mdbq 3.9.7__py3-none-any.whl → 3.9.8__py3-none-any.whl

mdbq/__version__.py

@@ -1 +1 @@
-VERSION = '3.9.7'
+VERSION = '3.9.8'

mdbq/log/mylogger.py

@@ -9,6 +9,8 @@ import threading
 import queue
 from typing import Optional, Dict, Any, List, Callable, Union
 import atexit
+import traceback
+import inspect
 
 try:
     import psutil
@@ -123,6 +125,7 @@ class MyLogger:
         # 定时刷新相关
         self._flush_thread = None
         self._last_flush_time = 0
+        self._start_flush_thread()
 
         # 创建日志记录器
         self.logger = logging.getLogger(name)
@@ -219,12 +222,9 @@ class MyLogger:
                         'timestamp': datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S"),
                         'level': record.levelname,
                         'message': record.getMessage(),
-                        # 'module': record.module,
-                        'function': record.funcName,
-                        # 'line': record.lineno,
-                        # 'thread': record.threadName,
-                        # 'process': record.processName,
                         'name': record.name,
+                        # 'module': record.module,
+                        # 'function': record.funcName,
                     }
 
                     # 添加额外字段
@@ -347,6 +347,45 @@ class MyLogger:
         )
         self._async_thread.start()
 
+    def log_error_handler(retry_times=0, fallback_level='error'):
+        """
+        日志错误处理装饰器
+
+        参数:
+        - retry_times: 异常时重试次数
+        - fallback_level: 降级日志级别
+        """
+
+        def decorator(log_method):
+            def wrapper(self, level: str, message: str, extra: Optional[Dict] = None):
+                last_exception = None
+                for attempt in range(retry_times + 1):
+                    try:
+                        return log_method(self, level, message, extra)
+                    except Exception as e:
+                        last_exception = e
+                        if attempt < retry_times:
+                            time.sleep(0.1 * (attempt + 1))  # 简单的退避策略
+                            continue
+
+                        try:
+                            # 降级处理
+                            logging.basicConfig()
+                            fallback_logger = logging.getLogger(f"{getattr(self, 'name', 'mylogger')}_fallback")
+                            fallback_msg = f"[降级处理] {message}"[:1000]
+                            getattr(fallback_logger, fallback_level)(
+                                f"日志记录失败(尝试{attempt + 1}次): {e}\n原始消息: {fallback_msg}"
+                            )
+                        except:
+                            sys.stderr.write(f"严重: 日志系统完全失败 - {last_exception}\n")
+
+                return None
+
+            return wrapper
+
+        return decorator
+
+    @log_error_handler(retry_times=1, fallback_level='warning')
     def _sync_log(self, level: str, message: str, extra: Optional[Dict] = None):
         """同步日志记录"""
         if not hasattr(self.logger, level.lower()):
@@ -375,7 +414,8 @@ class MyLogger:
             log_extra['context_data'] = self._context.data.copy()
 
         # 添加敏感字段过滤
-        log_extra['过滤'] = self.sensitive_fields
+        if self.sensitive_fields:
+            log_extra['过滤'] = self.sensitive_fields
 
         # 应用日志采样
         if self.sample_rate < 1.0 and level.lower() in ('debug', 'info'):
@@ -441,28 +481,6 @@ class MyLogger:
         if hasattr(self._context, 'data'):
             self._context.data.clear()
 
-    def shutdown(self):
-        """关闭日志记录器，确保所有日志被刷新"""
-        if self.enable_async:
-            self._stop_event.set()
-            # 等待队列清空
-            while not self._log_queue.empty():
-                time.sleep(0.1)
-            if self._async_thread and self._async_thread.is_alive():
-                self._async_thread.join(timeout=2)
-            if self._flush_thread and self._flush_thread.is_alive():
-                self._flush_thread.join(timeout=2)
-
-        # 确保所有handler被刷新
-        self._flush_handlers()
-
-        # 关闭所有handler
-        for handler in self.logger.handlers:
-            try:
-                handler.close()
-            except:
-                pass
-
     def debug(self, message: str, extra: Optional[Dict] = None):
         """记录调试信息"""
         self.log('debug', message, extra)
@@ -487,41 +505,46 @@ class MyLogger:
         """记录异常信息"""
         if not extra:
             extra = {}
-        # # 获取异常发生的实际位置
+
+        # # 获取完整的异常堆栈
         # tb = exc_info.__traceback__
+        # while tb.tb_next:
+        #     tb = tb.tb_next  # 获取最内层的堆栈帧
         #
-        # if tb:
-        #     extra.update({
-        #         'module': tb.tb_frame.f_globals.get('__name__', ''),
-        #         'function': tb.tb_frame.f_code.co_name,
-        #         'line': tb.tb_lineno,
-        #         'file': tb.tb_frame.f_code.co_filename
-        #     })
-        # extra['异常'] = str(exc_info)
-        # extra['类型'] = exc_info.__class__.__name__
-        # self.log('error', message, extra)
-
-        # 获取完整的异常堆栈
-        tb = exc_info.__traceback__
-        while tb.tb_next:
-            tb = tb.tb_next  # 获取最内层的堆栈帧
-
-        extra.update({
-            'module': tb.tb_frame.f_globals.get('__name__', ''),
-            'function': tb.tb_frame.f_code.co_name,
-            'line': tb.tb_lineno,
-            'file': tb.tb_frame.f_code.co_filename,
-            '异常': str(exc_info),
-            '类型': exc_info.__class__.__name__,
-            '堆栈': self._format_traceback(exc_info)
-        })
+        # extra.update({
+        #     'module': tb.tb_frame.f_globals.get('__name__', ''),
+        #     'function': tb.tb_frame.f_code.co_name,
+        #     'line': tb.tb_lineno,
+        #     'file': tb.tb_frame.f_code.co_filename,
+        #     '异常': str(exc_info),
+        #     '类型': exc_info.__class__.__name__,
+        #     '堆栈': self._format_traceback(exc_info)
+        # })
+
+        # 使用inspect获取调用栈
+        frame = inspect.currentframe()
+        try:
+            # 向上追溯2层（1层是exception方法本身，2层是实际调用位置）
+            caller_frame = frame.f_back.f_back
+            extra.update({
+                'module': caller_frame.f_globals.get('__name__', ''),
+                'function': caller_frame.f_code.co_name,
+                'line': caller_frame.f_lineno,
+                'file': caller_frame.f_code.co_filename,
+                '异常': str(exc_info),
+                '类型': exc_info.__class__.__name__,
+                '堆栈': self._format_traceback(exc_info)
+            })
+        finally:
+            del frame  # 避免循环引用
 
         # 直接使用logger的error方法记录，保留原始调用栈
         self.log('error', message, extra)
 
     def _format_traceback(self, exc_info):
         """格式化异常堆栈"""
-        import traceback
+        if exc_info is None:
+            return ""
         return ''.join(traceback.format_exception(type(exc_info), exc_info, exc_info.__traceback__))
 
     def timeit(self, message: str = "Execution time"):
@@ -579,17 +602,40 @@ class MyLogger:
                 except:
                     pass
 
+    def shutdown(self):
+        """关闭日志记录器，确保所有日志被刷新"""
+        if self.enable_async:
+            self._stop_event.set()
+            # 等待队列清空
+            while not self._log_queue.empty():
+                time.sleep(0.1)
+            if self._async_thread and self._async_thread.is_alive():
+                self._async_thread.join(timeout=0.5)
+
+        # 确保所有handler被刷新
+        if self._flush_thread:
+            self._flush_handlers()
+            if self._flush_thread.is_alive():
+                self._flush_thread.join(timeout=0.5)
+
+        # 关闭所有handler
+        for handler in self.logger.handlers:
+            try:
+                handler.close()
+            except:
+                pass
+
 def main():
     # 创建日志记录器
     logger = MyLogger(
         name='my_app',
         logging_mode='both',
         log_level='DEBUG',
-        log_file='/Users/xigua/Downloads/my_app.log',
+        log_file='my_app.log',
         log_format='json',
         max_log_size=50,
         backup_count=5,
-        enable_async=True,  # 是否启用异步日志
+        enable_async=False,  # 是否启用异步日志
         sample_rate=1,  # 采样50%的DEBUG/INFO日志
         sensitive_fields=[],  #  敏感字段列表
         enable_metrics=False,  # 是否启用性能指标
@@ -603,4 +649,3 @@ def main():
 
 if __name__ == '__main__':
     pass
-    main()

mdbq/mysql/deduplicator.py

@@ -89,6 +89,9 @@ class MySQLDeduplicator:
         :param retry_interval: 重试间隔(秒)
         :param pool_size: 连接池大小
         """
+        # 连接池状态标志
+        self._closed = False
+
         # 初始化连接池
         self.pool = PooledDB(
             creator=pymysql,
@@ -120,6 +123,8 @@ class MySQLDeduplicator:
 
     def _get_connection(self):
         """从连接池获取连接"""
+        if self._closed:
+            raise ConnectionError("连接池已关闭")
         try:
             conn = self.pool.connection()
             logger.debug("成功获取数据库连接")
@@ -263,7 +268,9 @@ class MySQLDeduplicator:
 
             # 构建去重SQL
             column_list = ', '.join([f'`{col}`' for col in use_columns])
-            temp_table = f"temp_{table}_{int(time.time())}"
+            # temp_table = f"temp_{table}_{int(time.time())}"
+            temp_table = f"temp_{table}_dedup_{os.getpid()}"  # 使用进程ID构建临时表
+            temp_table = re.sub(r'[^a-zA-Z0-9_]', '_', temp_table)  # 确保表名合法
 
             # 使用临时表方案处理去重，避免锁表问题
             create_temp_sql = f"""
@@ -556,13 +563,12 @@ class MySQLDeduplicator:
     def close(self):
         """关闭连接池"""
         try:
-            if hasattr(self, 'pool') and self.pool:
+            if hasattr(self, 'pool') and self.pool and not self._closed:
                 self.pool.close()
+                self._closed = True
                 logger.info("数据库连接池已关闭")
         except Exception as e:
             logger.error(f"关闭连接池时出错: {str(e)}", {'error_type': type(e).__name__})
-        finally:
-            self.pool = None
 
     def __enter__(self):
         return self

mdbq/mysql/uploader.py

@@ -661,7 +661,10 @@ class MySQLUploader:
             else:
                 return 'BIGINT'
         elif isinstance(value, float):
-            return 'DECIMAL(10,2)'
+            # 计算小数位数
+            num_str = str(value)
+            _, decimal_places = count_decimal_places(num_str)
+            return f'DECIMAL(20,{min(decimal_places, 6)})'  # 限制最大6位小数
         elif isinstance(value, (datetime.datetime, pd.Timestamp)):
             return 'DATETIME'
         elif isinstance(value, datetime.date):
@@ -694,7 +697,7 @@ class MySQLUploader:
             data: Union[Dict, List[Dict], pd.DataFrame],
             set_typ: Dict[str, str],
             allow_null: bool = False
-    ) -> List[Dict]:
+    ) -> Tuple[List[Dict], Dict[str, str]]:
         """
         准备要上传的数据，验证并转换数据类型
 
@@ -821,7 +824,7 @@ class MySQLUploader:
                 raise ValueError(error_msg)
 
             # 准备数据
-            prepared_data, set_typ = self._prepare_data(data, set_typ, allow_null)
+            prepared_data, filtered_set_typ = self._prepare_data(data, set_typ, allow_null)
 
             # 检查数据库是否存在
             if not self._check_database_exists(db_name):
@@ -861,7 +864,7 @@ class MySQLUploader:
                 for part_table, part_data in partitioned_data.items():
                     try:
                         self._upload_to_table(
-                            db_name, part_table, part_data, set_typ,
+                            db_name, part_table, part_data, filtered_set_typ,
                             primary_keys, check_duplicate, duplicate_columns,
                             allow_null, auto_create, partition_date_column,
                             indexes, batch_id
@@ -875,7 +878,7 @@ class MySQLUploader:
             else:
                 # 不分表，直接上传
                 self._upload_to_table(
-                    db_name, table_name, prepared_data, set_typ,
+                    db_name, table_name, prepared_data, filtered_set_typ,
                     primary_keys, check_duplicate, duplicate_columns,
                     allow_null, auto_create, partition_date_column,
                     indexes, batch_id
@@ -1093,10 +1096,10 @@ class MySQLUploader:
                         if attempt < max_retries - 1:
                             time.sleep(delay * (attempt + 1))
                             continue
-                        raise MySQLUploaderError(f"操作重试{max_retries}次后失败") from e
+                        raise logger.error(f"操作重试{max_retries}次后失败")
                     except Exception as e:
-                        raise MySQLUploaderError(f"操作失败: {str(e)}") from e
-                raise last_exception if last_exception else MySQLUploaderError("未知错误")
+                        raise logger.error(f"操作失败: {str(e)}")
+                raise last_exception if last_exception else logger.error("操作重试失败，未知错误")
 
             return wrapper
 

{mdbq-3.9.7.dist-info → mdbq-3.9.8.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: mdbq
-Version: 3.9.7
+Version: 3.9.8
 Home-page: https://pypi.org/project/mdbq
 Author: xigua, 
 Author-email: 2587125111@qq.com

{mdbq-3.9.7.dist-info → mdbq-3.9.8.dist-info}/RECORD

@@ -1,18 +1,18 @@
 mdbq/__init__.py,sha256=Il5Q9ATdX8yXqVxtP_nYqUhExzxPC_qk_WXQ_4h0exg,16
-mdbq/__version__.py,sha256=-IZp-xcG1aWVfit9XDMRHuI2-WEcYyXeULDeK2w-mPI,17
+mdbq/__version__.py,sha256=cWQRevRu-qUG7cakaKmnu0uw_arEQikDebOc5jW6PWs,17
 mdbq/aggregation/__init__.py,sha256=EeDqX2Aml6SPx8363J-v1lz0EcZtgwIBYyCJV6CcEDU,40
 mdbq/aggregation/optimize.py,sha256=2oalzD9weZhDclUC22OLxYa8Zj7KnmsGUoUau_Jlyc4,19796
 mdbq/aggregation/query_data.py,sha256=5_OzjGR5Sq00q-EgAYmSE5V9i4Solw9y4hkldl4mvt8,179808
 mdbq/config/__init__.py,sha256=jso1oHcy6cJEfa7udS_9uO5X6kZLoPBF8l3wCYmr5dM,18
 mdbq/config/config.py,sha256=eaTfrfXQ65xLqjr5I8-HkZd_jEY1JkGinEgv3TSLeoQ,3170
 mdbq/log/__init__.py,sha256=Mpbrav0s0ifLL7lVDAuePEi1hJKiSHhxcv1byBKDl5E,15
-mdbq/log/mylogger.py,sha256=T1s94-2ERWjFkzSKJKH0EgtqWdy0DE_OGb5-Ds5vYgk,22596
+mdbq/log/mylogger.py,sha256=jHCVO7KPQrg2kcCaIrakHivZmFBJyy-24sIn2rsbK4Y,24440
 mdbq/log/spider_logging.py,sha256=-ozWWEGm3HVv604ozs_OOvVwumjokmUPwbaodesUrPY,1664
 mdbq/mysql/__init__.py,sha256=A_DPJyAoEvTSFojiI2e94zP0FKtCkkwKP1kYUCSyQzo,11
-mdbq/mysql/deduplicator.py,sha256=sd5R-6Y00yGQ2PFTW3jkXPvJ-_OFEQZCjXM99nRHa8Q,21670
+mdbq/mysql/deduplicator.py,sha256=do1OqZtGwdf_KrRU_3LK4ZlKpQFsH6Pjef8eguoY4Xo,22009
 mdbq/mysql/mysql.py,sha256=jTcizvUtRdwMhWK2i_LA9yDPmcifLjUzVhwTbC3wfJk,119785
 mdbq/mysql/s_query.py,sha256=X055aLRAgxVvueXx4NbfNjp6MyBI02_XBb1pTKw09L0,8660
-mdbq/mysql/uploader.py,sha256=tb84-rJtrwla7SeBXL9EWzn0vIKnsE_9FebGNIoKrmU,45043
+mdbq/mysql/uploader.py,sha256=Zfp1xeCD0oNuBMnSmg2AKFqa2SToPkKstBUOH_h5vBM,45260
 mdbq/other/__init__.py,sha256=jso1oHcy6cJEfa7udS_9uO5X6kZLoPBF8l3wCYmr5dM,18
 mdbq/other/download_sku_picture.py,sha256=YU8DxKMXbdeE1OOKEA848WVp62jYHw5O4tXTjUdq9H0,44832
 mdbq/other/otk.py,sha256=iclBIFbQbhlqzUbcMMoePXBpcP1eZ06ZtjnhcA_EbmE,7241
@@ -25,7 +25,7 @@ mdbq/redis/__init__.py,sha256=YtgBlVSMDphtpwYX248wGge1x-Ex_mMufz4-8W0XRmA,12
 mdbq/redis/getredis.py,sha256=Uk8-cOWT0JU1qRyIVqdbYokSLvkDIAfcokmYj1ebw8k,24104
 mdbq/spider/__init__.py,sha256=RBMFXGy_jd1HXZhngB2T2XTvJqki8P_Fr-pBcwijnew,18
 mdbq/spider/aikucun.py,sha256=OhyEv1VyAKTOHjLDM37iNDQeRg5OnrNoKODoG2VxHes,19806
-mdbq-3.9.7.dist-info/METADATA,sha256=yX7vEbqUQIMoaJXh6VGpWKyYa5Ge0-ePKoxET8Y6LBM,363
-mdbq-3.9.7.dist-info/WHEEL,sha256=jB7zZ3N9hIM9adW7qlTAyycLYW9npaWKLRzaoVcLKcM,91
-mdbq-3.9.7.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
-mdbq-3.9.7.dist-info/RECORD,,
+mdbq-3.9.8.dist-info/METADATA,sha256=Ebjy7F7OmVoOel2poNGxFCX7l6AniJrupqwmuqTOJpA,363
+mdbq-3.9.8.dist-info/WHEEL,sha256=jB7zZ3N9hIM9adW7qlTAyycLYW9npaWKLRzaoVcLKcM,91
+mdbq-3.9.8.dist-info/top_level.txt,sha256=2FQ-uLnCSB-OwFiWntzmwosW3X2Xqsg0ewh1axsaylA,5
+mdbq-3.9.8.dist-info/RECORD,,