[](https://github.com/opendatalab/MinerU)
[](https://github.com/opendatalab/MinerU)
[](https://github.com/opendatalab/MinerU/issues)
[](https://github.com/opendatalab/MinerU/issues)
[](https://badge.fury.io/py/magic-pdf)
[](https://pepy.tech/project/magic-pdf)
[](https://pepy.tech/project/magic-pdf)
[](https://opendatalab.com/OpenSourceTools/Extractor/PDF)
[](https://huggingface.co/spaces/opendatalab/MinerU)
[](https://www.modelscope.cn/studios/OpenDataLab/MinerU)
[](https://colab.research.google.com/gist/papayalove/b5f4913389e7ff9883c6b687de156e78/mineru_demo.ipynb)
[](https://arxiv.org/pdf/2409.18839?)
[English](README.md) | [็ฎไฝไธญๆ](README_zh-CN.md)
PDF-Extract-Kit: High-Quality PDF Extraction Toolkit๐ฅ๐ฅ๐ฅ
๐ join us on Discord and WeChat
# Changelog
- 2024/10/31 0.9.0 released. This is a major new version with extensive code refactoring, addressing numerous issues, improving performance, reducing hardware requirements, and enhancing usability:
- Refactored the sorting module code to use [layoutreader](https://github.com/ppaanngggg/layoutreader) for reading order sorting, ensuring high accuracy in various layouts.
- Refactored the paragraph concatenation module to achieve good results in cross-column, cross-page, cross-figure, and cross-table scenarios.
- Refactored the list and table of contents recognition functions, significantly improving the accuracy of list blocks and table of contents blocks, as well as the parsing of corresponding text paragraphs.
- Refactored the matching logic for figures, tables, and descriptive text, greatly enhancing the accuracy of matching captions and footnotes to figures and tables, and reducing the loss rate of descriptive text to near zero.
- Added multi-language support for OCR, supporting detection and recognition of 84 languages.For the list of supported languages, see [OCR Language Support List](https://paddlepaddle.github.io/PaddleOCR/latest/en/ppocr/blog/multi_languages.html#5-support-languages-and-abbreviations).
- Added memory recycling logic and other memory optimization measures, significantly reducing memory usage. The memory requirement for enabling all acceleration features except table acceleration (layout/formula/OCR) has been reduced from 16GB to 8GB, and the memory requirement for enabling all acceleration features has been reduced from 24GB to 10GB.
- Optimized configuration file feature switches, adding an independent formula detection switch to significantly improve speed and parsing results when formula detection is not needed.
- Integrated [PDF-Extract-Kit 1.0](https://github.com/opendatalab/PDF-Extract-Kit):
- Added the self-developed `doclayout_yolo` model, which speeds up processing by more than 10 times compared to the original solution while maintaining similar parsing effects, and can be freely switched with `layoutlmv3` via the configuration file.
- Upgraded formula parsing to `unimernet 0.2.1`, improving formula parsing accuracy while significantly reducing memory usage.
- Due to the repository change for `PDF-Extract-Kit 1.0`, you need to re-download the model. Please refer to [How to Download Models](docs/how_to_download_models_en.md) for detailed steps.
- 2024/09/27 Version 0.8.1 released, Fixed some bugs, and providing a [localized deployment version](projects/web_demo/README.md) of the [online demo](https://opendatalab.com/OpenSourceTools/Extractor/PDF/) and the [front-end interface](projects/web/README.md).
- 2024/09/09: Version 0.8.0 released, supporting fast deployment with Dockerfile, and launching demos on Huggingface and Modelscope.
- 2024/08/30: Version 0.7.1 released, add paddle tablemaster table recognition option
- 2024/08/09: Version 0.7.0b1 released, simplified installation process, added table recognition functionality
- 2024/08/01: Version 0.6.2b1 released, optimized dependency conflict issues and installation documentation
- 2024/07/05: Initial open-source release
| Operating System |
| Ubuntu 22.04 LTS |
Windows 10 / 11 |
macOS 11+ |
| CPU |
x86_64(unsupported ARM Linux) |
x86_64(unsupported ARM Windows) |
x86_64 / arm64 |
| Memory |
16GB or more, recommended 32GB+ |
| Python Version |
3.10(Please make sure to create a Python 3.10 virtual environment using conda) |
| Nvidia Driver Version |
latest (Proprietary Driver) |
latest |
None |
| CUDA Environment |
Automatic installation [12.1 (pytorch) + 11.8 (paddle)] |
11.8 (manual installation) + cuDNN v8.7.0 (manual installation) |
None |
| GPU Hardware Support List |
Minimum Requirement 8G+ VRAM |
3060ti/3070/4060
8G VRAM enables layout, formula recognition acceleration and OCR acceleration |
None |
| Recommended Configuration 10G+ VRAM |
3080/3080ti/3090/3090ti/4070/4070ti/4070tisuper/4080/4090
10G VRAM or more can enable layout, formula recognition, OCR acceleration and table recognition acceleration simultaneously
|
### Online Demo
Stable Version (Stable version verified by QA):
[](https://opendatalab.com/OpenSourceTools/Extractor/PDF)
Test Version (Synced with dev branch updates, testing new features):
[](https://huggingface.co/spaces/opendatalab/MinerU)
[](https://www.modelscope.cn/studios/OpenDataLab/MinerU)
### Quick CPU Demo
#### 1. Install magic-pdf
```bash
conda create -n MinerU python=3.10
conda activate MinerU
pip install -U magic-pdf[full] --extra-index-url https://wheels.myhloli.com
```
#### 2. Download model weight files
Refer to [How to Download Model Files](docs/how_to_download_models_en.md) for detailed instructions.
#### 3. Modify the Configuration File for Additional Configuration
After completing the [2. Download model weight files](#2-download-model-weight-files) step, the script will automatically generate a `magic-pdf.json` file in the user directory and configure the default model path.
You can find the `magic-pdf.json` file in your ใuser directoryใ.
> The user directory for Windows is "C:\\Users\\username", for Linux it is "/home/username", and for macOS it is "/Users/username".
You can modify certain configurations in this file to enable or disable features, such as table recognition:
> If the following items are not present in the JSON, please manually add the required items and remove the comment content (standard JSON does not support comments).
```json
{
// other config
"layout-config": {
"model": "layoutlmv3" // Please change to "doclayout_yolo" when using doclayout_yolo.
},
"formula-config": {
"mfd_model": "yolo_v8_mfd",
"mfr_model": "unimernet_small",
"enable": true // The formula recognition feature is enabled by default. If you need to disable it, please change the value here to "false".
},
"table-config": {
"model": "tablemaster", // When using structEqTable, please change to "struct_eqtable".
"enable": false, // The table recognition feature is disabled by default. If you need to enable it, please change the value here to "true".
"max_time": 400
}
}
```
### Using GPU
If your device supports CUDA and meets the GPU requirements of the mainline environment, you can use GPU acceleration. Please select the appropriate guide based on your system:
- [Ubuntu 22.04 LTS + GPU](docs/README_Ubuntu_CUDA_Acceleration_en_US.md)
- [Windows 10/11 + GPU](docs/README_Windows_CUDA_Acceleration_en_US.md)
- Quick Deployment with Docker
> Docker requires a GPU with at least 16GB of VRAM, and all acceleration features are enabled by default.
>
> Before running this Docker, you can use the following command to check if your device supports CUDA acceleration on Docker.
>
> ```bash
> docker run --rm --gpus=all nvidia/cuda:12.1.0-base-ubuntu22.04 nvidia-smi
> ```
```bash
wget https://github.com/opendatalab/MinerU/raw/master/Dockerfile
docker build -t mineru:latest .
docker run --rm -it --gpus=all mineru:latest /bin/bash
magic-pdf --help
```
## Usage
### Command Line
```bash
magic-pdf --help
Usage: magic-pdf [OPTIONS]
Options:
-v, --version display the version and exit
-p, --path PATH local pdf filepath or directory [required]
-o, --output-dir PATH output local directory [required]
-m, --method [ocr|txt|auto] the method for parsing pdf. ocr: using ocr
technique to extract information from pdf. txt:
suitable for the text-based pdf only and
outperform ocr. auto: automatically choose the
best method for parsing pdf from ocr and txt.
without method specified, auto will be used by
default.
-l, --lang TEXT Input the languages in the pdf (if known) to
improve OCR accuracy. Optional. You should
input "Abbreviation" with language form url: ht
tps://paddlepaddle.github.io/PaddleOCR/latest/en
/ppocr/blog/multi_languages.html#5-support-languages-
and-abbreviations
-d, --debug BOOLEAN Enables detailed debugging information during
the execution of the CLI commands.
-s, --start INTEGER The starting page for PDF parsing, beginning
from 0.
-e, --end INTEGER The ending page for PDF parsing, beginning from
0.
--help Show this message and exit.
## show version
magic-pdf -v
## command line example
magic-pdf -p {some_pdf} -o {some_output_dir} -m auto
```
`{some_pdf}` can be a single PDF file or a directory containing multiple PDFs.
The results will be saved in the `{some_output_dir}` directory. The output file list is as follows:
```text
โโโ some_pdf.md # markdown file
โโโ images # directory for storing images
โโโ some_pdf_layout.pdf # layout diagram (Include layout reading order)
โโโ some_pdf_middle.json # MinerU intermediate processing result
โโโ some_pdf_model.json # model inference result
โโโ some_pdf_origin.pdf # original PDF file
โโโ some_pdf_spans.pdf # smallest granularity bbox position information diagram
โโโ some_pdf_content_list.json # Rich text JSON arranged in reading order
```
For more information about the output files, please refer to the [Output File Description](docs/output_file_en_us.md).
### API
Processing files from local disk
```python
image_writer = DiskReaderWriter(local_image_dir)
image_dir = str(os.path.basename(local_image_dir))
jso_useful_key = {"_pdf_type": "", "model_list": []}
pipe = UNIPipe(pdf_bytes, jso_useful_key, image_writer)
pipe.pipe_classify()
pipe.pipe_analyze()
pipe.pipe_parse()
md_content = pipe.pipe_mk_markdown(image_dir, drop_mode="none")
```
Processing files from object storage
```python
s3pdf_cli = S3ReaderWriter(pdf_ak, pdf_sk, pdf_endpoint)
image_dir = "s3://img_bucket/"
s3image_cli = S3ReaderWriter(img_ak, img_sk, img_endpoint, parent_path=image_dir)
pdf_bytes = s3pdf_cli.read(s3_pdf_path, mode=s3pdf_cli.MODE_BIN)
jso_useful_key = {"_pdf_type": "", "model_list": []}
pipe = UNIPipe(pdf_bytes, jso_useful_key, s3image_cli)
pipe.pipe_classify()
pipe.pipe_analyze()
pipe.pipe_parse()
md_content = pipe.pipe_mk_markdown(image_dir, drop_mode="none")
```
For detailed implementation, refer to:
- [demo.py Simplest Processing Method](demo/demo.py)
- [magic_pdf_parse_main.py More Detailed Processing Workflow](demo/magic_pdf_parse_main.py)
### Deploy Derived Projects
Derived projects include secondary development projects based on MinerU by project developers and community developers,
such as application interfaces based on Gradio, RAG based on llama, web demos similar to the official website, lightweight multi-GPU load balancing client/server ends, etc.
These projects may offer more features and a better user experience.
For specific deployment methods, please refer to the [Derived Project README](projects/README.md)
### Development Guide
TODO
# TODO
- ๐น Reading order based on the model
- ๐น Recognition of `index` and `list` in the main text
- ๐น Table recognition
- โ Code block recognition in the main text
- โ [Chemical formula recognition](docs/chemical_knowledge_introduction/introduction.pdf)
- โ Geometric shape recognition
# Known Issues
- Reading order is determined by the model based on the spatial distribution of readable content, and may be out of order in some areas under extremely complex layouts.
- Vertical text is not supported.
- Tables of contents and lists are recognized through rules, and some uncommon list formats may not be recognized.
- Only one level of headings is supported; hierarchical headings are not currently supported.
- Code blocks are not yet supported in the layout model.
- Comic books, art albums, primary school textbooks, and exercises cannot be parsed well.
- Table recognition may result in row/column recognition errors in complex tables.
- OCR recognition may produce inaccurate characters in PDFs of lesser-known languages (e.g., diacritical marks in Latin script, easily confused characters in Arabic script).
- Some formulas may not render correctly in Markdown.
# FAQ
[FAQ in Chinese](docs/FAQ_zh_cn.md)
[FAQ in English](docs/FAQ_en_us.md)
# All Thanks To Our Contributors