关于本地运行MinerU找不到路径的问题

[hmppt]

问题描述

我利用git lfs下载了PDF-Extract-Kit-1.0中的所有模型和文件，当我想在本地使用python调用mineru时(使用官方给出的python调用demo代码)，会出现如下的网络问题，我知道我的网络无法连接modelscope和huggingface，但是我不明白为什么设置了local、mineru.json和对应模型还是会需要网络，如何正确的进行离线使用呢？以下是我的运行情况，如果有需要补充的，请告诉我！

$ python minerua.py 
2025-09-24 19:29:04.707 | INFO     | mineru.backend.pipeline.pipeline_analyze:doc_analyze:126 - Batch 1/1: 175 pages/175 pages
2025-09-24 19:29:04.708 | INFO     | mineru.backend.pipeline.model_init:__init__:192 - DocAnalysis init, this may take some times......
2025-09-24 19:29:05.151 | ERROR    | __main__:parse_doc:220 - HTTPSConnectionPool(host='api.github.com', port=443): Max retries exceeded with url: /repos/ultralytics/assets/releases/tags/v8.3.0 (Caused by SSLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self-signed certificate in certificate chain (_ssl.c:1032)')))
urllib3.exceptions.SSLError: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self-signed certificate in certificate chain (_ssl.c:1032)

环境

Windows 11

minerua.py

...
if __name__ == '__main__':
    # args
    input_dir = "D:/mission/agent/documents"  # 修改为您的PDF文件目录
    output_dir = "D:/mission/agent/output"  # 修改为您想要的输出目录

    doc_path_list = []
    for doc_path in Path(input_dir).glob('*.pdf'):  # 只处理PDF文件
        doc_path_list.append(doc_path)

    os.environ['MINERU_MODEL_SOURCE'] = "local"

    """Use pipeline mode if your environment does not support VLM"""
    parse_doc(doc_path_list, output_dir, backend="pipeline")

    """To enable VLM mode, change the backend to 'vlm-xxx'"""
    # parse_doc(doc_path_list, output_dir, backend="vlm-transformers")  # more general.
    # parse_doc(doc_path_list, output_dir, backend="vlm-vllm-engine")  # faster(engine).
    # parse_doc(doc_path_list, output_dir, backend="vlm-http-client", server_url="http://127.0.0.1:30000/")  # faster(client).

mineru.json

{
    "bucket_info":{
        "bucket-name-1":["ak", "sk", "endpoint"],
        "bucket-name-2":["ak", "sk", "endpoint"]
    },
    "latex-delimiter-config": {
        "display": {
            "left": "$$",
            "right": "$$"
        },
        "inline": {
            "left": "$",
            "right": "$"
        }
    },
    "llm-aided-config": {
        "title_aided": {
            "api_key": "your_api_key",
            "base_url": "https://dashscope.aliyuncs.com/compatible-mode/v1",
            "model": "qwen2.5-32b-instruct",
            "enable": false
        }
    },
    "models-dir": {
        "pipeline": "D:/mission/PDF-Extract-Kit-1.0",
        "vlm": ""
    },
    "config_version": "1.3.0"
}

相关文件的路径

demo的路径为/d/mission/agent
PDF-Extract-Kit-1.0的路径为/d/mission/
mineru.json的路径为/d/mission/agent，与demo在同一个根目录下

[myhloli]

"pipeline": "D:/mission/PDF-Extract-Kit-1.0",这个目录要指定到models层级

[hmppt]

感谢回答，但是当我把mineru.json修改为如下时，仍然不起作用

"models-dir": {
        "pipeline": "D:/mission/PDF-Extract-Kit-1.0/models",
        "vlm": ""
    },

报错：

2025-09-25 11:01:26.836 | INFO     | mineru.backend.pipeline.pipeline_analyze:doc_analyze:126 - Batch 1/1: 175 pages/175 pages
2025-09-25 11:01:26.837 | INFO     | mineru.backend.pipeline.model_init:__init__:192 - DocAnalysis init, this may take some times......
2025-09-25 11:01:27.080 | ERROR    | __main__:parse_doc:216 - HTTPSConnectionPool(host='api.github.com', port=443): Max retries exceeded with url: /repos/ultralytics/assets/releases/tags/v8.3.0 (Caused by SSLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self-signed certificate in certificate chain (_ssl.c:1032)')))
urllib3.exceptions.SSLError: [SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: self-signed certificate in certificate chain (_ssl.c:1032)

个人认为可能是没有读到json，在demo中不需要显式的指出配置文件的目录吗？

[myhloli]

json文件需要放在用户目录下

[hmppt]

这个也放了，因为我是在windows系统运行的，所以json放在了C:\Users\xxxx下，同时也在/d/mission/agent放了相同的json

[zhongmmming]

官方文档写的，我按以上配了以后加了这个好了，希望对你有帮助！
指定目录后您可通过配置环境变量export MINERU_MODEL_SOURCE=local来使用本地模型。

[hmppt]

我在minerua.py里面添加了os.environ['MINERU_MODEL_SOURCE'] = "local"，但仍然没有用

[FlyDut]

我看了代码，实际上指向应当指向 .cache\modelscope\hub\models\OpenDataLab\PDF-Extract-Kit-1___0 ，并且要将环境变量设置为MINERU_MODEL_SOURCE=local。官方文档没有说明这一点，这也太粗糙了……