blade
/
crypto_quant
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

refresh Trump articles on Truth Social every 5 minutes

This commit is contained in:
blade 2025-10-21 18:19:51 +08:00
parent b1d6e62ca9
commit 8e2313ed76
2 changed files with 36 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
auto_fetch_truth_social.py
View File

@ -26,8 +26,9 @@ def run_script():
end_time = time.time() end_time = time.time()
logger.info(f"Script execution time: {end_time - start_time} seconds") logger.info(f"Script execution time: {end_time - start_time} seconds")
# 设置每天上午09:00:00 运行一次 # 设置每天上午09:00:00 运行一次
schedule.every().day.at("09:00:00").do(run_script) # schedule.every().day.at("09:00:00").do(run_script)
# schedule.every(60).seconds.do(run_script) # 设置每5分钟运行一次
schedule.every(5).minutes.do(run_script)
# 保持程序运行并检查调度 # 保持程序运行并检查调度
logger.info("Scheduler started. Press Ctrl+C to stop.") logger.info("Scheduler started. Press Ctrl+C to stop.")

57
core/media/truth_social_retriever.py
View File

@ -126,20 +126,22 @@ class TruthSocialRetriever:
print("获取帖子失败，请检查 API 密钥或网络。") print("获取帖子失败，请检查 API 密钥或网络。")
if len(results) > 0: if len(results) > 0:
user_path = os.path.join(self.save_path, user_name) # user_path = os.path.join(self.save_path, user_name)
os.makedirs(user_path, exist_ok=True) # os.makedirs(user_path, exist_ok=True)
now_date_time = datetime.now().strftime("%Y%m%d%H%M%S") # now_date_time = datetime.now().strftime("%Y%m%d%H%M%S")
json_file_name = os.path.join(user_path, f"{user_name}_{now_date_time}.json") # json_file_name = os.path.join(user_path, f"{user_name}_{now_date_time}.json")
# 将results内容写入json_file_name文件中 # # 将results内容写入json_file_name文件中
with open(json_file_name, 'w', encoding='utf-8') as f: # with open(json_file_name, 'w', encoding='utf-8') as f:
json.dump(results, f, ensure_ascii=False, indent=2) # json.dump(results, f, ensure_ascii=False, indent=2)
logger.info(f"已将{len(results)}条数据保存到: {json_file_name}") # logger.info(f"已将{len(results)}条数据保存到: {json_file_name}")
result_df = pd.DataFrame(results) result_df = pd.DataFrame(results)
result_df = self.remove_duplicate_posts(result_df) result_df = self.remove_duplicate_posts(result_df)
self.db_truth_social_content.insert_data_to_mysql(result_df) if len(result_df) > 0:
logger.info(f"已将{len(result_df)}条数据插入到数据库") self.db_truth_social_content.insert_data_to_mysql(result_df)
self.send_wechat_message(result_df) logger.info(f"已将{len(result_df)}条数据插入到数据库")
self.send_wechat_message(result_df)
else:
logger.info(f"没有数据需要插入到数据库和发送企业微信消息")
except requests.exceptions.RequestException as e: except requests.exceptions.RequestException as e:
print(f"请求错误: {e}") print(f"请求错误: {e}")
except json.JSONDecodeError as e: except json.JSONDecodeError as e:
@ -150,20 +152,27 @@ class TruthSocialRetriever:
results = json.load(f) results = json.load(f)
result_df = pd.DataFrame(results) result_df = pd.DataFrame(results)
result_df = self.remove_duplicate_posts(result_df) result_df = self.remove_duplicate_posts(result_df)
self.send_wechat_message(result_df) if len(result_df) > 0:
self.send_wechat_message(result_df)
else:
logger.info(f"没有数据需要发送企业微信消息")
def remove_duplicate_posts(self, result_df: pd.DataFrame): def remove_duplicate_posts(self, result_df: pd.DataFrame):
duplicate_index_list = [] try:
for index, row in result_df.iterrows(): duplicate_index_list = []
article_id = row["article_id"] for index, row in result_df.iterrows():
exist_data = self.db_truth_social_content.query_data_by_article_id(article_id) article_id = row["article_id"]
if exist_data: exist_data = self.db_truth_social_content.query_data_by_article_id(article_id)
duplicate_index_list.append(index) if exist_data:
# 删除重复的行 duplicate_index_list.append(index)
result_df = result_df.drop(duplicate_index_list) # 删除重复的行
result_df.sort_values(by="timestamp", ascending=True, inplace=True) result_df = result_df.drop(duplicate_index_list)
result_df.reset_index(drop=True, inplace=True) result_df.sort_values(by="timestamp", ascending=True, inplace=True)
logger.info(f"删除重复的行后，剩余{len(result_df)}条数据") result_df.reset_index(drop=True, inplace=True)
logger.info(f"删除重复的行后，剩余{len(result_df)}条数据")
except Exception as e:
result_df = pd.DataFrame([])
logger.error(f"删除重复的行失败: {e}")
return result_df return result_df
def send_wechat_message(self, result_df: pd.DataFrame): def send_wechat_message(self, result_df: pd.DataFrame):