持续运行

pp.py

@@ -1,7 +1,9 @@
 #!/usr/bin/env python3
 
+import os
 import io
 import oss2
+import time
 import json
 import base64
 import dotenv
@@ -46,7 +48,11 @@ def download_image(url: str, max_size=32767) -> Image.Image:
         if url.startswith('http://image.gameuiux.cn/') or url.startswith('https://image.gameuiux.cn/'):
             url = url.replace('http://image.gameuiux.cn/', '').replace('https://image.gameuiux.cn/', '')
             oss_auth = oss2.Auth(config['OSS_ACCESS_KEY_ID'], config['OSS_ACCESS_KEY_SECRET'])
-            img = Image.open(io.BytesIO(oss2.Bucket(oss_auth, f'http://{config["OSS_HOST"]}', config['OSS_BUCKET_NAME']).get_object(url).read()))
+            if os.path.exists(url):
+                img = Image.open(url)
+            else:
+                print(f'从OSS下载图片 {url}')
+                img = Image.open(io.BytesIO(oss2.Bucket(oss_auth, f'http://{config["OSS_HOST"]}', config['OSS_BUCKET_NAME']).get_object(url).read()))
         else:
             response = requests.get(url)
             img = Image.open(io.BytesIO(response.content))
@@ -65,10 +71,6 @@ def connect_to_mysql():
     return pymysql.connect(host=config['MYSQL_HOST'], user=config['MYSQL_USER'], password=config['MYSQL_PASSWORD'], database=config['MYSQL_NAME'], cursorclass=pymysql.cursors.SSDictCursor)
 
 
-def save_text(conn, id: int, text: str):
-    with conn.cursor() as cursor:
-        cursor.execute("UPDATE web_images SET text = %s WHERE id = %s", (text, id))
-
 
 # 中英日韩俄
 EN = PaddleOCR(use_angle_cls=True, lang="en")
@@ -79,7 +81,7 @@ RU = PaddleOCR(use_angle_cls=True, lang="ru")
 
 def process_images(conn):
     with conn.cursor(pymysql.cursors.SSCursor) as cursor:
-        cursor.execute("SELECT id, content FROM web_images WHERE id>222193 AND text='' LIMIT 10")
+        cursor.execute("SELECT id, content FROM web_images WHERE text='' AND text!='[]' AND article_category_top_id=22 LIMIT 10")
         for id, content in cursor.fetchall():
             image = download_image(content)
             if image is None:
@@ -173,23 +175,22 @@ def process_images(conn):
                 data.append({'text': 文本[0], 'confidence': 文本[1], 'coordinate': 坐标 })
             
             # 转换为字符串存储到索引库
-            text = ' '.join([x['text'] for x in data])
+            obj = { "_id": str(id), "text": ' '.join([x['text'] for x in data]) }
+            res = requests.put(zinc_url, headers=headers, data=json.dumps(obj), proxies={'http': '', 'https': ''})
+            print("\033[1;32m{}\033[0m".format(id) if json.loads(res.text)['message'] == 'ok' else id, text)
 
             # 转换为 JSON 存储到数据库
-            data = json.dumps(data, ensure_ascii=False, cls=MyEncoder)
-            
-            print(id, text)
-
-            save_text(conn, id, data)
-            res = requests.put(zinc_url, headers=headers, data=json.dumps(data), proxies={'http': '', 'https': ''})
-            print("\033[1;32m{}\033[0m".format(id) if json.loads(res.text)['message'] == 'ok' else id, text)
-            conn.commit()
+            with conn.cursor() as cursor:
+                data = json.dumps(data, ensure_ascii=False, cls=MyEncoder)
+                cursor.execute("UPDATE web_images SET text = %s WHERE id = %s", (data, id))
+                conn.commit()
 
 def main():
     conn = connect_to_mysql()
-    process_images(conn)
+    while True:
+        process_images(conn)
+        time.sleep(10)
 
 
 if __name__ == "__main__":
-    for _ in range(1):
-        main()
+    main()