Python使用MinerU的简单的示例

# 官网地址
https://mineru.readthedocs.io/en/latest/index.html

# Github地址
https://github.com/opendatalab/mineru

# API接口地址
https://mineru.readthedocs.io/en/latest/user_guide/quick_start/convert_pdf.html

# 模型下载脚本地址
# 从ModelScope下载模型：download_models.py
# 从HuggingFace下载模型: download_models_hf.py
https://github.com/opendatalab/MinerU/tree/master/scripts

# 我的版本是：magic-pdf==1.1.0
pip install -U "magic-pdf[full]" -i https://pypi.tuna.tsinghua.edu.cn/simple

# 官网下载方法地址
https://github.com/opendatalab/MinerU/blob/master/docs/how_to_download_models_zh_cn.md

# 安装modelscope
pip install modelscope

# 下载文件
wget https://gcore.jsdelivr.net/gh/opendatalab/MinerU@master/scripts/download_models.py -O download_models.py
# 也可以到下面的地址，找到download_models.py下载
https://github.com/opendatalab/MinerU/tree/master/scripts

# 执行下载模型
# 为了方便使用模型，我修改了download_models.py，添加了设置模型的位置。
python download_models.py

import json
import os

import requests
from modelscope import snapshot_download


def download_json(url):
    # 下载JSON文件
    response = requests.get(url)
    response.raise_for_status()  # 检查请求是否成功
    return response.json()


def download_and_modify_json(url, local_filename, modifications):
    if os.path.exists(local_filename):
        data = json.load(open(local_filename))
        config_version = data.get('config_version', '0.0.0')
        if config_version < '1.1.1':
            data = download_json(url)
    else:
        data = download_json(url)

    # 修改内容
    for key, value in modifications.items():
        data[key] = value

    # 保存修改后的内容
    with open(local_filename, 'w', encoding='utf-8') as f:
        json.dump(data, f, ensure_ascii=False, indent=4)


if __name__ == '__main__':
    mineru_patterns = [
        "models/Layout/LayoutLMv3/*",
        "models/Layout/YOLO/*",
        "models/MFD/YOLO/*",
        "models/MFR/unimernet_small_2501/*",
        "models/TabRec/TableMaster/*",
        "models/TabRec/StructEqTable/*",
    ]

    # 设置模型下载的位置
    local_dir="E:/mineru"

    # 下载模型
    model_dir = snapshot_download('opendatalab/PDF-Extract-Kit-1.0', allow_patterns=mineru_patterns, local_dir=local_dir)
    layoutreader_model_dir = snapshot_download('ppaanngggg/layoutreader', local_dir=local_dir)
    model_dir = model_dir + '/models'
    print(f'model_dir is: {model_dir}')
    print(f'layoutreader_model_dir is: {layoutreader_model_dir}')

    json_url = 'https://gcore.jsdelivr.net/gh/opendatalab/MinerU@master/magic-pdf.template.json'
    config_file_name = 'magic-pdf.json'
    home_dir = os.path.expanduser('~')
    config_file = os.path.join(home_dir, config_file_name)

    json_mods = {
        'models-dir': model_dir,
        'layoutreader-model-dir': layoutreader_model_dir,
    }

    download_and_modify_json(json_url, config_file, json_mods)
    print(f'The configuration file has been configured successfully, the path is: {config_file}')

import os

from magic_pdf.data.data_reader_writer import FileBasedDataWriter, FileBasedDataReader
from magic_pdf.data.dataset import PymuDocDataset
from magic_pdf.model.doc_analyze_by_custom_model import doc_analyze
from magic_pdf.config.enums import SupportedPdfParseMethod

# pdf文件路径
pdf_file_path = "E:/hello/test-5-2.pdf"
# 获取没有后缀的pdf文件名称
pdf_file_path_without_suff = pdf_file_path.split(".")[0]
print(pdf_file_path_without_suff)

# 文件所在的目录
pdf_file_path_parent_dir = os.path.dirname(pdf_file_path)
image_dir = os.path.join(pdf_file_path_parent_dir, "images")
print(image_dir)

# Markdown的写入实例
# markdown_dir = "./output/markdown"
# writer_markdown = FileBasedDataWriter(markdown_dir)
writer_markdown = FileBasedDataWriter()
# 读取图片
writer_image = FileBasedDataWriter(image_dir)

# 读取文件流
reader_pdf = FileBasedDataReader("")
bytes_pdf = reader_pdf.read(pdf_file_path)

# 处理数据
dataset_pdf = PymuDocDataset(bytes_pdf)

# 判断是否支持ocr
if dataset_pdf.classify() == SupportedPdfParseMethod.OCR:
    # 支持OCR
    infer_result = dataset_pdf.apply(doc_analyze, ocr=True)
    pipe_result = infer_result.pipe_ocr_mode(writer_image)
else:
    # 不支持OCR
    infer_result = dataset_pdf.apply(doc_analyze, ocr=False)
    pipe_result = infer_result.pipe_txt_mode(writer_image)

# 在每一页上都使用模型解析文本
infer_result.draw_model(pdf_file_path)

# 获取模型处理后的结果
model_inference_result = infer_result.get_infer_res()
print(model_inference_result)

# 为pdf生成含有颜色标注布局的pdf文件
pipe_result.draw_layout(f"{pdf_file_path_without_suff}_layout.pdf")

# 为pdf生成含有颜色标注文本行的pdf文件
pipe_result.draw_span(f"{pdf_file_path_without_suff}_spans.pdf")

# 获取markdown的内容
markdown_content = pipe_result.get_markdown(image_dir)
print(markdown_content)

# 保存markdown文件
# pipe_result.dump_md(writer_markdown, f"{pdf_file_path_without_suff}.md", image_dir)
pipe_result.dump_md(writer_markdown, f"{pdf_file_path_without_suff}.md", image_dir)

# json文本列表
# 数据类型包括type、text、text_level、page_idx、img_path等
content_list_content = pipe_result.get_content_list(image_dir)
print(content_list_content)

# 保存json文本列表
pipe_result.dump_content_list(writer_markdown, f"{pdf_file_path_without_suff}_content_list.json", image_dir)

# 获取含有位置信息的json文本
middle_json_content = pipe_result.get_middle_json()

# 保存含有位置信息的json文本
pipe_result.dump_middle_json(writer_markdown, f'{pdf_file_path_without_suff}_middle.json')