taskStream';
|
||||
import { applyPhaseEvent } from '../engine/phaseEventMapper';
|
||||
import { applyDetailEvent } from '../engine/processEvent';
|
||||
import { resolveTaskTemplateForMessage, wrongWorkflowGuidance } from '../engine/templateMatch';
|
||||
import { GOVERNANCE_BLOCKED_HINT, isBlockedGovernanceStatus } from '../lib/blockedGovernanceCopy';
|
||||
import {
|
||||
@@ -501,6 +502,14 @@ export function useTaskRunAttempts(input: TaskRunAttemptsInput) {
|
||||
}));
|
||||
break;
|
||||
}
|
||||
case 'detail': {
|
||||
// 过程帧只进 processItems，不拼进气泡正文。
|
||||
patchAsst((m) => ({
|
||||
status: m.status === 'pending' ? 'streaming_status' : m.status,
|
||||
processItems: applyDetailEvent(m.processItems || {}, ev.detail),
|
||||
}));
|
||||
break;
|
||||
}
|
||||
case 'status': {
|
||||
if (ev.status?.message) {
|
||||
patchAsst({ status: 'streaming_status', livePhaseLabel: ev.status.message });
|
||||
|
||||
@@ -0,0 +1,22 @@
|
||||
import assert from 'node:assert/strict';
|
||||
import test from 'node:test';
|
||||
import { defaultMemberRole, memberRunStatus } from './squadMembers';
|
||||
import type { PhaseEvent } from '../types/workbench';
|
||||
|
||||
const phases: PhaseEvent[] = [
|
||||
{ step: 'step_executing#1', icon: '⚙️', label: '法规', status: 'completed', role: 'regulatory' },
|
||||
{ step: 'step_executing#2', icon: '⚙️', label: '撰写', status: 'running', role: 'writer' },
|
||||
];
|
||||
|
||||
test('memberRunStatus prefers running over older completed', () => {
|
||||
assert.equal(memberRunStatus('writer', phases), 'running');
|
||||
assert.equal(memberRunStatus('regulatory', phases), 'completed');
|
||||
assert.equal(memberRunStatus('leader', phases), 'waiting');
|
||||
});
|
||||
|
||||
test('defaultMemberRole selects the running member', () => {
|
||||
assert.equal(
|
||||
defaultMemberRole([{ role: 'regulatory' }, { role: 'writer' }, { role: 'leader' }], phases),
|
||||
'writer',
|
||||
);
|
||||
});
|
||||
@@ -0,0 +1,24 @@
|
||||
import type { PhaseEvent } from '../types/workbench';
|
||||
|
||||
export type MemberRunStatus = 'waiting' | 'running' | 'completed' | 'failed';
|
||||
|
||||
export function memberRunStatus(role: string, phases: PhaseEvent[]): MemberRunStatus {
|
||||
const related = phases.filter((phase) => phase.role === role);
|
||||
if (related.some((phase) => phase.status === 'running')) return 'running';
|
||||
if (related.some((phase) => phase.status === 'failed')) return 'failed';
|
||||
if (related.some((phase) => phase.status === 'completed')) return 'completed';
|
||||
return 'waiting';
|
||||
}
|
||||
|
||||
/** 未指定 role 时：当前执行中，否则最近完成，否则编制第一人。 */
|
||||
export function defaultMemberRole(
|
||||
members: { role: string }[],
|
||||
phases: PhaseEvent[],
|
||||
): string | undefined {
|
||||
const running = members.find((member) => memberRunStatus(member.role, phases) === 'running');
|
||||
if (running) return running.role;
|
||||
for (let i = members.length - 1; i >= 0; i -= 1) {
|
||||
if (memberRunStatus(members[i].role, phases) === 'completed') return members[i].role;
|
||||
}
|
||||
return members[0]?.role;
|
||||
}
|
||||
@@ -1,3 +1,4 @@
|
||||
import type { ProcessItem } from '../engine/processEvent';
|
||||
import { DocLifecycleState, UserRole } from './docgov';
|
||||
import { DocVisibilityRule } from './rbac';
|
||||
import { MessageContentKind, MessageExportCapability } from './replyExport';
|
||||
@@ -57,6 +58,8 @@ export interface PhaseEvent {
|
||||
status: 'pending' | 'running' | 'completed' | 'failed';
|
||||
/** 后端 phase.message 原文（刷新快照 / 时间线当前行展开用） */
|
||||
detail?: string;
|
||||
/** 专员 role（step_executing 等按成员归集状态） */
|
||||
role?: string;
|
||||
}
|
||||
|
||||
/** 服务端写入 chat_messages.metadata.task_phase_snapshot 的折叠进度快照 */
|
||||
@@ -218,6 +221,11 @@ export interface ChatMessage {
|
||||
branchTag?: string;
|
||||
timestamp: string;
|
||||
phaseEvents?: PhaseEvent[];
|
||||
/**
|
||||
* 过程 detail 聚合（思考/工具/步骤产出）。只在气泡摘要与小队子页面使用，
|
||||
* 不得拼进 content。
|
||||
*/
|
||||
processItems?: Record<string, ProcessItem>;
|
||||
/** 任务总耗时（毫秒）；done 事件或 snapshot.duration_ms */
|
||||
taskDurationMs?: number;
|
||||
/** 历史消息带来的 phase 快照标记（有则刷新不必再 replay） */
|
||||
|
||||
Reference in New Issue
Block a user