README.md

InspaCy

Spacy pipeline running tagger, parser and Entity Ruler components, packaged in a web app using Flask.

Quick start

Set up a python virtual environment :

cd Inspacy
mkdir venv
python -m venv ./venv

Install dependencies :

source venv/bin/activate
pip install -r requirements.txt
python -m spacy download fr_core_news_md

Setup environment variables & run the app :

source init.sh
flask run

Go to localhost:5000/demo and enjoy !

What is it ?

A Python Web App for Named Entity Recognition using SpaCy !

The App is made up of three components : the pipeline itself, running both SpaCy components and custom components (in pipe.py), the knowledge base used in this pipeline, and the packaging in a web app (inspacy.py).

Here is a description of the pipeline :

It runs a custom component, Rule Builder, which takes as input an entity label, and generate a rule used by Entity Ruler, to recognize the entity in texts. Rules have several levels :

• strict : match the exact entity name.
• common : match groups of words which have the same part-of-speech tags and lemmas as the entity name.
• permis : same as common, but the group of word can contains additionnal adjectives