QihuAI Design
04a0641202
#876 OpenTelemetry 可观测（双 Claw 标签）： - app/core/otel.py: 新增 set_otel_ctx/get_otel_ctx(ContextVar 三元组 tenant_id+claw_type+ trace_id)；current_span_cm 自动注入双 Claw 强制标签，全部 13 个 span 调用点零签名侵入带维度； record_span_metrics 指标维度补 tenant_id+claw_type（trace_id 因高基数不计入指标标签）。 - app/core/llm_router.py: set_route_ctx 桥接转调 set_otel_ctx（单次入口注入双端共享： 路由上下文 + OTel 标签），失败安全。 - app/api/chat.py: send_message 与 send_message_stream 入口注入 set_route_ctx(web)， stream 经 asyncio.create_task 子协程自动传播 contextvars，使 _send_message_core 内 所有 span 均带 tenant_id+claw_type。 - app/api/im_llm.py: _call_real_llm 入口最早注入 set_route_ctx(im_wechat)，并显式开启 im.send_message span（此前 IM Claw 完全无 span，Grafana 无 IM 数据点），finally 关闭。 - tests/test_dual_claw.py: R-DUAL-6 由 skip 转为确定性断言（Web/IM 双 Claw span 必带 tenant_id+claw_type），作为持续哨兵。 - 交付 Grafana 双 Claw 对比看板 JSON（aiflow_span_* / aiflow_llm_route_total 按 claw_type 分组：P50/P95 延迟、成功率、QPS、租户健康 Top10）。Alloy OTLP 仅桥接 metrics→Prometheus， trace 无 pipeline（#800 已压制 UNIMPLEMENTED 噪声），看板基于 metrics 出图。 safe-align: 仅复制 5 个 git-tracked 文件，未触碰 secrets/venv/node_modules/caches。本地提交不 push。
532 lines
27 KiB
Python
532 lines
27 KiB
Python
"""
|
||
IM Bot 真实 LLM 调用。
|
||
|
||
原实现位于 channels.py（通用频道管理模块），该模块已移除。
|
||
为保持 IM Bot（微信机器人等）正常生成回复，将 _call_real_llm 迁移至此独立模块。
|
||
|
||
行为（对齐 AI 聊天 — 2026-07-23 同步至最新）：
|
||
- Issue 7：图像/视频/语音意图优先走平台媒体模型（与 AI 聊天同一套 Auto 路由），
|
||
不再对"制作图片"这类请求只回文本。
|
||
- Issue 8：默认 Auto 模型；每轮先自动发现本账号已安装且平台已部署的技能，
|
||
多个适用时由模型自行判断选 1 个最优，通过 function calling 执行。
|
||
- 2026-07-15：修复 trust_env=False + proxy 处理，避免容器坏代理导致调用失败；
|
||
媒体生成前先经 _optimize_media_prompt 精炼提示词，提升 AGNES 生成质量。
|
||
- 2026-07-23：对齐 Web Chat 最新能力：
|
||
* 模型 failover（Auto 模式自动切换备选模型）
|
||
* 平台知识库注入（get_platform_knowledge，技能清单实时同步）
|
||
"""
|
||
import asyncio
|
||
import os
|
||
import json
|
||
import logging
|
||
import re as _re
|
||
from datetime import datetime, timezone, timedelta
|
||
from typing import Optional
|
||
from urllib.parse import urlparse as _urlparse
|
||
|
||
import httpx
|
||
from app.core.llm_client import chat_complete # 统一模型层(07-24): web/IM 共用标准 OpenAI SDK 客户端
|
||
from sqlalchemy.orm import Session
|
||
|
||
from ..capability_registry import resolve_auto_model
|
||
from ..models import AIModel, ModelProvider, TokenUsageLog, gen_uuid
|
||
from app.core.observability import get_trace_id, span
|
||
# #875 LLM Router：统一模型路由层（影子模式先行，收敛 litellm 主密钥硬编码）
|
||
from ..core.llm_router import (
|
||
route as _llm_route,
|
||
set_route_ctx as _set_route_ctx,
|
||
SHADOW as _LLM_ROUTER_SHADOW,
|
||
resolve_api_key as _llm_resolve_api_key,
|
||
)
|
||
|
||
logger = logging.getLogger(__name__)
|
||
|
||
|
||
# 香港出网代理(实测对 AGNES 延迟更低，2026-07-15 用户实测)，与 chat.py 同值。
|
||
_HK_PROXY = "http://qyyfw:XHR521yhf@103.242.14.208:11495"
|
||
|
||
|
||
def _proxy_for_url(url: str) -> str | None:
|
||
"""按目标域名选择是否走代理（与 chat.py 同逻辑）。
|
||
|
||
2026-07-15 调整：AGNES 改走香港代理(实测延迟更低)，不再直连。
|
||
国内域名(尊重 NO_PROXY)直连，更稳定。
|
||
"""
|
||
import os
|
||
host = (_urlparse(url).hostname or "").lower()
|
||
if "agnes-ai.com" in host or "agnes-ai.space" in host:
|
||
return _HK_PROXY
|
||
# 2026-07-30 修复 3+: 不代理 Docker 内部 hostname 和 localhost
|
||
if "qihuyun.com" in host or "localhost" in host or host.startswith("127.") or "qihu-aiflow" in host:
|
||
return None
|
||
_no_proxy = [h.strip().lower() for h in os.environ.get("NO_PROXY", "").split(",") if h.strip()]
|
||
for _np in _no_proxy:
|
||
if host == _np or host.endswith("." + _np) or _np in host:
|
||
return None
|
||
return os.environ.get("HTTPS_PROXY") or os.environ.get("HTTP_PROXY") or "http://172.18.0.1:7890"
|
||
|
||
|
||
_DEBUG_LOG_PATH = "/tmp/im_llm_debug.log"
|
||
|
||
def _debug_log(msg: str):
|
||
"""写磁盘调试日志（不依赖 uvicorn logging 配置）。"""
|
||
try:
|
||
with open(_DEBUG_LOG_PATH, "a", encoding="utf-8") as _f:
|
||
_f.write(datetime.now(timezone(timedelta(hours=8))).isoformat() + " " + msg + "\n")
|
||
except Exception:
|
||
pass
|
||
|
||
|
||
def _run_async(coro, timeout=None):
|
||
"""在同步上下文中运行协程，兼容「无事件循环」与「已有事件循环」两种场景。
|
||
|
||
IM 个人微信在独立线程中调用（无事件循环）→ asyncio.run；
|
||
IM 多平台 Webhook 可能在 FastAPI 协程内调用（已有事件循环）→ 丢到线程池执行。
|
||
|
||
timeout: 可选，秒数；超时抛出 asyncio.TimeoutError（由调用方捕获）。
|
||
"""
|
||
try:
|
||
loop = asyncio.get_running_loop()
|
||
except RuntimeError:
|
||
loop = None
|
||
|
||
if timeout is not None:
|
||
async def _timed():
|
||
return await asyncio.wait_for(coro, timeout=timeout)
|
||
target = _timed()
|
||
else:
|
||
target = coro
|
||
|
||
if loop is not None and loop.is_running():
|
||
import concurrent.futures
|
||
with concurrent.futures.ThreadPoolExecutor(1) as ex:
|
||
return ex.submit(lambda: asyncio.run(target)).result(timeout=(timeout + 10) if timeout else None)
|
||
return asyncio.run(target)
|
||
|
||
|
||
def _log_media_usage(db: Session, user_id: str, ai_model, media_type: str, prompt: str):
|
||
"""IM Bot 媒体生成记账：把图片/视频消耗写入 TokenUsageLog，供驾驶舱统计。
|
||
|
||
此前 IM Bot 路径完全不写 TokenUsageLog，导致微信生成的图片/视频在驾驶舱里"查无此项"。
|
||
媒体 API 不返回 token 用量，这里用与 chat.py _generate_media_task 一致的估算。
|
||
"""
|
||
if not user_id or not ai_model:
|
||
return
|
||
try:
|
||
base = 1200 if media_type == "video" else 400
|
||
est = base + (len(prompt or "") // 4)
|
||
db.add(TokenUsageLog(
|
||
id=gen_uuid(),
|
||
user_id=user_id,
|
||
model_id=ai_model.id,
|
||
model_name=ai_model.name,
|
||
trace_id=get_trace_id(),
|
||
prompt_tokens=est,
|
||
completion_tokens=0,
|
||
total_tokens=est,
|
||
created_at=datetime.now(timezone(timedelta(hours=8))).replace(tzinfo=None),
|
||
))
|
||
db.commit()
|
||
except Exception as e:
|
||
try:
|
||
db.rollback()
|
||
except Exception:
|
||
pass
|
||
logger.warning(f"[im_llm] 媒体 Token 记账失败(忽略): {e}")
|
||
|
||
|
||
def _call_real_llm(db: Session, user=None, prompt: str = "", platform: str = "im", user_id: str = None) -> "str | dict":
|
||
"""调用七虎AI 生成回复，供 IM Bot 复用。支持媒体意图路由、已安装技能自动发现、模型 failover。"""
|
||
# #876 双 Claw 维度：入口最早注入 tenant_id + claw_type，确保下方所有 span / metric
|
||
# 均带双 Claw 标签（此前 IM Claw 完全无 span，Grafana 无 IM 数据点）。
|
||
_set_route_ctx(tenant_id=getattr(user, "tenant_id", "") or "", claw_type="im_wechat")
|
||
# #876：IM Claw 此前完全无 span，Grafana 无 IM 数据点。此处显式开启一个 im.send_message
|
||
# span（手动 enter/exit 以避免整段重缩进），使 IM 链路在双 Claw 对比看板出图。
|
||
_im_span = span("im.send_message", platform=platform, user_id=user_id)
|
||
_im_span.__enter__()
|
||
try:
|
||
from app.api.chat import (
|
||
_generate_media,
|
||
detect_media_intent,
|
||
_discover_installed_skills,
|
||
_build_system_prompt,
|
||
_execute_skill_tool,
|
||
_list_enabled_mcp_servers,
|
||
)
|
||
# 2026-08-04 双 Claw 对齐：能力发现收敛为唯一实现，Web / IM 共用
|
||
from app.core.capability_discovery import (
|
||
build_mcp_tool_defs,
|
||
discover_mcp_servers,
|
||
discover_skills,
|
||
)
|
||
from app.mcp_client import call_mcp_tool
|
||
|
||
# ===== 1) 媒体意图路由（Issue 7）=====
|
||
# 2026-07-30 fix: 只对当前用户消息做媒体意图检测，避免历史中的图像请求污染
|
||
_cur_msg = prompt
|
||
_marker = "用户当前问题："
|
||
if _marker in prompt:
|
||
_cur_msg = prompt[prompt.rfind(_marker) + len(_marker):].strip()
|
||
intent = detect_media_intent(_cur_msg)
|
||
if intent:
|
||
ai_model = resolve_auto_model(intent, db)
|
||
if ai_model:
|
||
try:
|
||
result = _run_async(_generate_media(db, ai_model, prompt, user_id), timeout=300)
|
||
except Exception:
|
||
# 任何未预期异常(如 AGNES 连接重置)统一兜底为友好提示，
|
||
# 绝不直接把原始异常(如 "Server disconnected")甩给用户
|
||
result = "⚠️ 媒体生成暂时失败（连接或服务波动），请稍后重试。"
|
||
if result:
|
||
# 媒体生成成功：返回结构化对象，交由 im_bot_user 在微信内联发送图片/视频
|
||
s = result.strip()
|
||
if s.startswith("<img") or s.startswith("<video"):
|
||
m = _re.search(r'src="([^"]+)"', s)
|
||
url = m.group(1) if m else ""
|
||
media_type = "image" if s.startswith("<img") else "video"
|
||
if not url:
|
||
return "（生成完成，但未能提取媒体地址）"
|
||
# IM Bot 路径此前完全不记账 → 补写媒体消耗，驾驶舱才看得到图片/视频数据
|
||
_log_media_usage(db, user_id, ai_model, media_type, prompt)
|
||
return {"__media__": True, "media_type": media_type, "url": url}
|
||
# 媒体生成失败：_generate_media 已返回友好中文（如限频/队列繁忙），
|
||
# 直接透传，绝不把原始提示词/标签/JSON 透传给用户
|
||
if s.startswith("⚠️") or s.startswith("（媒体"):
|
||
return s
|
||
return "（媒体生成未能返回有效内容，请重试）"
|
||
|
||
# ===== 1.5) EngineCore 薄壳（#873 · 灰阶门，默认关闭）=====
|
||
# IM_USE_ENGINE_CORE=1 时，文本回复改走与 Web Claw 同一份 EngineCore
|
||
# (_send_message_core)，经 BufferAdapter 缓冲装配为整条回复；上方媒体快路径
|
||
# 已提前返回，不受此门影响。默认 0/未设置 = 走下方既有自包含逻辑（零改动）。
|
||
if os.environ.get("IM_USE_ENGINE_CORE") == "1":
|
||
return _run_async(
|
||
_call_engine_core(db, user_id, prompt, platform, user),
|
||
timeout=300,
|
||
)
|
||
|
||
# ===== 2) 文本回复：模型 failover + 技能自动发现 + 平台知识库注入 =====
|
||
from ..platform_knowledge import get_platform_knowledge
|
||
|
||
# 构建候选模型列表（对齐 Web Chat failover）
|
||
ai_model = resolve_auto_model("text", db)
|
||
_try_models = [ai_model] if (ai_model and ai_model.is_enabled) else []
|
||
# Auto 模式附加备选（IM Bot 始终 Auto）
|
||
# 2026-07-31 根因修复：仅选用「纯文本」模型作为 failover 备选。
|
||
# 此前按 is_enabled 顺序取前 4 个，会混入图像/视频/语音(ASR)模型
|
||
# （这些模型 capability 含 "text" 但本质是媒体模型，收到文本 prompt 返回空内容、
|
||
# 不抛异常 → 被误判为「模型失败」，最终导致「已尝试N个模型均调用失败」）。
|
||
_MEDIA_CAPS = {"image", "video", "audio", "asr", "embedding", "rerank", "ocr"}
|
||
def _is_text_only(m):
|
||
_caps = m.capabilities or []
|
||
return "text" in _caps and not any(c in _MEDIA_CAPS for c in _caps)
|
||
_fallbacks = db.query(AIModel).filter(
|
||
AIModel.is_enabled == True,
|
||
AIModel.id != (ai_model.id if ai_model else ""),
|
||
).order_by(AIModel.is_default.desc()).all()
|
||
_fallbacks = [m for m in _fallbacks if _is_text_only(m)]
|
||
_try_models.extend([m for m in _fallbacks[:4] if m not in _try_models])
|
||
|
||
# ===== LLM Router（#875 影子模式）=====
|
||
# 计算决策并打 claw_type 标签 metrics；影子模式下仅观测，不改流量。
|
||
# 切主路径：LLM_ROUTER_SHADOW=0 时改用 Router 候选链（已含文本过滤，等同此处逻辑）。
|
||
try:
|
||
_rm = _llm_route("text", db=db, legacy_primary=_try_models[0].name if _try_models else None)
|
||
if not _LLM_ROUTER_SHADOW and _rm.models:
|
||
_try_models = _rm.models
|
||
_debug_log(f"ROUTER_LIVE models={[m.name for m in _try_models]}")
|
||
except Exception as _e:
|
||
logger.warning(f"[im_llm] LLM Router 观测失败(降级忽略): {_e}")
|
||
|
||
_debug_log(f"TRY_MODELS user={user_id} count={len(_try_models)} names={[m.name for m in _try_models]}")
|
||
|
||
if not _try_models:
|
||
return "（未配置可用文本模型，无法生成回复）"
|
||
|
||
# 能力自动发现（提前执行一次，不必每个候选模型重复做）
|
||
# 2026-08-04 根因修复：
|
||
# 旧实现 `_discover_installed_skills(db, user_id)[:15]` 强依赖
|
||
# app_installations 表，线上该表 0 行 → 97 个 published skill
|
||
# 在 Web / IM 双端同时发现不到；`[:15]` 截断的是一个空列表。
|
||
# 现改走统一发现层（安装优先 → published 降级 → 相关性裁剪）。
|
||
_query_hint = _cur_msg or prompt
|
||
skills = []
|
||
try:
|
||
skills = discover_skills(db, user_id, query=_query_hint, claw="im")
|
||
except Exception as _e:
|
||
logger.warning(f"[im_llm] 技能发现失败(降级为空): {_e}")
|
||
skills = []
|
||
|
||
# MCP 自动挂载：IM Bot Claw 没有勾选 UI，自动发现即其等价通道。
|
||
# （Web Claw 靠用户在界面勾选 → effective_mcp_ids）
|
||
mcp_servers = []
|
||
mcp_function_definitions = []
|
||
mcp_tool_map = {}
|
||
try:
|
||
mcp_servers = discover_mcp_servers(db, None, claw="im")
|
||
if mcp_servers:
|
||
mcp_function_definitions, mcp_tool_map = _run_async(
|
||
build_mcp_tool_defs(mcp_servers), timeout=45
|
||
)
|
||
except Exception as _e:
|
||
logger.warning(f"[im_llm] MCP 发现失败(降级为空): {_e}")
|
||
mcp_servers, mcp_function_definitions, mcp_tool_map = [], [], {}
|
||
|
||
_debug_log(
|
||
f"CAP skills={len(skills)} mcp_servers={len(mcp_servers)} "
|
||
f"mcp_tools={len(mcp_function_definitions)}"
|
||
)
|
||
|
||
# 平台知识库注入（对齐 Web Chat——微信用户提问"平台有哪些技能"也能得到完整清单）
|
||
kb_context = ""
|
||
try:
|
||
pk = get_platform_knowledge(db=db)
|
||
if pk:
|
||
kb_context = f"【平台知识库——优先使用以下知识回答用户问题】\n{pk}"
|
||
except Exception:
|
||
pass
|
||
|
||
# 尝试每个候选模型
|
||
_retry_log = []
|
||
for _try_model in _try_models:
|
||
ai_model = _try_model
|
||
|
||
provider = db.query(ModelProvider).filter(ModelProvider.id == ai_model.provider_id).first()
|
||
api_key = ai_model.api_key or (provider.api_key if provider else None)
|
||
api_url = ai_model.endpoint
|
||
# 2026-07-30: LiteLLM 端点统一用 master key（收敛进 LLM Router，两端唯一来源）
|
||
api_key = _llm_resolve_api_key(api_url, api_key, provider.api_key if provider else None)
|
||
if not api_url and provider and provider.api_base_url:
|
||
api_url = f"{provider.api_base_url.rstrip('/')}/v1/chat/completions"
|
||
if not (api_key and api_url):
|
||
_retry_log.append(f"{ai_model.name}: 缺少密钥或端点")
|
||
continue
|
||
|
||
# agents 暂留空：线上 market_apps 中 type='agent' 记录为 0，
|
||
# 且 IM 侧尚无 agent 执行分支；待薄壳化接入 EngineCore 后自动获得。
|
||
system_prompt, function_definitions = _build_system_prompt(
|
||
skills=skills,
|
||
agents=[],
|
||
kb_context=kb_context,
|
||
user_content=prompt,
|
||
mcp=mcp_servers,
|
||
mcp_functions=mcp_function_definitions,
|
||
auto_discovered=bool(skills),
|
||
all_mcp=_list_enabled_mcp_servers(db),
|
||
db=db,
|
||
)
|
||
messages = [
|
||
{"role": "system", "content": system_prompt},
|
||
{"role": "user", "content": prompt},
|
||
]
|
||
|
||
# 尝试调用
|
||
try:
|
||
max_rounds = 3
|
||
ai_content = ""
|
||
_http_ok = True
|
||
for round_num in range(max_rounds):
|
||
request_body = {"model": ai_model.name, "messages": messages, "max_tokens": 2048, "enable_thinking": False, "temperature": 0.7}
|
||
if function_definitions:
|
||
request_body["tools"] = [
|
||
{"type": "function", "function": fd} for fd in function_definitions
|
||
]
|
||
request_body["tool_choice"] = "auto"
|
||
try:
|
||
import logging as _l5
|
||
_l5.warning(f'[im_llm REQ] {ai_model.name}: url={api_url[:60]}, proxy={_proxy_for_url(api_url)}, api_key={api_key[:20]}..., messages_count={len(request_body.get("messages", []))}, tools_count={len(request_body.get("tools", []))}')
|
||
_debug_log(f"REQ model={ai_model.name} url={api_url} proxy={_proxy_for_url(api_url)} body={json.dumps(request_body, ensure_ascii=False)[:600]}")
|
||
data = chat_complete(
|
||
api_url,
|
||
api_key,
|
||
request_body,
|
||
proxy=_proxy_for_url(api_url),
|
||
timeout=ai_model.timeout or 120,
|
||
)
|
||
_l5.warning(f'[im_llm RES] {ai_model.name}: data_keys={list(data.keys()) if data else None}, resp_type={type(data).__name__}')
|
||
_debug_log(f"RES model={ai_model.name} data={json.dumps(data, ensure_ascii=False)[:600]}")
|
||
except Exception as _exc:
|
||
_retry_log.append(f"{ai_model.name}: {str(_exc)[:60]}")
|
||
import traceback as _tb
|
||
logger.warning(f"[im_llm] 模型 {ai_model.name} 异常: {_exc}")
|
||
logger.warning(f"[im_llm] {ai_model.name} 异常详情: {_tb.format_exc()[:2000]}")
|
||
_debug_log(f"EXC model={ai_model.name} err={str(_exc)[:200]} trace={_tb.format_exc()[:500]}")
|
||
_http_ok = False
|
||
break
|
||
message = data.get("choices", [{}])[0].get("message", {})
|
||
tool_calls = message.get("tool_calls", [])
|
||
# 2026-07-30 CRITICAL DEBUG
|
||
import logging as _lcr
|
||
_raw_content = message.get("content")
|
||
_raw_reasoning = message.get("reasoning_content")
|
||
_msg_keys = list(message.keys()) if message else []
|
||
_tool_calls = tool_calls
|
||
_lcr.warning(f'[CRITICAL] MODEL={ai_model.name} msg_keys={_msg_keys} content_len={len(_raw_content or "")} reasoning_len={len(_raw_reasoning or "")} has_tool_calls={bool(_tool_calls)}')
|
||
_debug_log(f"PARSE model={ai_model.name} content_len={len(_raw_content or '')} reasoning_len={len(_raw_reasoning or '')} content_preview={repr(_raw_content)[:120]}")
|
||
ai_content = message.get("content", "") or message.get("reasoning_content") or ""
|
||
if not ai_content and (message.get("reasoning_content") or ""):
|
||
ai_content = f"（AI 思考：{message.get('reasoning_content')[:500]}...）"
|
||
if tool_calls and (skills or mcp_tool_map):
|
||
messages.append({
|
||
"role": "assistant",
|
||
"content": message.get("content") or message.get("reasoning_content") or "",
|
||
"tool_calls": tool_calls,
|
||
})
|
||
for tc in tool_calls:
|
||
func_name = tc.get("function", {}).get("name", "")
|
||
try:
|
||
func_args = json.loads(tc.get("function", {}).get("arguments", "{}"))
|
||
except Exception:
|
||
func_args = {"input": tc.get("function", {}).get("arguments", "")}
|
||
if func_name.startswith("mcp__"):
|
||
# MCP 工具：命名规则与 Web Claw 完全一致
|
||
# mcp__{server}__{tool}__{idx}
|
||
_info = mcp_tool_map.get(func_name)
|
||
if not _info:
|
||
tool_result = (
|
||
"[MCP 工具未找到] 可用工具: "
|
||
+ str(list(mcp_tool_map.keys())[:5])
|
||
)
|
||
else:
|
||
_srv = next(
|
||
(s for s in mcp_servers
|
||
if s.get("id") == _info.get("server_id")),
|
||
None,
|
||
)
|
||
if not _srv:
|
||
tool_result = f"[MCP 服务未找到: {_info.get('server_name')}]"
|
||
else:
|
||
try:
|
||
tool_result = _run_async(
|
||
call_mcp_tool(_srv, _info["name"], func_args),
|
||
timeout=90,
|
||
)
|
||
except Exception as _mexc:
|
||
tool_result = (
|
||
f"[MCP 工具 {_info.get('name')} 执行失败："
|
||
f"{str(_mexc)[:160]}]"
|
||
)
|
||
_debug_log(
|
||
f"MCP_CALL name={func_name} "
|
||
f"result_len={len(tool_result or '')}"
|
||
)
|
||
else:
|
||
tool_result = _execute_skill_tool(db, func_name, func_args, skills, prompt)
|
||
messages.append({
|
||
"role": "tool",
|
||
"tool_call_id": tc.get("id", "call_1"),
|
||
"content": tool_result,
|
||
})
|
||
if round_num >= max_rounds - 1:
|
||
break
|
||
continue
|
||
break
|
||
|
||
# 检查本轮调用是否成功
|
||
import logging as _log2
|
||
_log2.warning(f'[im_llm trace] {ai_model.name}: http_ok={_http_ok}, ai_content_len={len(ai_content) if ai_content else 0}, ai_content_preview={repr(ai_content)[:200]}')
|
||
_debug_log(f"CHECK model={ai_model.name} http_ok={_http_ok} ai_content_len={len(ai_content) if ai_content else 0}")
|
||
if _http_ok and ai_content:
|
||
_debug_log(f"SUCCESS model={ai_model.name} returning len={len(ai_content)}")
|
||
return ai_content or "（AI 回复为空）"
|
||
# 本轮失败：记录日志，切换备选
|
||
logger.warning('[im_llm] 模型 %s 失败 (http_ok=%s, content=%s), 切换备选...' % (
|
||
ai_model.name, _http_ok, repr(ai_content)[:100]
|
||
))
|
||
_debug_log(f"FAIL model={ai_model.name} http_ok={_http_ok} content_preview={repr(ai_content)[:100]}")
|
||
continue
|
||
except Exception as exc:
|
||
_retry_log.append(f"{ai_model.name}: {str(exc)[:60]}")
|
||
logger.warning(f"[im_llm] 模型 {ai_model.name} 异常: {exc}")
|
||
_debug_log(f"OUTER_EXC model={ai_model.name} err={str(exc)[:120]}")
|
||
continue
|
||
|
||
# 全部失败
|
||
_summary = "; ".join(_retry_log[-3:])
|
||
_debug_log(f"ALL_FAILED try_models={len(_try_models)} retry_log={_retry_log}")
|
||
return f"（已尝试 {len(_try_models)} 个模型均调用失败。最后错误: {_summary}。请稍后重试。）"
|
||
except Exception as e:
|
||
return f"（AI 回复生成失败：{e}）"
|
||
finally:
|
||
# #876：关闭 im.send_message span（记录耗时 + 指标，强制带 tenant_id + claw_type）
|
||
try:
|
||
_im_span.__exit__(None, None, None)
|
||
except Exception:
|
||
pass
|
||
|
||
|
||
async def _call_engine_core(db, user_id, prompt: str, platform: str, user=None) -> "str | dict":
|
||
"""（#873 薄壳）复用 Web Claw 同一份 EngineCore 生成 IM 回复。
|
||
|
||
单引擎原则：IM Bot 不再维护第二份「模型 failover + 技能/MCP 执行」实现，
|
||
而是把已发现的 skill_ids/mcp_ids 交给 _send_message_core，经 BufferAdapter
|
||
缓冲装配成整条回复（媒体结构 __media__ 由 BufferAdapter 原样保留）。
|
||
默认不启用（见 _call_real_llm 中的 IM_USE_ENGINE_CORE 灰阶门）。
|
||
"""
|
||
from app.api.chat_service import _send_message_core
|
||
from app.api.chat_utils import SendMessageRequest
|
||
from app.core.claw_adapters import BufferAdapter
|
||
from ..models import Conversation, User
|
||
|
||
# 解析用户：IM 调用方常传 user=None，仅给 user_id
|
||
real_user = user
|
||
if real_user is None and user_id:
|
||
real_user = db.query(User).filter(User.id == user_id).first()
|
||
if real_user is None:
|
||
return "（EngineCore 模式：无法解析当前用户，请检查 IM 账号绑定）"
|
||
|
||
# 能力发现（与 IM 自包含逻辑同一套，claw="im"）
|
||
from app.core.capability_discovery import (
|
||
discover_mcp_servers,
|
||
discover_skills,
|
||
)
|
||
_query_hint = prompt
|
||
_marker = "用户当前问题："
|
||
if _marker in prompt:
|
||
_query_hint = prompt[prompt.rfind(_marker) + len(_marker):].strip()
|
||
skills = []
|
||
try:
|
||
skills = discover_skills(db, user_id, query=_query_hint, claw="im") or []
|
||
except Exception as _e:
|
||
logger.warning(f"[im_llm.engine] 技能发现失败(降级为空): {_e}")
|
||
mcp_servers = []
|
||
try:
|
||
mcp_servers = discover_mcp_servers(db, None, claw="im") or []
|
||
except Exception as _e:
|
||
logger.warning(f"[im_llm.engine] MCP 发现失败(降级为空): {_e}")
|
||
|
||
skill_ids = [
|
||
(s.get("id") if isinstance(s, dict) else getattr(s, "id", None))
|
||
for s in skills
|
||
]
|
||
skill_ids = [s for s in skill_ids if s]
|
||
mcp_ids = [
|
||
(m.get("id") if isinstance(m, dict) else getattr(m, "id", None))
|
||
for m in mcp_servers
|
||
]
|
||
mcp_ids = [m for m in mcp_ids if m]
|
||
|
||
# 构造本回合专用对话：EngineCore 要求已存在且归属当前用户的 Conversation
|
||
conv = Conversation(
|
||
id=gen_uuid(),
|
||
user_id=real_user.id,
|
||
title=f"IM对话-{platform}",
|
||
source="im_bot",
|
||
)
|
||
db.add(conv)
|
||
db.commit()
|
||
|
||
req = SendMessageRequest(
|
||
content=prompt,
|
||
skill_ids=skill_ids or None,
|
||
mcp_ids=mcp_ids or None,
|
||
use_tools=bool(skill_ids or mcp_ids),
|
||
agent_ids=None,
|
||
knowledge_base_ids=None,
|
||
attachments=None,
|
||
)
|
||
|
||
adapter = BufferAdapter(tenant_id=str(getattr(real_user, "tenant_id", "") or ""))
|
||
await _send_message_core(conv.id, req, db, real_user, stream=True, on_event=adapter)
|
||
return adapter.assemble()
|