TextProc

Toy REST API for basic text processing

This was made as a weekend project to learn FastAPI/Flask and basic NLP.

It's a FastAPI API with CRUD capabilities that can also do some basic text processing.

• Using SQLAlchemy/Pydantic for DB
• And simple text processing functions that use NLTK/Gensim.

There's also a flask version (flask.py):

• Using SQLAlchemy and Flask-Marshmallow for DB
• Flask-RESTful for the actual REST resources
• And simple text processing functions that use NLTK.

Create env

conda env create -f environment.yml

import nltk

nltk.download('punkt_tab')
nltk.download('stopwords')
nltk.download('wordnet')

Init db and run api:

uvicorn src.app:app --reload

Usage examples (cURL):

• Create new user:

curl -X POST http://127.0.0.1:8000/users \
     -H "Content-Type: application/json" \
     -d '{"name":"João","content":"I must explain to you how all this mistaken idea of denouncing pleasure and praising pain was born and I will give you a complete account of the system, and expound the actual teachings of the great explorer of the truth, the master-builder of human happiness."}'

• Get user 1's bag-of-words:

curl --header "Content-Type: application/json" --request GET  http://localhost:8000/users/1/getbow

• Get the word most similar to 'pleasure' in user 1's vocabulary:

curl -X POST http://127.0.0.1:8000/users/1/wordvec \
     -H "Content-Type: application/json" \
     -d '{"word": "pleasure"}'