jira:NYJ-1540 desc: add my_logger.py

 import os
 import time
-import logging
+
+
+
 from aabd.base.enhance_dict import value_or_default
 
-logger = logging.getLogger(__name__)
+
 
 from utils.football_replay_video_event_by_llm import FootballReplayVideoEvent
 from utils.football_replay_match_live import FootballReplayMatchLive
 
+from utils.my_logger import get_logger
+
+logger = get_logger()
+
 class FootballReplayMatch:
     def __init__(self, settings):
         self.settings = settings
@@ -93,25 +99,25 @@ class FootballReplayMatch:
             start_time = time.time()
             replay_det_info = self.det_goal_replay(replay_info, task_id)
             end_time = time.time()
-            logger.info(f"判断是否是进球耗时： {end_time - start_time} 秒")
-            print(f"判断是否是进球耗时： {end_time - start_time} 秒")
+            logger.info(f"Goal_recognition_all_cost_time:{end_time-start_time:.2f}seconds")
+            # print(f"判断是否是进球耗时： {end_time - start_time} 秒")
             replay_event_name = replay_det_info.get('event_name', '')
             if replay_event_name == '进球':
                 matched_event = self.match_by_llm(replay_info, goal_live_events, task_id)
                 if matched_event is None:
-                    logger.info(f"{task_id}_LLM认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 是进球但是未能找到匹配的事件")
-                    print(f"{task_id}_LLM认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 是进球但是未能找到匹配的事件")
+                    logger.info(f"{task_id}_LLM_think_{replay_info['url']}_{replay_start_utc}_{replay_end_utc}_is_goal_but_no_matched_event")
+                    # print(f"{task_id}_LLM认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 是进球但是未能找到匹配的事件")
                 else:
                     matched_event_id = matched_event.get('video_id')
-                    logger.info(f"{task_id}_LLM认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 对应的进球的事件视频是{matched_event_id}")
-                    print(f"{task_id}_LLM认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 对应的进球的事件视频是{matched_event_id}")
+                    logger.info(f"{task_id}_LLM_think_{replay_info['url']}_{replay_start_utc}_{replay_end_utc}_matched_with_{matched_event_id}")
+                    # print(f"{task_id}_LLM认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 对应的进球的事件视频是{matched_event_id}")
             else:
-                logger.info(f"{task_id}_LLM判断{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 不是进球，无需匹配")
-                print(f"{task_id}_LLM判断{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 不是进球，无需匹配")
+                logger.info(f"{task_id}_LLM_think_{replay_info['url']}_{replay_start_utc}_{replay_end_utc}_is_not_goal_no_need_match")
+                # print(f"{task_id}_LLM判断{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 不是进球，无需匹配")
         else:
             matched_event_id = matched_event.get('id')
-            logger.info(f"{task_id}_通过时间找到匹配认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 对应的进球的事件视频是{matched_event_id}")
-            print(f"{task_id}_通过时间找到匹配认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 对应的进球的事件视频是{matched_event_id}")
+            logger.info(f"{task_id}_matched_by_time{replay_info['url']}_{replay_start_utc}_{replay_end_utc}_matched_with_{matched_event_id}")
+            # print(f"{task_id}_通过时间找到匹配认为视频{replay_info['url']}_{replay_start_utc}_{replay_end_utc} 对应的进球的事件视频是{matched_event_id}")
         return {
                 "id": task_id,
                 "match_id": match_id,

@@ -20,6 +20,10 @@ try:
 except ImportError:
     from .api import FootballReplayMatch
     
+    
+    
+    
+
 from pathlib import Path
 def get_video_files_pathlib(folder_path, extensions=None):
     if extensions is None:
@@ -32,7 +36,7 @@ def get_video_files_pathlib(folder_path, extensions=None):
                    if f.is_file() and f.suffix.lower() in extensions]
     return sorted(image_files)
 
-def _fake_settings(match_by_time_threshold: int = 30):
+def _fake_settings(match_by_time_threshold: int = 30, cache_dir_tag: str = "test"):
     """构造一个 settings 伪对象，兼容 api.py 中的 value_or_default 读取逻辑。"""
     return SimpleNamespace(
         match_by_time_threshold=match_by_time_threshold,
@@ -41,18 +45,34 @@ def _fake_settings(match_by_time_threshold: int = 30):
                             temperature=0.7, 
                             api_key='no_key'
                             ),
-        common=SimpleNamespace(cache_dir="/root/lzw/tmp_0527_replay_cache_08/videos_cache", 
-                               task_log_dir="/root/lzw/tmp_0527_replay_cache_08/tasks_log"
+        common=SimpleNamespace(cache_dir=f"/root/lzw/tmp_0528_replay_cache_{cache_dir_tag}/videos_cache", 
+                               task_log_dir=f"/root/lzw/tmp_0528_replay_cache_{cache_dir_tag}/tasks_log"
                                ),
         save_frames_enable=True,
     )
 
 
+def get_replays_and_events_pair():
+    
+    videos_dir = rf"/root/lzw/finished"
+    
+    dir_pairs = []
+    for video_name in sorted(os.listdir(videos_dir)):
 
+        live_dir = os.path.join(videos_dir, video_name, 'live')
+        live_video_dir = os.path.join(live_dir, 'videos')
         
-def get_event_list():
+        replays_dir = os.path.join(videos_dir, video_name, 'replays')
+        replay_videos_dir = os.path.join(replays_dir, 'videos')
+        
+        dir_pairs.append((replay_videos_dir, live_video_dir))
+    return dir_pairs
+
+
+
+def get_event_list(events_mp4_dir):
     # url = "http://video.mam.miguvideo.com/mnt6/fastclip3/wsc/2026/04/29/df760b8d15394cbc9ed0d0a4a9c4b786_1080PS/29133605/vodtmp/3b3e99cf4ca84c3782503d8817242de2.m3u8"
-    events_mp4_dir = r"/root/lzw/69e4ed7c872d573c9554510b-auto-10154973/goal"
+    # events_mp4_dir = r"/root/lzw/finished/69dd5845dd0412067b8d5587-auto-1776074760653/live/videos"
     
     events_mp4_names = get_video_files_pathlib(events_mp4_dir)
     
@@ -60,7 +80,7 @@ def get_event_list():
     for mp4 in events_mp4_names:
         mp4_path = os.path.join(events_mp4_dir, mp4)
         event_lis.append({
-            "id": f"event_{mp4.split('.')[0]}",
+            "id": mp4,
             "type": "1",
             "url": mp4_path,
             # "event_utc": None
@@ -73,7 +93,7 @@ def _base_data(i, replay_path, event_lis):
         "id": f"task_{str(i).zfill(3)}",
         "match_id": f"match_{str(i).zfill(3)}",
         "replay": {
-            "id": f"replay_{str(i).zfill(3)}",
+            "id": os.path.basename(replay_path),
             "url": replay_path,
             # "start_utc": None,
             # "end_utc": None
@@ -82,24 +102,36 @@ def _base_data(i, replay_path, event_lis):
     }
 
 
-class TestReplayMatchEvent(unittest.TestCase):
 
-    @patch("utils.football_replay_match_live.FootballReplayMatchLive")
-    @patch("utils.football_replay_video_event_by_llm.FootballReplayVideoEvent")
-    def test_match_by_time_success(self, mock_video_cls, mock_live_cls):
+
+
+
+def test_match_by_time_success():
     # """场景1：时间差在阈值内，直接通过 match_by_time 命中。"""
-        settings = _fake_settings(match_by_time_threshold=10)  # threshold = 30 * 1000 = 30000 ms
+    
+    replays_and_events_pair = get_replays_and_events_pair()
+    
+    for replays_mp4_dir, events_mp4_dir in replays_and_events_pair:
+        print(100*"=")
+        print(f"replays_mp4_dir: {replays_mp4_dir}, events_mp4_dir: {events_mp4_dir}")
+        
+        tag = os.path.basename(os.path.dirname(os.path.dirname(replays_mp4_dir))) 
+        settings = _fake_settings(match_by_time_threshold=30, cache_dir_tag=tag)  # threshold = 30 * 1000 = 30000 ms
+        
+        
         frm = FootballReplayMatch(settings)
 
-        replays_mp4_dir = "/root/lzw/69e4ed7c872d573c9554510b-auto-10154973/replay"
+        # replays_mp4_dir = "/root/lzw/finished/69dd5845dd0412067b8d5587-auto-1776074760653/replays/videos"
         relays_mp4_names = get_video_files_pathlib(replays_mp4_dir)
         
-        event_list = get_event_list()
+        
+        # events_mp4_dir = r"/root/lzw/finished/69dd5845dd0412067b8d5587-auto-1776074760653/live/videos"
+        event_list = get_event_list(events_mp4_dir)
         
         all_time = 0.0
         max_time = 0.0
         for i, replay_name in enumerate(relays_mp4_names):
-            print(100*"*" + f"replay_name: {replay_name}")
+            print(100*"*")
             replay_path = os.path.join(replays_mp4_dir, replay_name)
             data = _base_data(i+1, replay_path, event_list)
             
@@ -115,13 +147,31 @@ class TestReplayMatchEvent(unittest.TestCase):
             print(result)
             print()
             print()
-            print()
             
         print(100*"#")
         print(f"max_time: {max_time}")
         print(f"all_time: {all_time}")
         # exit()
 
+
+
+
+if __name__ == "__main__":
+    test_match_by_time_success()
+
+
+
+
+
+
+
+
+
+
+
+
+
+
     #     self.assertEqual(result["id"], "task_001")
     #     self.assertEqual(result["match_id"], "match_001")
     #     self.assertEqual(result["replay_id"], "replay_001")
@@ -245,7 +295,3 @@ class TestReplayMatchEvent(unittest.TestCase):
 
     #     with self.assertRaises(KeyError):
     #         frm.replay_match_event(data)
-
-
-if __name__ == "__main__":
-    unittest.main()

 import re
 import json
+import time
 import os.path
 from pathlib import Path
 from datetime import timedelta
@@ -9,9 +10,14 @@ from langchain_openai import ChatOpenAI
 
 try:
     from .llm_image import Video2Frame
+    from .my_logger import get_logger
 except:
     from llm_image import Video2Frame
+    from my_logger import get_logger
 
+
+
+logger = get_logger()
 # req_prompt = """
 # # Role
 # 你是一名拥有20年经验的足球视频技术分析师，擅长结合**视觉画面**与**解说音频（ASR）**进行跨镜头的事件匹配。你能通过解说员的描述锁定时间背景，并通过视觉特征确认物理细节。
@@ -192,11 +198,14 @@ class FootballReplayMatchLive:
         user_contents.append({'type': "text", "text": "\n请根据上述片段进行匹配并按要求输出结果/no_think\n"})
         system_message = SystemMessage(content=req_prompt)
         user_message = HumanMessage(content=user_contents)
-        result = self.model.invoke([system_message, user_message]).content
         
+        start_time = time.time()
+        result = self.model.invoke([system_message, user_message]).content
+        end_time = time.time()
+        logger.info(f"match_once_invoke_cost_time:{end_time-start_time:.2f}seconds")
         
         # print(100*"3")
-        # print(result)
+        # logger.info(f"match once invoke result: {result}")
         # print(100*"4")
         # try:
         #     result_json = json.loads(result)
@@ -299,7 +308,7 @@ class FootballReplayMatchLive:
                                                     start=event_start, 
                                                     end=event_end, 
                                                     roi=None, 
-                                                    max_px_area=400_000,
+                                                    max_px_area=300_000,
                                                     prompt_start="\n【回放片段信息】\n",
                                                     prompt_end=f"\n回放解说内容：无\n"
                                                     )

 import json
+import time
 import os.path
 from pathlib import Path
 
+
+
 from langchain_core.messages import SystemMessage, HumanMessage
 from langchain_openai import ChatOpenAI
 
 try:
     from .llm_image import Video2Frame
+    from .my_logger import get_logger
 except:
     from llm_image import Video2Frame
+    from my_logger import get_logger
+
+
+
+logger = get_logger()
 
 
 req_prompt = """
@@ -108,7 +117,12 @@ class FootballReplayVideoEvent:
         system_message = SystemMessage(content=req_prompt)
         video_message = HumanMessage(content=contents)
         # asr_message = HumanMessage(content=f"解说内容：{asr_text}")
+        
+        start_time = time.time()
         result = self.model.invoke([system_message, video_message]).content
+        end_time = time.time()
+        logger.info(f"video_event_invoke_cost_time:{end_time-start_time:.2f}seconds")
+        # logger.info(f"video event invoke result: {result}")
         try:
             result_json = json.loads(result)
         except json.JSONDecodeError:

+import os
+import time
+import logging
+from functools import lru_cache
+
+@lru_cache(maxsize=1)
+def get_logger():
+    name = "football_replay_match"
+    logger = logging.getLogger(name)
+    logger.setLevel(logging.INFO)
+
+    # 1. 配置日志输出格式
+    # 修改点：增加了 %(filename)s:%(lineno)d 来显示文件名和行号
+    formatter = logging.Formatter(
+        '%(asctime)s.%(msecs)03d - %(name)s - %(levelname)s - [%(filename)s:%(lineno)d] - %(message)s',
+        datefmt='%Y-%m-%d %H:%M:%S'
+    )
+
+    # 2. 创建控制台处理器
+    console_handler = logging.StreamHandler()
+    console_handler.setFormatter(formatter)
+    logger.addHandler(console_handler)
+
+    # 3. 创建文件处理器
+    log_dir = "logs"
+    if not os.path.exists(log_dir):
+        os.makedirs(log_dir)
+    
+    log_file = os.path.join(log_dir, f"run_{time.strftime('%Y%m%d')}.log")
+    file_handler = logging.FileHandler(log_file, encoding='utf-8')
+    file_handler.setFormatter(formatter)
+    logger.addHandler(file_handler)
+
+    return logger
\ No newline at end of file