新增日志存储逻辑

Co-Authored-By: ouzhuowei <190020754@qq.com>

cmd_arg/arg.py

@@ -84,6 +84,16 @@ class InitDbOptionEnum(str, Enum):
     POSTGRES = "postgres"
 
 
+class LogLevelEnum(str, Enum):
+    """Log level enumeration"""
+
+    DEBUG = "DEBUG"
+    INFO = "INFO"
+    WARNING = "WARNING"
+    ERROR = "ERROR"
+    CRITICAL = "CRITICAL"
+
+
 def _to_bool(value: bool | str) -> bool:
     if isinstance(value, bool):
         return value
@@ -299,6 +309,31 @@ async def parse_cmd(argv: Optional[Sequence[str]] = None):
                 rich_help_panel="Proxy Configuration",
             ),
         ] = config.IP_PROXY_PROVIDER_NAME,
+        log_save_enable: Annotated[
+            str,
+            typer.Option(
+                "--log_save_enable",
+                help="Whether to save logs to file, supports yes/true/t/y/1 or no/false/f/n/0",
+                rich_help_panel="Log Configuration",
+                show_default=True,
+            ),
+        ] = str(config.LOG_SAVE_ENABLE),
+        log_save_path: Annotated[
+            str,
+            typer.Option(
+                "--log_save_path",
+                help="Log file save path, default is ./logs",
+                rich_help_panel="Log Configuration",
+            ),
+        ] = config.LOG_SAVE_PATH,
+        log_save_level: Annotated[
+            LogLevelEnum,
+            typer.Option(
+                "--log_save_level",
+                help="Log save level (DEBUG | INFO | WARNING | ERROR | CRITICAL)",
+                rich_help_panel="Log Configuration",
+            ),
+        ] = _coerce_enum(LogLevelEnum, config.LOG_SAVE_LEVEL, LogLevelEnum.INFO),
     ) -> SimpleNamespace:
         """MediaCrawler 命令行入口"""
 
@@ -306,6 +341,7 @@ async def parse_cmd(argv: Optional[Sequence[str]] = None):
         enable_sub_comment = _to_bool(get_sub_comment)
         enable_headless = _to_bool(headless)
         enable_ip_proxy_value = _to_bool(enable_ip_proxy)
+        enable_log_save = _to_bool(log_save_enable)
         init_db_value = init_db.value if init_db else None
 
         # Parse specified_id and creator_id into lists
@@ -330,6 +366,9 @@ async def parse_cmd(argv: Optional[Sequence[str]] = None):
         config.ENABLE_IP_PROXY = enable_ip_proxy_value
         config.IP_PROXY_POOL_COUNT = ip_proxy_pool_count
         config.IP_PROXY_PROVIDER_NAME = ip_proxy_provider_name
+        config.LOG_SAVE_ENABLE = enable_log_save
+        config.LOG_SAVE_PATH = log_save_path
+        config.LOG_SAVE_LEVEL = log_save_level.value
 
         # Set platform-specific ID lists for detail/creator mode
         if specified_id_list:

config/base_config.py

@@ -120,6 +120,14 @@ FONT_PATH = "./docs/STZHONGS.TTF"
 # 爬取间隔时间
 CRAWLER_MAX_SLEEP_SEC = 2
 
+# 日志配置
+# 是否保存日志
+LOG_SAVE_ENABLE = False
+# 日志保存路径
+LOG_SAVE_PATH = "./logs"
+# 日志保存级别
+LOG_SAVE_LEVEL = "INFO"
+
 from .bilibili_config import *
 from .xhs_config import *
 from .dy_config import *

tools/utils.py

@@ -20,6 +20,9 @@
 
 import argparse
 import logging
+import os
+from logging.handlers import RotatingFileHandler
+from datetime import datetime
 
 from .crawler_util import *
 from .slider_util import *
@@ -27,15 +30,55 @@ from .time_util import *
 
 
 def init_loging_config():
+    # 导入配置
+    try:
+        from config.base_config import LOG_SAVE_ENABLE, LOG_SAVE_PATH, LOG_SAVE_LEVEL
+    except ImportError:
+        LOG_SAVE_ENABLE = False
+        LOG_SAVE_PATH = "./logs"
+        LOG_SAVE_LEVEL = "INFO"
+    
     level = logging.INFO
+    log_format = "%(asctime)s %(name)s %(levelname)s (%(filename)s:%(lineno)d) - %(message)s"
+    date_format = '%Y-%m-%d %H:%M:%S'
+    
+    # 配置基础日志
     logging.basicConfig(
         level=level,
-        format="%(asctime)s %(name)s %(levelname)s (%(filename)s:%(lineno)d) - %(message)s",
-        datefmt='%Y-%m-%d %H:%M:%S'
+        format=log_format,
+        datefmt=date_format
     )
     _logger = logging.getLogger("MediaCrawler")
     _logger.setLevel(level)
 
+    # 如果启用日志保存，添加文件处理器
+    if LOG_SAVE_ENABLE and LOG_SAVE_PATH:
+        try:
+            # 确保日志目录存在
+            log_dir = os.path.abspath(LOG_SAVE_PATH)
+            os.makedirs(log_dir, exist_ok=True)
+            
+            # 日志文件名：按日期命名
+            log_filename = os.path.join(log_dir, f"mediacrawler-{datetime.now().strftime('%Y-%m-%d')}.log")
+            
+            # 转换日志级别字符串为logging级别
+            file_level = getattr(logging, LOG_SAVE_LEVEL.upper(), logging.INFO)
+            
+            # 创建文件处理器
+            file_handler = RotatingFileHandler(
+                log_filename,
+                encoding='utf-8'
+            )
+            file_handler.setLevel(file_level)
+            file_handler.setFormatter(logging.Formatter(log_format, date_format))
+            
+            # 添加到logger
+            _logger.addHandler(file_handler)
+            
+        except Exception as e:
+            # 如果文件日志配置失败，不影响控制台日志
+            _logger.warning(f"日志文件保存配置失败: {e}")
+
     # Disable httpx INFO level logs
     logging.getLogger("httpx").setLevel(logging.WARNING)