多语言调试

.gitignore

@@ -121,6 +121,7 @@ celerybeat.pid
 *.sage.py
 
 # Environments
+data
 .env
 .venv
 env/

pp.py

@@ -1,15 +1,91 @@
 #!/usr/bin/env python3
-from paddleocr import PaddleOCR, draw_ocr
+
+import io
+import requests
+import oss2
+import pymysql
+import json
+import numpy as np
+import warnings
+from PIL import Image, ImageFile
+from dotenv import dotenv_values
+from elasticsearch import Elasticsearch
+from paddleocr import PaddleOCR
+
+warnings.filterwarnings("ignore")
+ImageFile.LOAD_TRUNCATED_IMAGES = True
+config = dotenv_values(".env")
+oss2.defaults.connection_pool_size = 100
+
+class MyEncoder(json.JSONEncoder):
+    def default(self, obj):
+        if isinstance(obj, np.float32):
+            return int(obj)
+        if isinstance(obj, np.ndarray):
+            return obj.astype(int).tolist()
+        return super(MyEncoder, self).default(obj)
+
+def download_image(url:str) -> Image.Image:
+    try:
+        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'])
+            return 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)
+            return Image.open(io.BytesIO(response.content))
+    except Exception:
+        print(f'图片从{url}下载失败')
+        return None
+
+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))
+
+CH = PaddleOCR(use_angle_cls=True, lang="ch")
+JP = PaddleOCR(use_angle_cls=True, lang="japan")
+KR = PaddleOCR(use_angle_cls=True, lang="korean")
+
+def process_images(conn, es):
+    with conn.cursor(pymysql.cursors.SSCursor) as cursor:
+        cursor.execute("SELECT id, content FROM web_images WHERE text='' LIMIT 10")
+        for id, content in cursor.fetchall():
+            image = download_image(content)
+            if image is None:
+                continue
+            print(CH.ocr(image, cls=True))
+            print(JP.ocr(image, cls=True))
+            print(KR.ocr(image, cls=True))
+        #    item = [x for x in ocr.ocr(image) if x['text'] and not x['text'].isdigit() and len(x['text']) > 1]
+        #    text = ' '.join([x['text'] for x in item])
+        #    print(id, text)
+        #    save_text(conn, id, json.dumps(item, ensure_ascii=False, cls=MyEncoder))
+        #    es.index(index='web_images', id=id, body={'content': text})
+        #conn.commit()
+
+def main():
+    es = Elasticsearch(config['ELASTICSEARCH_HOST'], basic_auth=(config['ELASTICSEARCH_USERNAME'], config['ELASTICSEARCH_PASSWORD']), verify_certs=False)
+    if not es.indices.exists(index='web_images'):
+        es.indices.create(index='web_images')
+    conn = connect_to_mysql()
+    process_images(conn, es)
+
+if __name__ == "__main__":
+    main()
+
 
 # Paddleocr目前支持的多语言语种可以通过修改lang参数进行切换
 # 例如`ch`, `en`, `fr`, `german`, `korean`, `japan`
-ocr = PaddleOCR(use_angle_cls=True, lang="ch")  # need to run only once to download and load model into memory
-img_path = './imgs/14.jpg'
-result = ocr.ocr(img_path, cls=True)
-for idx in range(len(result)):
-    res = result[idx]
-    for line in res:
-        print(line)
+#ocr = PaddleOCR(use_angle_cls=True, lang="ch")
+#img_path = './imgs/14.jpg'
+#result = ocr.ocr(img_path, cls=True)
+#for idx in range(len(result)):
+#    res = result[idx]
+#    for line in res:
+#        print(line)
 
 # 显示结果
 #from PIL import Image
@@ -21,4 +97,3 @@ for idx in range(len(result)):
 #im_show = draw_ocr(image, boxes, txts, scores, font_path='./fonts/simfang.ttf')
 #im_show = Image.fromarray(im_show)
 #im_show.save('result.jpg')
-