2.19 判断是否为招标文件：创建进程池+热启动，解决内存泄漏问题的同时提升速度

flask_app/logger_setup.py

@@ -66,9 +66,9 @@ def create_logger(app, subfolder):
         file_formatter = CSTFormatter('%(asctime)s - %(name)s - %(levelname)s - %(message)s')
         file_handler.setFormatter(file_formatter)
         logger.addHandler(file_handler)
-        # stream_handler = logging.StreamHandler()
+        stream_handler = logging.StreamHandler()
-        # stream_handler.setFormatter(logging.Formatter('%(message)s'))
+        stream_handler.setFormatter(logging.Formatter('%(message)s'))
-        # logger.addHandler(stream_handler)
+        logger.addHandler(stream_handler)
         logger.setLevel(logging.INFO)  #Logger 只会处理大于或等于 INFO 级别的日志消息（例如 INFO、WARNING、ERROR、CRITICAL），而 DEBUG 级别的消息会被忽略。
         logger.propagate = False
     g.logger = logger

flask_app/routes/judge_zbfile.py

@@ -1,9 +1,10 @@
+import multiprocessing
 import os
 import threading
 import time
 from enum import Enum
 from typing import Any
-from flask import Blueprint, g
+from flask import Blueprint, g, current_app
 from flask_app.general.format_change import download_file
 from flask_app.routes.判断是否是招标文件 import judge_zbfile_exec
 from flask_app.routes.utils import validate_and_setup_logger, create_response_normal, log_error_unique_id
@@ -26,7 +27,13 @@ def judge_zbfile() -> Any:              #判断是否是招标文件
     unique_id = g.unique_id
     result = [None]  # 用于存储结果的可变对象
     done = threading.Event()  # 标志判断是否完成
-
+    pool = current_app.process_pool  # 使用全局的进程池
+    def judge_zbfile_exec_sub(file_path):
+        result = pool.apply(
+            judge_zbfile_exec,  # 你的实际执行函数
+            args=(file_path,)
+        )
+        return result
     def wrapper() -> None:
         """
         包装整个 judge_zbfile 的函数逻辑
@@ -38,7 +45,6 @@ def judge_zbfile() -> Any:              #判断是否是招标文件
             downloaded_filepath, file_type = download_file(file_url, downloaded_filename,True)
 
             if not downloaded_filepath or file_type == 4:
-                logger.error("下载地址不存在或不支持的文件类型！")
                 log_error_unique_id(unique_id, 4)
                 result[0] = JudgeResult.ERROR
                 return
@@ -46,7 +52,7 @@ def judge_zbfile() -> Any:              #判断是否是招标文件
             logger.info(f"Local file path: {downloaded_filepath}")
 
             # 调用实际的判断函数
-            judge_result = judge_zbfile_exec(downloaded_filepath)
+            judge_result = judge_zbfile_exec_sub(downloaded_filepath)
             judge = JudgeResult.YES if judge_result else JudgeResult.NO
 
             end_time = time.time()

flask_app/routes/判断是否是招标文件.py

@@ -44,6 +44,8 @@ def judge_zbfile_exec(file_path):
             except Exception:
                 print("DOCX 文件读取失败，可能为乱码文件")
                 return False  # 解析失败直接返回 False
+        pre_endtime=time.time()
+        print(f"judge_zbfile_exec预处理耗时：{pre_endtime - start_time:.2f} 秒")
         # 使用大模型进行判断
         user_query = """该文件是否属于招标文件？如果是的话，请返回'是',如果不是的话，返回'否'。请不要返回其他解释或内容。
         以下是常见的招标文件类型：

flask_app/run_serve.py

@@ -1,8 +1,24 @@
 #flask_app/run_serve.py
+import threading
+import time
+
+import requests
 from waitress import serve
 from flask_app.start_up import create_app
-
+def warmup_request():
+    # 等待服务器完全启动，例如等待 1-2 秒
+    time.sleep(5)
+    try:
+        url = "http://127.0.0.1:5000/judge_zbfile"
+        #url必须为永久地址，完成热启动，创建进程池
+        payload = {"file_url": "https://intellbid-open.oss-cn-wuhan-lr.aliyuncs.com/test/094%E5%AE%9A%E7%A8%BF-%E6%B9%96%E5%8C%97%E5%B7%A5%E4%B8%9A%E5%A4%A7%E5%AD%A6%E8%BD%BB%E6%AD%A6%E5%99%A8%E6%A8%A1%E6%8B%9F%E5%B0%84%E5%87%BB%E8%AE%BE%E5%A4%87%E9%87%87%E8%B4%AD%E9%A1%B9%E7%9B%AE%E6%8B%9B%E6%A0%87%E6%96%87%E4%BB%B6.pdf"}  # 根据实际情况设置 file_url
+        headers = {"Content-Type": "application/json"}
+        response = requests.post(url, json=payload, headers=headers)
+        print(f"Warm-up 请求发送成功，状态码：{response.status_code}")
+    except Exception as e:
+        print(f"Warm-up 请求出错：{e}")
 
 if __name__ == "__main__":
     app = create_app()
+    threading.Thread(target=warmup_request, daemon=True).start()
     serve(app, host='0.0.0.0', port=5000)

flask_app/start_up.py

@@ -1,6 +1,6 @@
 # flask_app/start_up.py
-import gc
+from concurrent.futures import ProcessPoolExecutor
-import os
+from multiprocessing import Pool
 
 from flask import Flask, g
 from flask_app.ConnectionLimiter import ConnectionLimiter
@@ -13,11 +13,10 @@ from flask_app.routes.test_zbparse import test_zbparse_bp
 from flask_app.general.llm.清除file_id import delete_file_by_ids,read_file_ids
 from flask_app.routes.judge_zbfile import judge_zbfile_bp
 from flask_app.routes.test_preprocess import test_process_bp
-
-
 def create_app():
     # 创建全局日志记录器
     app = Flask(__name__)
+    app.process_pool = Pool(processes=10, maxtasksperchild=3)
     app.global_logger = create_logger_main('model_log')  # 全局日志记录器
     # 注册蓝图
     app.register_blueprint(get_deviation_bp)