NER (Named Entity Recognition)

This repository implements a toy example on how to find sensitive information in a PDF and highlight it.

Requirements

to install the requirements use pip install -r requirments.txt

Install new NER models (spacy)

Spacy allows loading different pre-trained models. Before running the scrips in this repository you'd need to install the German package: python -m spacy download de_core_news_sm

Run the algo

Simply have a PDF document ready and run python process_document --input_pdf_path <you-path>

This will generate an output.pdf document with sensitive information highlighted. To extend the script (e.g., you want to blur/hide the sensitive information), modify the code in _draw_rectangles_on_pdf(...)

Examples

See docs

Have fun!