新增功能: - execution_log 模块:自动记录引擎调用耗时、成功/失败、错误详情 - 所有搜索引擎自动跟踪每次调用的性能指标 - CLI 搜索后自动显示执行摘要 - health 命令使用 ParallelSearch 检测所有引擎连通性 - 新增 log 命令查看历史执行日志
233 lines
8.1 KiB
Python
233 lines
8.1 KiB
Python
"""
|
||
执行日志模块 — 记录各引擎/接口的调用耗时、报错信息。
|
||
|
||
用于跟踪各个接口的执行效果,便于后续优化。
|
||
|
||
Usage:
|
||
from scripts.core.execution_log import execution_log
|
||
|
||
# 自动记录耗时(推荐用法)
|
||
with execution_log.track("bing", "search") as ctx:
|
||
results = await engine.search(query)
|
||
ctx.result_count = len(results)
|
||
|
||
# 手动记录
|
||
execution_log.record(EngineCallLog(
|
||
engine="bing",
|
||
operation="search",
|
||
duration_ms=320,
|
||
success=True,
|
||
result_count=5,
|
||
))
|
||
|
||
# 打印摘要
|
||
execution_log.print_summary()
|
||
"""
|
||
|
||
from __future__ import annotations
|
||
|
||
import json
|
||
import time
|
||
import logging
|
||
from contextlib import contextmanager
|
||
from dataclasses import dataclass, field, asdict
|
||
from datetime import datetime
|
||
from typing import Any, Optional
|
||
|
||
logger = logging.getLogger(__name__)
|
||
|
||
|
||
@dataclass
|
||
class EngineCallLog:
|
||
"""单次引擎调用记录。"""
|
||
engine: str # 引擎名(bing, sogou, baidu, ...)
|
||
operation: str = "search" # 操作类型(search, fetch, health, ...)
|
||
timestamp: str = "" # 调用时间 ISO 格式
|
||
start_time: float = 0.0 # 起始时间戳
|
||
duration_ms: int = 0 # 耗时(毫秒)
|
||
success: bool = True # 是否成功
|
||
result_count: int = 0 # 返回结果数
|
||
error: Optional[str] = None # 错误信息
|
||
error_type: Optional[str] = None # 错误类型(异常类名)
|
||
status_code: Optional[int] = None # HTTP 状态码(如有)
|
||
detail: dict[str, Any] = field(default_factory=dict) # 额外信息
|
||
|
||
|
||
class _ExecutionLogCollector:
|
||
"""全局执行日志收集器。"""
|
||
|
||
def __init__(self):
|
||
self._logs: list[EngineCallLog] = []
|
||
self._session_start = time.time()
|
||
|
||
def record(self, log: EngineCallLog):
|
||
"""记录一次引擎调用。"""
|
||
self._logs.append(log)
|
||
|
||
# 同时写 logger
|
||
status = "✓" if log.success else "✗"
|
||
msg = (
|
||
f"{status} [{log.engine}] {log.operation} "
|
||
f"{log.duration_ms}ms"
|
||
)
|
||
if log.result_count > 0:
|
||
msg += f" → {log.result_count} results"
|
||
if log.error:
|
||
msg += f" | error: {log.error}"
|
||
if log.status_code and log.status_code >= 400:
|
||
msg += f" | HTTP {log.status_code}"
|
||
|
||
if log.success:
|
||
logger.info(msg)
|
||
else:
|
||
logger.warning(msg)
|
||
|
||
@contextmanager
|
||
def track(self, engine: str, operation: str = "search",
|
||
**extra):
|
||
"""上下文管理器,自动记录耗时和异常。
|
||
|
||
Example:
|
||
with execution_log.track("bing", "search", region="cn") as ctx:
|
||
results = await engine.search(query)
|
||
ctx.result_count = len(results)
|
||
"""
|
||
entry = EngineCallLog(
|
||
engine=engine,
|
||
operation=operation,
|
||
timestamp=datetime.now().isoformat(),
|
||
start_time=time.time(),
|
||
detail=dict(extra),
|
||
)
|
||
try:
|
||
yield entry
|
||
except Exception as e:
|
||
entry.success = False
|
||
entry.error = str(e)
|
||
entry.error_type = type(e).__name__
|
||
raise
|
||
finally:
|
||
entry.duration_ms = int((time.time() - entry.start_time) * 1000)
|
||
self.record(entry)
|
||
|
||
def get_logs(self, engine: str | None = None) -> list[EngineCallLog]:
|
||
"""获取日志,可按引擎名过滤。"""
|
||
if engine:
|
||
return [l for l in self._logs if l.engine == engine]
|
||
return list(self._logs)
|
||
|
||
def get_stats(self, engine: str | None = None) -> dict[str, Any]:
|
||
"""获取统计摘要。"""
|
||
logs = self.get_logs(engine)
|
||
if not logs:
|
||
return {"total_calls": 0}
|
||
|
||
total = len(logs)
|
||
success = sum(1 for l in logs if l.success)
|
||
failed = total - success
|
||
durations = [l.duration_ms for l in logs]
|
||
result_counts = [l.result_count for l in logs]
|
||
|
||
errors = []
|
||
for l in logs:
|
||
if l.error:
|
||
errors.append({
|
||
"engine": l.engine,
|
||
"operation": l.operation,
|
||
"error_type": l.error_type,
|
||
"error": l.error[:200],
|
||
"timestamp": l.timestamp,
|
||
})
|
||
|
||
return {
|
||
"total_calls": total,
|
||
"success": success,
|
||
"failed": failed,
|
||
"success_rate": f"{success/total*100:.1f}%",
|
||
"avg_duration_ms": sum(durations) // total,
|
||
"max_duration_ms": max(durations),
|
||
"min_duration_ms": min(durations),
|
||
"total_results": sum(result_counts),
|
||
"errors": errors,
|
||
}
|
||
|
||
def print_summary(self):
|
||
"""打印本次会话的执行摘要。"""
|
||
if not self._logs:
|
||
print("暂无执行记录")
|
||
return
|
||
|
||
total_time = int((time.time() - self._session_start) * 1000)
|
||
stats = self.get_stats()
|
||
|
||
lines = [
|
||
"═══════════════════════════════════════════",
|
||
" 执行日志摘要",
|
||
"═══════════════════════════════════════════",
|
||
f" 总调用次数: {stats['total_calls']}",
|
||
f" 成功/失败: {stats['success']}/{stats['failed']} ({stats['success_rate']})",
|
||
f" 平均耗时: {stats['avg_duration_ms']}ms",
|
||
f" 最长耗时: {stats['max_duration_ms']}ms",
|
||
f" 最短耗时: {stats['min_duration_ms']}ms",
|
||
f" 总结果数: {stats['total_results']}",
|
||
f" 总耗时: {total_time}ms",
|
||
"───────────────────────────────────────────",
|
||
]
|
||
|
||
# 按引擎分组
|
||
engine_names = sorted(set(l.engine for l in self._logs))
|
||
lines.append(" 引擎明细:")
|
||
for name in engine_names:
|
||
estat = self.get_stats(name)
|
||
status = "✓" if estat["failed"] == 0 else "✗"
|
||
lines.append(
|
||
f" {status} {name:15s} "
|
||
f"calls={estat['total_calls']} "
|
||
f"avg={estat['avg_duration_ms']}ms "
|
||
f"max={estat['max_duration_ms']}ms "
|
||
f"results={estat['total_results']}"
|
||
)
|
||
|
||
# 错误详情
|
||
if stats["errors"]:
|
||
lines.append("───────────────────────────────────────────")
|
||
lines.append(" 错误详情:")
|
||
for err in stats["errors"]:
|
||
lines.append(
|
||
f" ✗ [{err['engine']}] {err['error_type']}: {err['error'][:100]}"
|
||
)
|
||
|
||
lines.append("═══════════════════════════════════════════")
|
||
print("\n".join(lines))
|
||
|
||
def to_json(self) -> str:
|
||
"""导出为 JSON。"""
|
||
data = {
|
||
"session_start": datetime.fromtimestamp(self._session_start).isoformat(),
|
||
"total_calls": len(self._logs),
|
||
"stats": self.get_stats(),
|
||
"logs": [asdict(l) for l in self._logs],
|
||
}
|
||
return json.dumps(data, ensure_ascii=False, indent=2)
|
||
|
||
def save(self, path: str = "logs/search_log.jsonl"):
|
||
"""追加写入 JSONL 日志文件。"""
|
||
import os
|
||
os.makedirs(os.path.dirname(path) or ".", exist_ok=True)
|
||
with open(path, "a", encoding="utf-8") as f:
|
||
for log in self._logs:
|
||
record = asdict(log)
|
||
record["session_start"] = datetime.fromtimestamp(
|
||
self._session_start
|
||
).isoformat()
|
||
f.write(json.dumps(record, ensure_ascii=False) + "\n")
|
||
|
||
def clear(self):
|
||
"""清空日志。"""
|
||
self._logs.clear()
|
||
self._session_start = time.time()
|
||
|
||
|
||
# 全局单例
|
||
execution_log = _ExecutionLogCollector()
|