年报

comData/annualReport/fbs_annualreport.py

@@ -58,22 +58,22 @@ if __name__ == '__main__':
         'Accept-Encoding': 'gzip, deflate, br',
         'Accept-Language': 'zh-CN,zh;q=0.9,en;q=0.8,en-GB;q=0.7,en-US;q=0.6',
     }
-    query = "SELECT * FROM clb_sys_attachment WHERE type_id=1 AND source='证监会' AND id = '18703822757'"
+    query = "SELECT * FROM clb_sys_attachment WHERE type_id=1 AND create_time >= '2023-10-27' AND create_time <= '2023-10-28' AND source like '%官网' "
     cursor_.execute(query)
     results = cursor_.fetchall()
     for result in results:
         att_id = result[0]
         name = result[1]
-
+        source = result[18]
         social_code = result[3]
-        selectcom = f"select * from EnterpriseInfo where SocialCode = '{social_code}'"
-        cursor.execute(selectcom)
-        selects = cursor.fetchone()
-        com_name = selects[4]
-        if com_name:
-            pass
-        else:
-            com_name = selects[1]
+        # selectcom = f"select * from EnterpriseInfo where SocialCode = '{social_code}'"
+        # cursor.execute(selectcom)
+        # selects = cursor.fetchone()
+        # com_name = selects[4]
+        # if com_name:
+        #     pass
+        # else:
+        #     com_name = selects[1]
         full_path = 'http://zzsn.luyuen.com/' + result[19]
         year = result[9]
         create_time = result[13]
@@ -103,7 +103,7 @@ if __name__ == '__main__':
             'id': '',
             'keyWords': '',
             'lang': detect_language,
-            'origin': '证监会',
+            'origin': source,
             # 'origin': '雪球网',
             'publishDate': publish,
             'sid': '1684032033495392257',

comData/annualReport1014/report.py

@@ -29,7 +29,7 @@ type_id = 1
 create_by = 'XueLingKun'
 taskType = '企业年报'
 
-file_path = 'D:\\年报\\欧盟记分牌2500_年报补充_718_20231018'
+file_path = 'D:\\年报\\福布斯2000年报PDF下载-207'
 log.info(f'=============当前pid为{baseCore.getPID()}==============')
 
 def sendKafka(dic_news):
@@ -146,9 +146,9 @@ if __name__=='__main__':
         social_code = data[1]
         ename = data[2]
         cname = data[3]
-        file_name = ename + ':' + file_year + '年年度报告' + '.pdf'
+        file_name = cname + ':' + file_year + '年年度报告' + '.pdf'
         content = ''
-        origin = ename + '官网'
+        origin = cname + '官网'
         #解析文件页数和内容
         log.info(f"-----------正在处理{file_name}--------------")
         with open(pdf_path, 'rb') as file:
@@ -178,7 +178,7 @@ if __name__=='__main__':
             retData_f = uptoOBS(retData, pathType, taskType, start_time,file_name,pdf_path)
             if retData_f['state']:
                 #retData, com_name, year, pdf_name, num, pub_time
-                att_id= baseCore.tableUpdate(retData_f, cname,file_year,file_name, num,file_year+'-12-31',origin)
+                att_id= baseCore.tableUpdate(retData_f,file_year,file_name, num,file_year+'-12-31',origin)
                 if att_id:
                     detect_language = baseCore.detect_language(content)
                     dic_news = {

comData/annualReport1014/report1.py

@@ -211,7 +211,7 @@ if __name__=='__main__':
                         'sid': '1684032033495392257',
                         'sourceAddress': '',  # 原文链接
                         'summary': '',
-                        'title': file_name,
+                        'title': file_name.replace('.pdf',''),
                         'type': 1,
                         'socialCreditCode': social_code,
                         'year': file_year