OCR MCP Service

#!/usr/bin/env python3 """批量保存 OCR 结果到文件""" import json from pathlib import Path # OCR 结果数据（从之前的识别结果中提取） ocr_data = { 'IMG_20251124_220855': { 'text': '1\n法\n福\n山\n新州卫地品\n长江出版社\nCHANGJIANGPRESS\nViVO\nS20\nle]', 'confidence': 0.5232393205165863, 'boxes': [ {"x1": 1538.0, "y1": 2393.0, "x2": 1963.0, "y2": 2712.0}, {"x1": 2045.0, "y1": 2559.0, "x2": 2069.0, "y2": 2592.0}, {"x1": 1791.0, "y1": 2728.0, "x2": 1816.0, "y2": 2757.0}, {"x1": 2137.0, "y1": 2728.0, "x2": 2449.0, "y2": 3039.0}, {"x1": 3017.0, "y1": 2802.0, "x2": 3046.0, "y2": 3017.0}, {"x1": 352.0, "y1": 3848.0, "x2": 773.0, "y2": 3930.0}, {"x1": 366.0, "y1": 3930.0, "x2": 772.0, "y2": 3975.0}, {"x1": 17.0, "y1": 4448.0, "x2": 297.0, "y2": 4536.0}, {"x1": 337.0, "y1": 4442.0, "x2": 628.0, "y2": 4536.0}, {"x1": 2355.0, "y1": 4492.0, "x2": 3071.0, "y2": 4536.0} ], 'processing_time': 34.443774938583374 }, # 其他图片的数据...（由于数据量大，这里只保存关键信息） } def save_ocr_result(image_name, text, confidence, boxes=None, processing_time=None): """保存单张图片的 OCR 结果""" base_dir = Path(__file__).parent.parent / "东野圭吾图片测试集" ocr_dir = base_dir / "ocr_results" ocr_dir.mkdir(exist_ok=True) # 保存 JSON 格式 json_data = { "source_image": f"东野圭吾图片测试集/{image_name}.jpg", "text": text, "boxes": boxes or [], "confidence": confidence, "engine": "paddleocr", "processing_time": processing_time or 0.0 } json_path = ocr_dir / f"{image_name}_ocr.json" with open(json_path, 'w', encoding='utf-8') as f: json.dump(json_data, f, ensure_ascii=False, indent=2) # 保存 TXT 格式 txt_path = ocr_dir / f"{image_name}_ocr.txt" with open(txt_path, 'w', encoding='utf-8') as f: f.write(f"## 📷 源图片\n") f.write(f"- [{image_name}.jpg](东野圭吾图片测试集/{image_name}.jpg)\n\n") f.write(text) print(f"已保存: {image_name}") if __name__ == "__main__": # 这里可以批量处理 pass