feat: 重构数据处理与程序退出逻辑

   - 修复了多选字段（如剧本、平台）因包含 ,、/ 等分隔符而导致的同步失败问题。新逻辑能智能区分不同字段和分隔符，确保数据被正确分割。

Author: baiheyyds

clients/dlsite_client.py

@@ -125,10 +125,16 @@ async def get_game_detail(self, url):
                     key = th.get_text(strip=True)
 
                     def extract_list_from_td(table_cell):
+                        # Replace <br> tags with a common separator
                         for br in table_cell.find_all("br"):
-                            br.replace_with("/")
-                        all_text = table_cell.get_text(separator="/", strip=True)
-                        return [name.strip() for name in all_text.split("/") if name.strip()]
+                            br.replace_with(",")
+                        
+                        # Get all text, using the common separator
+                        all_text = table_cell.get_text(separator=",")
+                        
+                        # Standardize all separators to the common one and then split
+                        processed_text = all_text.replace('、', ',').replace('/', ',').replace('，', ',')
+                        return [name.strip() for name in processed_text.split(',') if name.strip()]
 
                     if key in self.STAFF_MAPPING:
                         details[self.STAFF_MAPPING[key]] = extract_list_from_td(td)

clients/notion_client.py

@@ -375,14 +375,34 @@ async def create_or_update_game(self, properties_schema: dict, page_id=None, **i
                 options = []
                 values_to_process = value if isinstance(value, list) else [value]
 
+                # For tags, trust the TagManager and don't split further.
                 if notion_prop_name == FIELDS["tags"]:
                     for item in values_to_process:
                         if item:
                             options.append(str(item))
                 else:
+                    # For other multi-selects, use a smart splitting heuristic.
                     for item in values_to_process:
-                        if item:
-                            options.append(str(item))
+                        if not isinstance(item, str):
+                            if item:
+                                options.append(str(item))
+                            continue
+
+                        # Heuristic: Only split by '/' if it doesn't create very short segments.
+                        # This helps differentiate "Artist A / Artist B" from "Windows 7 / 8".
+                        use_slash_as_separator = True
+                        if '/' in item:
+                            # Threshold for a segment to be considered "too short"
+                            MIN_SEGMENT_LEN = 3 
+                            if any(len(part.strip()) < MIN_SEGMENT_LEN for part in item.split('/')):
+                                use_slash_as_separator = False
+
+                        # Standardize separators
+                        processed_item = item.replace('，', ',').replace('、', ',')
+                        if use_slash_as_separator:
+                            processed_item = processed_item.replace('/', ',')
+                        
+                        options.extend([opt.strip() for opt in processed_item.split(',') if opt.strip()])
 
                 if options:
                     unique_options = list(dict.fromkeys(options))

core/driver_factory.py

@@ -89,26 +89,42 @@ async def get_driver(self, driver_key: str) -> Optional[WebDriver]:
         logger.warn(f"{driver_key} 既未创建也无创建任务。可能需要先调用 start_background_creation。")
         return None
 
-    def shutdown(self):
-        """关闭所有驱动并停止后台事件循环。"""
+    def shutdown_sync(self):
+        """同步关闭所有驱动并停止后台事件循环。会阻塞调用线程。"""
         if not self._loop:
             return
         logger.system("正在关闭驱动工厂...")
-        # 提交关闭所有驱动的任务
         if self._drivers or self._creation_futures:
             future = asyncio.run_coroutine_threadsafe(self.close_all_drivers(), self._loop)
             try:
-                future.result(timeout=10) # 等待关闭完成
+                # 在同步上下文中，我们阻塞等待，直到驱动程序关闭
+                future.result()
             except Exception as e:
                 logger.error(f"关闭驱动时发生错误: {e}")
 
-        # 停止事件循环
         if self._loop.is_running():
             self._loop.call_soon_threadsafe(self._loop.stop)
 
-        # 等待线程结束
         if self._thread:
-            self._thread.join(timeout=5)
+            self._thread.join()
+        logger.system("驱动工厂已关闭。")
+
+    async def shutdown_async(self):
+        """异步关闭所有驱动并停止后台事件循环。"""
+        if not self._loop:
+            return
+        logger.system("正在关闭驱动工厂...")
+        if self._drivers or self._creation_futures:
+            await self.close_all_drivers()
+
+        if self._loop.is_running():
+            self._loop.call_soon_threadsafe(self._loop.stop)
+        
+        if self._thread:
+            # 在异步函数中，为了不阻塞事件循环，我们不能直接join
+            # 但由于这是程序退出的最后一步，短暂的阻塞是可以接受的
+            # 更好的方法是使用 to_thread，但对于退出逻辑，join是清晰的
+            await asyncio.to_thread(self._thread.join)
         logger.system("驱动工厂已关闭。")
 
     async def close_all_drivers(self):

core/init.py

@@ -15,7 +15,7 @@ async def init_context():
 
 async def close_context(context: dict):
     # Shutdown browser drivers first
-    driver_factory.shutdown()
+    await driver_factory.shutdown_async()
 
     # Close loop-specific resources
     if context.get("async_client"):

utils/tag_manager.py

@@ -164,6 +164,18 @@ async def process_tags(
         interaction_provider: InteractionProvider,
     ) -> List[str]:
         async with self._interaction_lock:
+            
+            def _split_tags(tags: List[str]) -> List[str]:
+                processed_tags = []
+                for tag in tags:
+                    # 拆分可能包含逗号或分号的标签字符串
+                    processed_tags.extend([t.strip() for t in tag.replace('，', ',').replace('；', ',').split(',') if t.strip()])
+                return processed_tags
+
+            dlsite_tags = _split_tags(dlsite_tags)
+            fanza_tags = _split_tags(fanza_tags)
+            ggbases_tags = _split_tags(ggbases_tags)
+
             translated_tags = []
             source_maps = [
                 (dlsite_tags, self._jp_to_cn_map, TAG_JP_TO_CN_PATH, "DLsite"),
@@ -179,12 +191,14 @@ async def process_tags(
                             tag, source_map, map_path, source_name, interaction_provider
                         )
                     if translation:
-                        translated_tags.append(translation)
+                        # Also split the translated tags
+                        translated_tags.extend([t.strip() for t in translation.replace('，', ',').replace('；', ',').split(',') if t.strip()])
 
             for tag in ggbases_tags:
                 translated = self._ggbase_map.get(tag, tag) or tag
                 if translated:
-                    translated_tags.append(translated)
+                    # Also split the translated tags
+                    translated_tags.extend([t.strip() for t in translated.replace('，', ',').replace('；', ',').split(',') if t.strip()])
 
             final_tags_set: Set[str] = set()
             for concept in list(dict.fromkeys(translated_tags)):