联合国代码判重

gwzk/lhg.py

@@ -180,10 +180,10 @@ def doJob():
             publishDate = publishDate.strftime("%Y-%m-%d %H:%M:%S")
             if publishDate < '2023-01-20':
                 continue
-            # is_href = db_storage.find_one({'网址': href})
-            # if is_href:
-            #     log.info(f'{href}===已采集')
-            #     continue
+            is_href = db_storage.find_one({'网址': href})
+            if is_href:
+                log.info(f'{href}===已采集')
+                continue
             div.find_all('div')[0].extract()
             div.find('span', class_='Z3988').extract()
             contentWithTag = div