tzt
ddef8bec1c
- OpenAICompatChat 默认 stream=True：httpx 流式解析 OpenAI chunk， tool_calls 碎片按 index 组装（不在正文展示），include_usage 计量； 失败自动回退非流式一次（已有部分增量输出则如实抛出）； 补 raise_for_status（修复流式 4xx 误入回退的缺陷） - ToolLoop 增 on_delta（签名探测兼容 2/3 参 chat_fn）；_loads_json_object 宽松解析 - DeltaThrottle >=48 字符节流落 delta 事件；单模型与两级模式（含规划者）接线 - 前端：streamText 打字机渲染 + 光标动画；工具/阶段事件到达时清空归档 - 配套修复：AgentView 闭包持有 push 前原始对象导致响应式丢失、过程事件不渲染 - 测试 +7（SSE 解析/碎片组装/回退/部分失败抛出/on_delta/节流/service delta 事件）， 全量 296 passed
912 lines
40 KiB
Python
912 lines
40 KiB
Python
"""智能体服务（AgentService）—— zcode 式"模型操作工作区文件"的网关侧封装。
|
||
|
||
职责：
|
||
- OpenAICompatChat：OpenAI 兼容 /chat/completions 的工具调用客户端（ToolLoop 的 chat_fn），
|
||
支持 httpx transport/client 注入（测试用 MockTransport，对齐 D11 封闭性）。
|
||
- AgentService：运行一次智能体任务——事件逐条落盘 agent_runs/{id}/events.jsonl，
|
||
终态写 status.json；SSE 端点轮询事件文件增量推送（与 v3 workspace 监视同思路，
|
||
不侵入 router_system）。
|
||
- 模型来源：模型池 agent 角色（或显式 pool_id），否则回退经典 Architect 设置。
|
||
|
||
安全与护栏：
|
||
- 文件操作被 WorkspaceTools 关押在工作区根目录内
|
||
- 轮数上限（agent.max_rounds）与 token 熔断（agent.token_cap）双护栏
|
||
"""
|
||
from __future__ import annotations
|
||
|
||
import asyncio
|
||
import json
|
||
import time
|
||
import uuid
|
||
from dataclasses import dataclass, field
|
||
from pathlib import Path
|
||
from typing import Any, Awaitable, Callable, Dict, List, Optional
|
||
|
||
from router_system.tools import ToolLoop, WorkspaceTools
|
||
|
||
# 运行目录（与 runs/ 平级）
|
||
AGENT_RUNS_DIR = Path("agent_runs")
|
||
|
||
STATE_RUNNING = "running"
|
||
STATE_DONE = "done"
|
||
STATE_FAILED = "failed"
|
||
|
||
AGENT_SYSTEM_PROMPT = (
|
||
"你是端云协同 LLM 系统中的智能体（Agent），正在操作用户选择的**真实项目工作目录**。"
|
||
"你拥有的工具：list_dir（列目录）、read_file（读文件）、write_file（写文件/新建）、"
|
||
"edit_file（精确替换编辑：old_string 须唯一匹配）、search_files（跨文件搜索内容）、"
|
||
"run_command（执行 shell 命令，仅当系统开启 allow_shell 时可用，否则不要尝试）、"
|
||
"web_fetch（抓取公网 http/https 文档页面，私网地址会被拒绝）。"
|
||
"像编程助手一样工作：先列目录/搜索了解项目结构，读文件核对原文后再用 edit_file 小步修改"
|
||
"（或 write_file 新建），需要查外部资料时用 web_fetch，需要时运行命令验证。"
|
||
"任务完成或给出结论后，直接输出给用户的最终答复（中文，不要再调用工具）。"
|
||
"注意：不要反复以完全相同的参数调用同一工具——那不会带来新信息。"
|
||
)
|
||
|
||
# ── 两级智能体（D7）：规划者（大模型）+ 执行者（本地小模型），交接走 handoff 文档 ──
|
||
PLANNER_SYSTEM_PROMPT = (
|
||
"你是两级智能体中的**规划者**（大模型）。执行者是一个能力有限的本地小模型，"
|
||
"只能机械地使用工具。你的职责：把用户任务拆成执行者可照做的**具体指令**，"
|
||
"并在执行后审查其汇报。输出必须是合法 JSON 对象（不要 markdown 围栏）。"
|
||
)
|
||
EXECUTOR_SYSTEM_PROMPT = (
|
||
"你是两级智能体中的**执行者**（本地小模型）。规划者已给你具体指令，"
|
||
"你只负责用工具完成指令并在最后**汇报**：做了什么、结果如何、有什么问题。"
|
||
"严格遵守指令范围，不要自行扩大任务。汇报用中文，是给规划者看的，"
|
||
"要列出：修改的文件、关键命令输出、未完成项。"
|
||
)
|
||
|
||
# 规划者首轮：产出指令（JSON）
|
||
_PLAN_SCHEMA_HINT = {
|
||
"instructions": "string(给执行者的具体步骤指令，<=600字)",
|
||
"acceptance": "string(验收标准，<=200字)",
|
||
}
|
||
# 规划者审查轮：裁决（JSON）
|
||
_REVIEW_SCHEMA_HINT = {
|
||
"verdict": "enum(done|redo)",
|
||
"reply_to_executor": "string(verdict=redo 时给执行者的补充指令；done 时可空)",
|
||
"final_answer": "string(verdict=done 时给用户的最终答复)",
|
||
}
|
||
|
||
DEFAULT_MAX_HANDOFFS = 2 # 规划者<->执行者交接轮数上限
|
||
|
||
|
||
def _parse_json_loose(content: str) -> Dict[str, Any]:
|
||
"""宽松解析规划者的 JSON 输出（剥围栏/取首个对象）；失败返回 {}。"""
|
||
try:
|
||
from router_system.architect import ArchitectClient
|
||
return ArchitectClient._parse_json(content)
|
||
except Exception:
|
||
return {}
|
||
|
||
|
||
# ─────────────────────────────────────────────────────────────────────────────
|
||
# OpenAI 兼容工具调用客户端
|
||
# ─────────────────────────────────────────────────────────────────────────────
|
||
class OpenAICompatChat:
|
||
"""ToolLoop.chat_fn 的 OpenAI 兼容实现（支持 tools 参数）。"""
|
||
|
||
def __init__(
|
||
self,
|
||
base_url: str,
|
||
api_key: Optional[str],
|
||
model: str,
|
||
temperature: float = 0.3,
|
||
max_tokens: int = 4096,
|
||
timeout_s: float = 120.0,
|
||
stream: bool = True,
|
||
max_retries: int = 2,
|
||
retry_delay_s: float = 1.0,
|
||
transport: Any = None,
|
||
_client: Any = None,
|
||
):
|
||
self.base_url = base_url.rstrip("/")
|
||
self.api_key = api_key
|
||
self.model = model
|
||
self.temperature = temperature
|
||
self.max_tokens = max_tokens
|
||
self.timeout_s = timeout_s
|
||
self.stream = stream # D10：默认流式；解析失败自动回退非流式
|
||
self.max_retries = max(0, int(max_retries)) # 可重试错误的重试次数（dsh llm-retry 同款）
|
||
self.retry_delay_s = max(0.0, float(retry_delay_s)) # 指数退避基数
|
||
self._transport = transport
|
||
self._client = _client
|
||
self._owns = _client is None
|
||
|
||
def _get_client(self):
|
||
if self._client is None:
|
||
import httpx
|
||
kwargs: Dict[str, Any] = {"timeout": self.timeout_s}
|
||
if self._transport is not None:
|
||
kwargs["transport"] = self._transport
|
||
self._client = httpx.AsyncClient(**kwargs)
|
||
return self._client
|
||
|
||
async def aclose(self) -> None:
|
||
if self._owns and self._client is not None:
|
||
await self._client.aclose()
|
||
self._client = None
|
||
|
||
async def __call__(self, messages: List[Dict[str, Any]],
|
||
tools_spec: List[Dict[str, Any]],
|
||
on_delta: Optional[Callable[[str], None]] = None) -> Dict[str, Any]:
|
||
"""ToolLoop.chat_fn：默认流式（D10）；流式不可用时回退非流式（带重试退避）。"""
|
||
if self.stream:
|
||
try:
|
||
return await self._stream_call(messages, tools_spec, on_delta)
|
||
except Exception:
|
||
# 已有部分增量输出则如实抛出；否则回退非流式
|
||
if getattr(self, "_stream_partial", False):
|
||
raise
|
||
return await self._post_with_retry(messages, tools_spec)
|
||
|
||
@staticmethod
|
||
def _is_retryable(exc: Exception) -> bool:
|
||
"""可重试错误：网络传输类 / 408 / 429 / 5xx（dsh retryableCodes 同思路）。"""
|
||
import httpx
|
||
if isinstance(exc, httpx.TransportError):
|
||
return True
|
||
if isinstance(exc, httpx.HTTPStatusError):
|
||
code = exc.response.status_code
|
||
return code in (408, 429) or code >= 500
|
||
return False
|
||
|
||
async def _post_with_retry(self, messages: List[Dict[str, Any]],
|
||
tools_spec: List[Dict[str, Any]]) -> Dict[str, Any]:
|
||
"""非流式调用 + 指数退避重试（仅针对可重试错误）。"""
|
||
for attempt in range(self.max_retries + 1):
|
||
try:
|
||
return await self._post_once(messages, tools_spec)
|
||
except Exception as exc:
|
||
if attempt >= self.max_retries or not self._is_retryable(exc):
|
||
raise
|
||
await asyncio.sleep(self.retry_delay_s * (2 ** attempt))
|
||
|
||
async def _stream_call(self, messages: List[Dict[str, Any]],
|
||
tools_spec: List[Dict[str, Any]],
|
||
on_delta: Optional[Callable[[str], None]]) -> Dict[str, Any]:
|
||
"""流式调用：逐段转发 content 增量；tool_calls 碎片按 index 组装（不在正文展示）。"""
|
||
import json as _json
|
||
self._stream_partial = False # 每次调用前复位（防上次的标志影响本次回退判定）
|
||
body: Dict[str, Any] = {
|
||
"model": self.model,
|
||
"messages": messages,
|
||
"temperature": self.temperature,
|
||
"max_tokens": self.max_tokens,
|
||
"stream": True,
|
||
"stream_options": {"include_usage": True},
|
||
}
|
||
if tools_spec:
|
||
body["tools"] = tools_spec
|
||
body["tool_choice"] = "auto"
|
||
headers = {"Authorization": f"Bearer {self.api_key}"} if self.api_key else {}
|
||
client = self._get_client()
|
||
content_parts: List[str] = []
|
||
tc_slots: Dict[int, Dict[str, str]] = {}
|
||
usage: Dict[str, Any] = {}
|
||
async with client.stream("POST", f"{self.base_url}/chat/completions",
|
||
headers=headers, json=body) as resp:
|
||
resp.raise_for_status()
|
||
async for line in resp.aiter_lines():
|
||
line = line.strip()
|
||
if not line.startswith("data:"):
|
||
continue
|
||
payload = line[5:].strip()
|
||
if payload == "[DONE]":
|
||
break
|
||
try:
|
||
obj = _json.loads(payload)
|
||
except _json.JSONDecodeError:
|
||
continue
|
||
choices = obj.get("choices") or [{}]
|
||
delta = (choices[0].get("delta") or {}) if choices else {}
|
||
piece = delta.get("content")
|
||
if piece:
|
||
self._stream_partial = True
|
||
content_parts.append(piece)
|
||
if on_delta is not None:
|
||
try:
|
||
on_delta(piece)
|
||
except Exception:
|
||
pass
|
||
for tc in delta.get("tool_calls") or []:
|
||
idx = int(tc.get("index", 0))
|
||
slot = tc_slots.setdefault(idx, {"id": "", "name": "", "args": ""})
|
||
if tc.get("id"):
|
||
slot["id"] = tc["id"]
|
||
fn = tc.get("function") or {}
|
||
if fn.get("name"):
|
||
slot["name"] = fn["name"]
|
||
if fn.get("arguments"):
|
||
slot["args"] += fn["arguments"]
|
||
if obj.get("usage"):
|
||
usage = obj["usage"]
|
||
content = "".join(content_parts) or None
|
||
from router_system.tools import _loads_json_object
|
||
tool_calls = []
|
||
for idx in sorted(tc_slots):
|
||
slot = tc_slots[idx]
|
||
tool_calls.append({
|
||
"id": slot["id"] or f"call_{idx}",
|
||
"name": slot["name"],
|
||
"arguments": _loads_json_object(slot["args"]),
|
||
})
|
||
return {"content": content, "tool_calls": tool_calls, "usage": usage}
|
||
|
||
async def _post_once(self, messages: List[Dict[str, Any]],
|
||
tools_spec: List[Dict[str, Any]]) -> Dict[str, Any]:
|
||
"""非流式调用（回退路径）。"""
|
||
body: Dict[str, Any] = {
|
||
"model": self.model,
|
||
"messages": messages,
|
||
"temperature": self.temperature,
|
||
"max_tokens": self.max_tokens,
|
||
}
|
||
if tools_spec:
|
||
body["tools"] = tools_spec
|
||
body["tool_choice"] = "auto"
|
||
headers = {"Authorization": f"Bearer {self.api_key}"} if self.api_key else {}
|
||
client = self._get_client()
|
||
resp = await client.post(f"{self.base_url}/chat/completions",
|
||
headers=headers, json=body)
|
||
resp.raise_for_status()
|
||
data = resp.json()
|
||
msg = (data.get("choices") or [{}])[0].get("message") or {}
|
||
from router_system.tools import parse_tool_calls
|
||
return {
|
||
"content": msg.get("content"),
|
||
"tool_calls": parse_tool_calls(msg),
|
||
"usage": data.get("usage") or {},
|
||
}
|
||
|
||
|
||
# ─────────────────────────────────────────────────────────────────────────────
|
||
# 智能体服务
|
||
# ─────────────────────────────────────────────────────────────────────────────
|
||
@dataclass
|
||
class AgentRunInfo:
|
||
"""一次智能体运行的状态快照（内存 + status.json 双写）。"""
|
||
request_id: str
|
||
task: str = ""
|
||
model: str = ""
|
||
state: str = STATE_RUNNING
|
||
started_at: float = 0.0
|
||
finished_at: float = 0.0
|
||
error: Optional[str] = None
|
||
response: str = ""
|
||
rounds: int = 0
|
||
prompt_tokens: int = 0
|
||
completion_tokens: int = 0
|
||
pool_id: str = ""
|
||
workspace: str = "" # 本次运行使用的工作区根目录（绝对路径）
|
||
executor_model: str = "" # 两级模式：执行者模型名（空 = 单模型模式）
|
||
mode: str = "single" # single | dual
|
||
tool_calls: int = 0 # 本次运行的工具调用步数
|
||
asyncio_task: Optional[asyncio.Task] = field(default=None, repr=False)
|
||
|
||
def to_dict(self) -> Dict[str, Any]:
|
||
return {
|
||
"request_id": self.request_id,
|
||
"task": self.task,
|
||
"model": self.model,
|
||
"state": self.state,
|
||
"started_at": self.started_at,
|
||
"finished_at": self.finished_at,
|
||
"error": self.error,
|
||
"response": self.response,
|
||
"rounds": self.rounds,
|
||
"prompt_tokens": self.prompt_tokens,
|
||
"completion_tokens": self.completion_tokens,
|
||
"pool_id": self.pool_id,
|
||
"workspace": self.workspace,
|
||
"executor_model": self.executor_model,
|
||
"mode": self.mode,
|
||
"tool_calls": self.tool_calls,
|
||
}
|
||
|
||
|
||
class AgentService:
|
||
"""智能体运行服务：事件落盘 + 状态管理。"""
|
||
|
||
def __init__(self, run_dir: str | Path = AGENT_RUNS_DIR):
|
||
self.run_dir = Path(run_dir)
|
||
self._runs: Dict[str, AgentRunInfo] = {}
|
||
self.max_running = 5
|
||
|
||
# ---------- 路径 ----------
|
||
def _dir(self, request_id: str) -> Path:
|
||
return self.run_dir / request_id
|
||
|
||
def events_path(self, request_id: str) -> Path:
|
||
return self._dir(request_id) / "events.jsonl"
|
||
|
||
def status_path(self, request_id: str) -> Path:
|
||
return self._dir(request_id) / "status.json"
|
||
|
||
# ---------- 注册与查询 ----------
|
||
def register(self, request_id: str, task: str, model: str, pool_id: str,
|
||
workspace: str = "", executor_model: str = "",
|
||
mode: str = "single") -> Optional[AgentRunInfo]:
|
||
running = [r for r in self._runs.values() if r.state == STATE_RUNNING]
|
||
if len(running) >= self.max_running:
|
||
return None
|
||
info = AgentRunInfo(request_id=request_id, task=task, model=model,
|
||
pool_id=pool_id, workspace=workspace,
|
||
executor_model=executor_model, mode=mode,
|
||
started_at=time.time())
|
||
self._runs[request_id] = info
|
||
self._dir(request_id).mkdir(parents=True, exist_ok=True)
|
||
self._write_status(info)
|
||
return info
|
||
|
||
def get(self, request_id: str) -> Optional[AgentRunInfo]:
|
||
return self._runs.get(request_id)
|
||
|
||
# ---------- 执行 ----------
|
||
async def run(self, info: AgentRunInfo, chat: Any, workspace_dir: str | Path,
|
||
max_rounds: int = 8, token_cap: int = 0,
|
||
allow_shell: bool = False, shell_timeout_s: int = 20,
|
||
allow_net: bool = True,
|
||
executor_chat: Any = None,
|
||
max_handoffs: int = DEFAULT_MAX_HANDOFFS,
|
||
session: Optional["AgentSession"] = None,
|
||
approval_policy: str = "dangerous",
|
||
approval_timeout_s: int = 120) -> None:
|
||
"""执行智能体任务（由调用方包成后台协程）。
|
||
|
||
executor_chat 为空 = 单模型模式（chat 全程包办）；
|
||
提供时进入两级模式：chat 作规划者，executor_chat 作执行者（D7）。
|
||
session 提供时：既往轮次作为对话上下文，完成后把本轮追加进会话。
|
||
"""
|
||
history = self._history_from_session(session)
|
||
approval_mgr = ApprovalManager()
|
||
info._approval_manager = approval_mgr # 供 /approve 端点裁决（瞬态属性）
|
||
throttle = DeltaThrottle(lambda ev: self._append_event(info, ev))
|
||
|
||
async def approval_hook(name: str, args: Dict[str, Any]) -> bool:
|
||
"""按策略判定；需审批则挂起等用户裁决，超时 fail-closed。"""
|
||
if not needs_approval(approval_policy, name):
|
||
return True
|
||
aid = "ap" + uuid.uuid4().hex[:8]
|
||
ev = approval_mgr.open(aid)
|
||
self._append_event(info, {"type": "approval_request", "id": aid,
|
||
"name": name, "arguments": args,
|
||
"policy": approval_policy})
|
||
# 轮询等待（0.1s 步进）：不用 wait_for——portal 循环下其定时器不可靠
|
||
allowed = False
|
||
note = ""
|
||
deadline = time.time() + max(1, approval_timeout_s)
|
||
while time.time() < deadline:
|
||
if ev.is_set():
|
||
allowed = approval_mgr._pending.get(aid, {}).get("allowed", False)
|
||
break
|
||
await asyncio.sleep(0.1)
|
||
else:
|
||
note = f"超时（{approval_timeout_s}s）未响应，自动拒绝"
|
||
if ev.is_set() and not allowed:
|
||
note = note or "用户拒绝"
|
||
approval_mgr.close(aid)
|
||
self._append_event(info, {"type": "approval_decided", "id": aid,
|
||
"name": name, "allowed": allowed,
|
||
**({"note": note} if note else {})})
|
||
return allowed
|
||
|
||
try:
|
||
if executor_chat is not None:
|
||
result = await self.run_dual(
|
||
info, chat, executor_chat, workspace_dir,
|
||
max_rounds=max_rounds, token_cap=token_cap,
|
||
allow_shell=allow_shell, shell_timeout_s=shell_timeout_s,
|
||
allow_net=allow_net,
|
||
max_handoffs=max_handoffs,
|
||
approval_hook=approval_hook)
|
||
else:
|
||
tools = WorkspaceTools(workspace_dir, allow_shell=allow_shell,
|
||
shell_timeout_s=shell_timeout_s,
|
||
allow_net=allow_net)
|
||
loop = ToolLoop(tools, chat, max_rounds=max_rounds, token_cap=token_cap,
|
||
on_event=self._make_event_writer(info),
|
||
approval_hook=approval_hook,
|
||
on_delta=throttle.make_cb("executor"))
|
||
result = await loop.run(info.task, system=AGENT_SYSTEM_PROMPT,
|
||
history=history)
|
||
throttle.flush("executor")
|
||
self._apply_result(info, result)
|
||
except Exception as exc: # pragma: no cover
|
||
info.state = STATE_FAILED
|
||
info.error = f"{type(exc).__name__}: {exc}"
|
||
self._append_event(info, {"type": "final", "round": info.rounds,
|
||
"reason": "error", "error": info.error})
|
||
finally:
|
||
info.finished_at = time.time()
|
||
self._write_status(info)
|
||
if session is not None:
|
||
session.data["turns"].append({
|
||
"request_id": info.request_id,
|
||
"task": info.task,
|
||
"response": info.response,
|
||
"state": info.state,
|
||
"tool_calls": info.tool_calls,
|
||
"tokens": info.prompt_tokens + info.completion_tokens,
|
||
"error": info.error,
|
||
"ts": info.finished_at,
|
||
})
|
||
get_session_store().save(session)
|
||
|
||
@staticmethod
|
||
def _history_from_session(session: Optional["AgentSession"],
|
||
max_turns: int = 6,
|
||
max_chars: int = 1500) -> List[Dict[str, Any]]:
|
||
"""把会话既往轮次折叠成对话上下文（不含工具细节）。"""
|
||
if session is None:
|
||
return []
|
||
turns = [t for t in session.data.get("turns", [])
|
||
if t.get("state") == STATE_DONE and t.get("response")]
|
||
out: List[Dict[str, Any]] = []
|
||
for t in turns[-max_turns:]:
|
||
out.append({"role": "user", "content": str(t["task"])[:max_chars]})
|
||
out.append({"role": "assistant", "content": str(t["response"])[:max_chars]})
|
||
return out
|
||
|
||
def _apply_result(self, info: AgentRunInfo, result: Dict[str, Any]) -> None:
|
||
"""把循环结果落到运行状态（单/两级模式共用）。"""
|
||
info.response = result.get("response", "")
|
||
info.rounds = int(result.get("rounds", 0))
|
||
info.prompt_tokens = int(result.get("prompt_tokens", 0))
|
||
info.completion_tokens = int(result.get("completion_tokens", 0))
|
||
if result.get("reason") == "error":
|
||
info.state = STATE_FAILED
|
||
info.error = result.get("error")
|
||
elif result.get("reason") in ("token_cap", "max_rounds", "max_handoffs"):
|
||
# 触顶属于护栏行为：结果仍交付，但标记部分完成信息
|
||
info.state = STATE_DONE
|
||
info.error = result.get("error")
|
||
else:
|
||
info.state = STATE_DONE
|
||
|
||
# ---------- 两级模式（D7）：规划者 + 执行者 ----------
|
||
async def run_dual(self, info: AgentRunInfo, planner_chat: Any, executor_chat: Any,
|
||
workspace_dir: str | Path, max_rounds: int = 8,
|
||
token_cap: int = 0, allow_shell: bool = False,
|
||
shell_timeout_s: int = 20, allow_net: bool = True,
|
||
max_handoffs: int = DEFAULT_MAX_HANDOFFS,
|
||
approval_hook: Optional[Callable[[str, Dict[str, Any]], Awaitable[bool]]] = None
|
||
) -> Dict[str, Any]:
|
||
"""大模型拆解/审查 + 小模型执行工具轮，交接状态写 handoff.json（智能体版交流文本）。"""
|
||
info.mode = "dual"
|
||
tools = WorkspaceTools(workspace_dir, allow_shell=allow_shell,
|
||
shell_timeout_s=shell_timeout_s, allow_net=allow_net)
|
||
throttle = DeltaThrottle(lambda ev: self._append_event(info, ev))
|
||
handoff: Dict[str, Any] = {
|
||
"task": info.task, "planner_model": info.model,
|
||
"executor_model": info.executor_model, "workspace": info.workspace,
|
||
"instructions": "", "acceptance": "", "exchanges": [],
|
||
}
|
||
spent = {"in": 0, "out": 0}
|
||
total_rounds = 0
|
||
|
||
def _account(usage: Dict[str, Any] | None) -> None:
|
||
spent["in"] += int((usage or {}).get("prompt_tokens", 0))
|
||
spent["out"] += int((usage or {}).get("completion_tokens", 0))
|
||
|
||
def _save_handoff() -> None:
|
||
try:
|
||
(self._dir(info.request_id) / "handoff.json").write_text(
|
||
json.dumps(handoff, ensure_ascii=False, indent=2), encoding="utf-8")
|
||
except OSError:
|
||
pass
|
||
|
||
def _remaining_cap() -> int:
|
||
return (token_cap - spent["in"] - spent["out"]) if token_cap else 1
|
||
|
||
async def _planner_json(user_msg: str) -> Dict[str, Any]:
|
||
"""调规划者并解析 JSON；解析失败回喂重试一次，再失败降级为 {}（禁止带病继续的软版本）。"""
|
||
import inspect
|
||
messages = [{"role": "system", "content": PLANNER_SYSTEM_PROMPT},
|
||
{"role": "user", "content": user_msg}]
|
||
content = ""
|
||
for attempt in (1, 2):
|
||
# 规划者同样流式（前端弱化展示其 JSON 草稿）
|
||
try:
|
||
accepts = len(inspect.signature(planner_chat).parameters) >= 3
|
||
except (TypeError, ValueError):
|
||
accepts = False
|
||
if accepts:
|
||
resp = await planner_chat(messages, [], throttle.make_cb("planner"))
|
||
else:
|
||
resp = await planner_chat(messages, [])
|
||
_account(resp.get("usage"))
|
||
content = resp.get("content") or ""
|
||
throttle.flush("planner")
|
||
obj = _parse_json_loose(content)
|
||
if obj:
|
||
break
|
||
if attempt == 1:
|
||
messages += [{"role": "assistant", "content": content},
|
||
{"role": "user",
|
||
"content": "你的输出不是合法 JSON。请重新只输出合法 JSON 对象。"}]
|
||
self._append_event(info, {"type": "message", "role": "planner",
|
||
"content": content[:2000]})
|
||
return obj
|
||
|
||
try:
|
||
# ---- 阶段 1：规划（大模型拆解为执行者指令） ----
|
||
self._append_event(info, {"type": "phase", "phase": "plan", "model": info.model})
|
||
plan = await _planner_json(
|
||
f"用户任务：{info.task}\n\n"
|
||
"请产出给执行者的指令，仅输出符合如下结构的 JSON：\n"
|
||
+ json.dumps(_PLAN_SCHEMA_HINT, ensure_ascii=False))
|
||
instructions = (plan.get("instructions") or info.task).strip()
|
||
handoff["instructions"] = instructions
|
||
handoff["acceptance"] = str(plan.get("acceptance", ""))
|
||
_save_handoff()
|
||
|
||
final_text = ""
|
||
reason = "answer"
|
||
error = None
|
||
exec_rounds_total = 0
|
||
|
||
# ---- 阶段 2/3：执行 <-> 审查（有界交接） ----
|
||
for h in range(1, max_handoffs + 1):
|
||
# 执行（本地小模型跑工具轮）
|
||
self._append_event(info, {"type": "phase", "phase": "execute",
|
||
"handoff": h, "model": info.executor_model})
|
||
loop = ToolLoop(tools, executor_chat, max_rounds=max_rounds,
|
||
token_cap=max(1, _remaining_cap()),
|
||
on_event=self._make_event_writer(info),
|
||
emit_final=False,
|
||
approval_hook=approval_hook,
|
||
on_delta=throttle.make_cb("executor"))
|
||
exec_result = await loop.run(instructions, system=EXECUTOR_SYSTEM_PROMPT)
|
||
_account({"prompt_tokens": exec_result.get("prompt_tokens", 0),
|
||
"completion_tokens": exec_result.get("completion_tokens", 0)})
|
||
exec_rounds_total += int(exec_result.get("rounds", 0))
|
||
report = exec_result.get("response", "")
|
||
# 执行者汇报作为消息事件透出（前端可读）
|
||
self._append_event(info, {"type": "message", "role": "executor",
|
||
"handoff": h, "content": (report or "")[:4000]})
|
||
if exec_result.get("reason") == "error":
|
||
reason, error = "error", exec_result.get("error")
|
||
final_text = report
|
||
break
|
||
|
||
# 审查（大模型裁决）
|
||
self._append_event(info, {"type": "phase", "phase": "review",
|
||
"handoff": h, "model": info.model})
|
||
review = await _planner_json(
|
||
f"用户任务：{info.task}\n你之前给出的指令：{instructions}\n"
|
||
f"验收标准：{handoff['acceptance'] or '（未明确）'}\n\n"
|
||
f"执行者第 {h} 轮汇报：\n{report[:4000]}\n\n"
|
||
"请审查是否已按验收标准完成，仅输出符合如下结构的 JSON：\n"
|
||
+ json.dumps(_REVIEW_SCHEMA_HINT, ensure_ascii=False))
|
||
verdict = str(review.get("verdict", "done")).lower()
|
||
handoff["exchanges"].append({
|
||
"handoff": h, "executor_report": report,
|
||
"verdict": verdict,
|
||
"reply_to_executor": str(review.get("reply_to_executor", "")),
|
||
})
|
||
_save_handoff()
|
||
|
||
if verdict == "done":
|
||
final_text = str(review.get("final_answer") or report)
|
||
break
|
||
# redo：裁决意见作为下一轮执行者指令（带上一轮上下文）
|
||
instructions = str(review.get("reply_to_executor") or instructions)
|
||
if h == max_handoffs:
|
||
reason = "max_handoffs"
|
||
error = f"交接轮数达上限（{max_handoffs}），以执行者汇报收尾"
|
||
final_text = report
|
||
else:
|
||
final_text = final_text or ""
|
||
|
||
self._append_event(info, {"type": "final", "round": total_rounds + exec_rounds_total,
|
||
"reason": reason, "error": error})
|
||
return {"response": final_text, "rounds": total_rounds + exec_rounds_total,
|
||
"reason": reason, "error": error,
|
||
"prompt_tokens": spent["in"], "completion_tokens": spent["out"]}
|
||
except Exception as exc:
|
||
reason = "error"
|
||
error = f"{type(exc).__name__}: {exc}"
|
||
self._append_event(info, {"type": "final", "round": total_rounds,
|
||
"reason": reason, "error": error})
|
||
return {"response": "", "rounds": total_rounds, "reason": reason,
|
||
"error": error,
|
||
"prompt_tokens": spent["in"], "completion_tokens": spent["out"]}
|
||
|
||
# ---------- 事件 ----------
|
||
def _make_event_writer(self, info: AgentRunInfo):
|
||
def _on_event(ev: Dict[str, Any]) -> None:
|
||
if ev.get("type") == "tool_call":
|
||
info.tool_calls += 1 # 工具步数统计（单/两级模式统一在此）
|
||
self._append_event(info, ev)
|
||
return _on_event
|
||
|
||
def _append_event(self, info: AgentRunInfo, ev: Dict[str, Any]) -> None:
|
||
ev = {"ts": time.time(), **ev}
|
||
try:
|
||
with self.events_path(info.request_id).open("a", encoding="utf-8") as f:
|
||
f.write(json.dumps(ev, ensure_ascii=False) + "\n")
|
||
except OSError:
|
||
pass
|
||
|
||
def read_events(self, request_id: str) -> List[Dict[str, Any]]:
|
||
p = self.events_path(request_id)
|
||
if not p.exists():
|
||
return []
|
||
out = []
|
||
for line in p.read_text(encoding="utf-8").splitlines():
|
||
line = line.strip()
|
||
if not line:
|
||
continue
|
||
try:
|
||
out.append(json.loads(line))
|
||
except json.JSONDecodeError:
|
||
pass # 半行（正在写入）忽略
|
||
return out
|
||
|
||
# ---------- 状态 ----------
|
||
def _write_status(self, info: AgentRunInfo) -> None:
|
||
try:
|
||
self.status_path(info.request_id).write_text(
|
||
json.dumps(info.to_dict(), ensure_ascii=False, indent=2),
|
||
encoding="utf-8")
|
||
except OSError:
|
||
pass
|
||
|
||
async def watch_events(self, request_id: str, cancel_event: asyncio.Event,
|
||
poll_interval: float = 0.3, max_seconds: float = 900.0):
|
||
"""SSE 生成器：增量推送 events.jsonl 新行，直到终态/取消/超时。
|
||
|
||
从文件头开始回放（晚加入的订阅者也能看到完整过程）。
|
||
"""
|
||
p = self.events_path(request_id)
|
||
offset = 0
|
||
deadline = time.time() + max_seconds
|
||
while not cancel_event.is_set() and time.time() < deadline:
|
||
if p.exists():
|
||
try:
|
||
size = p.stat().st_size
|
||
if size > offset:
|
||
with p.open("r", encoding="utf-8") as f:
|
||
f.seek(offset)
|
||
new_text = f.read()
|
||
offset = f.tell()
|
||
for line in new_text.splitlines():
|
||
line = line.strip()
|
||
if not line:
|
||
continue
|
||
try:
|
||
ev = json.loads(line)
|
||
except json.JSONDecodeError:
|
||
continue
|
||
yield ev
|
||
if ev.get("type") == "final":
|
||
return
|
||
except OSError:
|
||
pass
|
||
info = self.get(request_id)
|
||
if info and info.state in (STATE_DONE, STATE_FAILED):
|
||
# 终态兜底：状态已结束但可能没有 final 事件（如注册即失败）
|
||
yield {"type": "final", "round": info.rounds,
|
||
"reason": "answer" if info.state == STATE_DONE else "error",
|
||
"error": info.error}
|
||
return
|
||
await asyncio.sleep(poll_interval)
|
||
yield {"type": "final", "round": 0, "reason": "error", "error": "订阅超时"}
|
||
|
||
|
||
# ---------- 全局单例 ----------
|
||
_service: Optional[AgentService] = None
|
||
|
||
|
||
def get_agent_service() -> AgentService:
|
||
global _service
|
||
if _service is None:
|
||
_service = AgentService()
|
||
return _service
|
||
|
||
|
||
def reset_agent_service() -> None:
|
||
"""测试用：重置全局智能体服务单例。"""
|
||
global _service
|
||
_service = None
|
||
|
||
|
||
def new_request_id() -> str:
|
||
return "ag" + uuid.uuid4().hex[:10]
|
||
|
||
|
||
# ─────────────────────────────────────────────────────────────────────────────
|
||
# 审批流（D9）：dsh 式 allow-once / deny，fail-closed
|
||
# ─────────────────────────────────────────────────────────────────────────────
|
||
READ_ONLY_TOOLS = {"list_dir", "read_file", "search_files", "web_fetch"}
|
||
|
||
|
||
def needs_approval(policy: str, tool_name: str) -> bool:
|
||
"""审批策略判定：off=全放行；all=全询问；dangerous=写/编辑/命令询问，只读放行。"""
|
||
if policy == "all":
|
||
return True
|
||
if policy == "dangerous":
|
||
return tool_name not in READ_ONLY_TOOLS
|
||
return False
|
||
|
||
|
||
class DeltaThrottle:
|
||
"""流式增量节流（D10）：积攒超过阈值才落一条 delta 事件，防事件爆炸。"""
|
||
|
||
THRESHOLD = 48
|
||
|
||
def __init__(self, append_event):
|
||
self._append = append_event # (ev: dict) -> None
|
||
self._buf: Dict[str, str] = {}
|
||
|
||
def make_cb(self, role: str):
|
||
def cb(text: str) -> None:
|
||
self.add(role, text)
|
||
return cb
|
||
|
||
def add(self, role: str, text: str) -> None:
|
||
buf = self._buf.get(role, "") + (text or "")
|
||
if len(buf) >= self.THRESHOLD:
|
||
self._flush(role, buf)
|
||
buf = ""
|
||
self._buf[role] = buf
|
||
|
||
def flush(self, role: Optional[str] = None) -> None:
|
||
roles = [role] if role else list(self._buf.keys())
|
||
for r in roles:
|
||
buf = self._buf.get(r, "")
|
||
if buf:
|
||
self._flush(r, buf)
|
||
self._buf[r] = ""
|
||
|
||
def _flush(self, role: str, text: str) -> None:
|
||
self._append({"type": "delta", "role": role, "text": text})
|
||
|
||
|
||
class ApprovalManager:
|
||
"""单次智能体运行内的审批挂起/裁决（asyncio Event 实现，dsh 式 allow-once）。"""
|
||
|
||
def __init__(self):
|
||
self._pending: Dict[str, Dict[str, Any]] = {}
|
||
|
||
def open(self, approval_id: str) -> asyncio.Event:
|
||
ev = asyncio.Event()
|
||
self._pending[approval_id] = {"event": ev, "allowed": False}
|
||
return ev
|
||
|
||
def decide(self, approval_id: str, allowed: bool) -> bool:
|
||
p = self._pending.get(approval_id)
|
||
if p is None:
|
||
return False
|
||
p["allowed"] = allowed
|
||
p["event"].set()
|
||
return True
|
||
|
||
def close(self, approval_id: str) -> None:
|
||
self._pending.pop(approval_id, None)
|
||
|
||
|
||
# ─────────────────────────────────────────────────────────────────────────────
|
||
# 会话（dsh 式：工作区内多轮对话，持久化到磁盘）
|
||
# ─────────────────────────────────────────────────────────────────────────────
|
||
SESSIONS_DIR = Path("agent_runs") / "sessions"
|
||
|
||
|
||
class AgentSession:
|
||
"""一个智能体会话：多轮任务 + 配置快照（磁盘持久化）。"""
|
||
|
||
def __init__(self, data: Dict[str, Any]):
|
||
self.data = data
|
||
|
||
@classmethod
|
||
def new(cls, sid: str, title: str, workspace: str,
|
||
pool_id: str = "", executor_pool_id: str = "") -> "AgentSession":
|
||
now = time.time()
|
||
return cls({
|
||
"id": sid, "title": title[:24] or "新会话", "workspace": workspace,
|
||
"pool_id": pool_id, "executor_pool_id": executor_pool_id,
|
||
"created_at": now, "updated_at": now, "busy": False,
|
||
"turns": [], # [{request_id, task, response, state, tool_calls, tokens}]
|
||
})
|
||
|
||
def to_dict(self) -> Dict[str, Any]:
|
||
return dict(self.data)
|
||
|
||
def view(self, include_turns: bool = True) -> Dict[str, Any]:
|
||
out = self.to_dict()
|
||
if not include_turns:
|
||
out["turns"] = len(self.data.get("turns", []))
|
||
return out
|
||
|
||
|
||
class SessionStore:
|
||
"""会话注册表（内存索引 + sessions/{sid}.json 持久化）。"""
|
||
|
||
def __init__(self, root: Path = SESSIONS_DIR):
|
||
self.root = Path(root)
|
||
self.root.mkdir(parents=True, exist_ok=True)
|
||
self._cache: Dict[str, AgentSession] = {}
|
||
|
||
def _path(self, sid: str) -> Path:
|
||
return self.root / f"{sid}.json"
|
||
|
||
def create(self, title: str, workspace: str,
|
||
pool_id: str = "", executor_pool_id: str = "") -> AgentSession:
|
||
sid = "as" + uuid.uuid4().hex[:10]
|
||
sess = AgentSession.new(sid, title or "新会话", workspace, pool_id, executor_pool_id)
|
||
self._cache[sid] = sess
|
||
self._save(sess)
|
||
return sess
|
||
|
||
def get(self, sid: str) -> Optional[AgentSession]:
|
||
if sid in self._cache:
|
||
return self._cache[sid]
|
||
p = self._path(sid)
|
||
if not p.exists():
|
||
return None
|
||
try:
|
||
sess = AgentSession(json.loads(p.read_text(encoding="utf-8")))
|
||
self._cache[sid] = sess
|
||
return sess
|
||
except (json.JSONDecodeError, OSError):
|
||
return None
|
||
|
||
def list(self) -> List[Dict[str, Any]]:
|
||
out = []
|
||
for p in sorted(self.root.glob("*.json"),
|
||
key=lambda x: x.stat().st_mtime, reverse=True):
|
||
try:
|
||
out.append(json.loads(p.read_text(encoding="utf-8")))
|
||
except (json.JSONDecodeError, OSError):
|
||
continue
|
||
return out
|
||
|
||
def delete(self, sid: str) -> bool:
|
||
self._cache.pop(sid, None)
|
||
p = self._path(sid)
|
||
if p.exists():
|
||
p.unlink()
|
||
return True
|
||
return False
|
||
|
||
def rename(self, sid: str, title: str) -> Optional[AgentSession]:
|
||
"""重命名会话标题（dsh session.rename 对齐）。"""
|
||
sess = self.get(sid)
|
||
if sess is None:
|
||
return None
|
||
title = (title or "").strip()
|
||
if not title:
|
||
return sess
|
||
sess.data["title"] = title[:24]
|
||
self.save(sess)
|
||
return sess
|
||
|
||
def save(self, sess: AgentSession) -> None:
|
||
self._cache[sess.data["id"]] = sess
|
||
self._save(sess)
|
||
|
||
def _save(self, sess: AgentSession) -> None:
|
||
sess.data["updated_at"] = time.time()
|
||
try:
|
||
self._path(sess.data["id"]).write_text(
|
||
json.dumps(sess.data, ensure_ascii=False, indent=2), encoding="utf-8")
|
||
except OSError:
|
||
pass
|
||
|
||
|
||
_session_store: Optional[SessionStore] = None
|
||
|
||
|
||
def get_session_store() -> SessionStore:
|
||
global _session_store
|
||
if _session_store is None:
|
||
_session_store = SessionStore()
|
||
return _session_store
|
||
|
||
|
||
def reset_session_store() -> None:
|
||
"""测试用。"""
|
||
global _session_store
|
||
_session_store = None
|