Tatodet for language detection/classification

This is a standalone python microservice to classify text into language classes using models trained on tatoeba's data

Requirements

Mainly pymc3, numpy, and sklearn. However many other packages are used and many more will be added in the future.

pip install -r requirements.txt

Make sure the model is built:

python3 build_model.py

To run the webapi:

python3 api.py

then send a query:

curl -X GET "http://localhost:8080/v1/det?sent=what+is+it&trials=20"

To run tests:

python3 -m pytest