fix: 连接 Pipeline 到爬取流程 + 夜间跳过 + 断网恢复

- CrawlService.run_spider 现在自动调用 PostCrawlPipeline（爬取→存储→筛选→去重→推送）
- 定时任务添加 22:00-06:00 夜间跳过逻辑
- 断网恢复后不会漏公告：爬虫每次从 API 按日期倒序抓取，未推送的公告下次自动补推
- Pipeline 的 _exclude_sent + mark_sent 防止重复推送

app/api/crawl.py

@@ -12,13 +12,21 @@ async def trigger_crawl(request: CrawlTriggerRequest):
     names = service.get_spider_names()
 
     all_results = []
+    total_stored = 0
+    total_notified = 0
     for name in names:
         results = await service.run_spider(name)
         all_results.extend(results)
+        for r in results:
+            if r.pipeline_result:
+                total_stored += r.pipeline_result.stored
+                total_notified += r.pipeline_result.notified
 
     return {
         "spiders_run": names,
         "total_announcements": sum(r.total_count for r in all_results),
+        "total_stored": total_stored,
+        "total_notified": total_notified,
         "errors": [r.error_message for r in all_results if not r.success],
     }