/nougat

Implementation of Nougat Neural Optical Understanding for Academic Documents

Primary LanguagePythonMIT LicenseMIT

Nougat: Neural Optical Understanding for Academic Documents

Paper GitHub PyPI Python 3.9+ Code style: black Hugging Face Spaces

This is the official repository for Nougat, the academic document PDF parser that understands LaTeX math and tables.

Project page: https://facebookresearch.github.io/nougat/

Install

From pip:

pip install nougat-ocr

From repository:

pip install git+https://github.com/facebookresearch/nougat

Note, on Windows: If you want to utilize a GPU, make sure you first install the correct PyTorch version. Follow instructions here

There are extra dependencies if you want to call the model from an API or generate a dataset. Install via

pip install "nougat-ocr[api]" or pip install "nougat-ocr[dataset]"

Get prediction for a PDF

CLI

To get predictions for a PDF run

$ nougat path/to/file.pdf -o output_directory
usage: nougat [-h] [--batchsize BATCHSIZE] [--checkpoint CHECKPOINT] [--out OUT] [--recompute] [--markdown] pdf [pdf ...]

positional arguments:
  pdf                   PDF(s) to process.

options:
  -h, --help            show this help message and exit
  --batchsize BATCHSIZE, -b BATCHSIZE
                        Batch size to use.
  --checkpoint CHECKPOINT, -c CHECKPOINT
                        Path to checkpoint directory.
  --out OUT, -o OUT     Output directory.
  --recompute           Recompute already computed PDF, discarding previous predictions.
  --markdown            Add postprocessing step for markdown compatibility.

In the output directory every PDF will be saved as a .mmd file, the lightweight markup language, mostly compatible with Mathpix Markdown (we make use of the LaTeX tables).

API

With the extra dependencies you use app.py to start an API. Call

$ nougat_api

To get a prediction of a PDF file by making a POST request to http://127.0.0.1:8503/predict/. It also accepts parameters start and stop to limit the computation to select page numbers (boundaries are included).

The response is a string with the markdown text of the document.

curl -X 'POST' \
  'http://127.0.0.1:8503/predict/' \
  -H 'accept: application/json' \
  -H 'Content-Type: multipart/form-data' \
  -F 'file=@<PDFFILE.pdf>;type=application/pdf'

To use the limit the conversion to pages 1 to 5, ue the start/stop parameters in the request URL: http://127.0.0.1:8503/predict/?start=1&stop=5

Dataset

Generate dataset

To generate a dataset you need

  1. A directory containing the PDFs
  2. A directory containing the .html files (processed .tex files by LaTeXML) with the same folder structure
  3. A binary file of pdffigures2 and a corresponding environment variable export PDFFIGURES_PATH="/path/to/binary.jar"

Next run

python -m nougat.dataset.split_htmls_to_pages --html path/html/root --pdfs path/pdf/root --out path/paired/output --figure path/pdffigures/outputs

Additional arguments include

Argument Description
--recompute recompute all splits
--markdown MARKDOWN Markdown output dir
--workers WORKERS How many processes to use
--dpi DPI What resolution the pages will be saved at
--timeout TIMEOUT max time per paper in seconds
--tesseract Tesseract OCR prediction for each page

Finally create a jsonl file that contains all the image paths, markdown text and meta information.

python -m nougat.dataset.create_index --dir path/paired/output --out index.jsonl

For each jsonl file you also need to generate a seek map for faster data loading:

python -m nougat.dataset.gen_seek file.jsonl

The resulting directory structure can look as follows:

root/
├── images
├── train.jsonl
├── train.seek.map
├── test.jsonl
├── test.seek.map
├── validation.jsonl
└── validation.seek.map

Note that the .mmd and .json files in the path/paired/output (here images) are no longer required. This can be useful for pushing to a S3 bucket by halving the amount of files.

Training

To train or fine tune a Nougat model, run

python train.py --config config/train_nougat.yaml

Evaluation

Run

python test.py --checkpoint path/to/checkpoint --dataset path/to/test.jsonl --save_path path/to/results.json

To get the results for the different text modalities, run

python -m nougat.metrics path/to/results.json

Citation

@misc{blecher2023nougat,
      title={Nougat: Neural Optical Understanding for Academic Documents}, 
      author={Lukas Blecher and Guillem Cucurull and Thomas Scialom and Robert Stojnic},
      year={2023},
      eprint={2308.13418},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}

Acknowledgments

This repository builds on top of the Donut repository.

License

Nougat codebase is licensed under MIT.

Nougat model weights are licensed under CC-BY-NC.