|
| 1 | +#!/usr/bin/env python |
| 2 | +# -*- coding: utf-8 -*- |
| 3 | +""" |
| 4 | +MinerU 独立工作进程脚本 |
| 5 | +用于在独立进程中运行 MinerU,避免阻塞 Celery Worker |
| 6 | +""" |
| 7 | +import sys |
| 8 | +import json |
| 9 | +from pathlib import Path |
| 10 | + |
| 11 | + |
| 12 | +def main(): |
| 13 | + """从命令行参数读取配置并执行 MinerU 转换""" |
| 14 | + if len(sys.argv) < 6: |
| 15 | + print("Usage: python mineru_worker.py <pdf_file_path> <temp_output_dir> <server_url> <backend> <result_json_path>", file=sys.stderr) |
| 16 | + sys.exit(1) |
| 17 | + |
| 18 | + pdf_file_path = sys.argv[1] |
| 19 | + temp_output_dir = sys.argv[2] |
| 20 | + server_url = sys.argv[3] |
| 21 | + backend = sys.argv[4] |
| 22 | + result_json_path = sys.argv[5] |
| 23 | + |
| 24 | + try: |
| 25 | + from mineru.cli.common import read_fn, prepare_env |
| 26 | + from mineru.data.data_reader_writer import FileBasedDataWriter |
| 27 | + from mineru.backend.vlm.vlm_analyze import doc_analyze as vlm_doc_analyze |
| 28 | + |
| 29 | + pdf_bytes = read_fn(Path(pdf_file_path)) |
| 30 | + pdf_file_name = Path(pdf_file_path).stem |
| 31 | + |
| 32 | + local_image_dir, local_md_dir = prepare_env(temp_output_dir, pdf_file_name, "vlm") |
| 33 | + image_writer = FileBasedDataWriter(local_image_dir) |
| 34 | + |
| 35 | + middle_json, _ = vlm_doc_analyze( |
| 36 | + pdf_bytes, |
| 37 | + image_writer=image_writer, |
| 38 | + backend=backend, |
| 39 | + server_url=server_url |
| 40 | + ) |
| 41 | + |
| 42 | + result = { |
| 43 | + "success": True, |
| 44 | + "middle_json": middle_json |
| 45 | + } |
| 46 | + |
| 47 | + with open(result_json_path, 'w', encoding='utf-8') as f: |
| 48 | + json.dump(result, f, ensure_ascii=False, default=str) |
| 49 | + |
| 50 | + sys.exit(0) |
| 51 | + |
| 52 | + except Exception as e: |
| 53 | + result = { |
| 54 | + "success": False, |
| 55 | + "error": str(e) |
| 56 | + } |
| 57 | + |
| 58 | + try: |
| 59 | + with open(result_json_path, 'w', encoding='utf-8') as f: |
| 60 | + json.dump(result, f, ensure_ascii=False) |
| 61 | + except: |
| 62 | + pass |
| 63 | + |
| 64 | + print(f"Error: {e}", file=sys.stderr) |
| 65 | + sys.exit(1) |
| 66 | + |
| 67 | + |
| 68 | +if __name__ == '__main__': |
| 69 | + main() |
| 70 | + |
0 commit comments