Update cut_image.py

magic_pdf/pre_proc/cut_image.py

@@ -24,7 +24,7 @@ def ocr_image(image_path,compress_image):
     logger.info(f'qwen解析{image_path}表格的内容为：{generated_text},耗时为：{end - start}')
     return generated_text
 
-def ocr_cut_image_and_table(ocr_status,config_path,local_image_dir,spans, page, page_id, pdf_bytes_md5, imageWriter):
+def ocr_cut_image_and_table(config_path,local_image_dir,spans, page, page_id, pdf_bytes_md5, imageWriter):
     def return_path(type):
         return join_path(pdf_bytes_md5, type)
 
@@ -40,6 +40,7 @@ def ocr_cut_image_and_table(ocr_status,config_path,local_image_dir,spans, page,
     url = config.get('server', 'ocr_server')
     workers = int(config.get('server', 'ocr_workers'))
     client = PredictClient(url)
+    ocr_status = client.check_health()
     pool = Pool(workers)
     if not ocr_status:
         logger.warning(f'Health check failed. The server at "{url}" is not responding as expected.')
@@ -59,7 +60,6 @@ def ocr_cut_image_and_table(ocr_status,config_path,local_image_dir,spans, page,
                                    imageWriter=imageWriter)
             image_path = join_path(local_image_dir, image_path)
             if ocr_status:
-
                 txt = pool.apply_async(ocr_image, args=(image_path,compress_image)).get()
                 span['image_path'] = str(txt)
             else: