---
license: openrail
---


<h3 align="center">PDF Paragraphs Extraction</h3>
<p align="center">A model for extracting paragraphs from PDFs</p>

This model uses features from the PDF to extract the text and paragraphs from it. It can be used as a service. 

The paragraphs contain the page number, the position in the page, the size, and the text.

We have created the better and more flexible version of this service, you can check here:
  
    https://huggingface.co/HURIDOCS/pdf-document-layout-analysis


## Quick Start

Download the service that uses the model:

    git clone https://github.com/huridocs/pdf_paragraphs_extraction.git
    cd pdf_paragraphs_extraction

Start the service:

    ./run start

Get the paragraphs from a PDF:

    curl -X GET -F 'file=@/PATH/TO/PDF/pdf_name.pdf' localhost:5051

To stop the server:

    ./run stop


## Performance

Accuracy: 93.9%

Speed: 0.15 seconds per page