This repo contains code for the Offensive Language Detection Model API, the requirements, tests as well as other set up files.

The API uses the offensive_language_detection_model created by me (link).

Find below the description of the different components & their respective uses:

app:

This is the main package in this directory and It contains a number of files that is used to develop our API.

1. Schemas: This folder contains files that are used to define the format of the inputs or outputs of the API's endpoints. There are two schema files: - predict.py: contains schema for the format of the inputs & outputs for the predict endpoint - health.py: contains schema for the format of the response of the health endpoint
2. api.py:This file contains code for the API's endpoints: health & predict.
3. config.py: This contains code for configuring the API like setting the logging level, setting up custome logging etc
4. main.py: This file contains code for creating & running the API
5. tests: contains scripts for testing the package
   • conftest.py - contains a fixture fxn which is used to provide input text to the other test.
   • test_api.py - used to test the predict endpoint of the api

Dockerfile

Contains commands used to build the docker image for this api. The image can be gotten by running "docker pull olaitanlawal/forecast_api"

requirements.txt:

Contains dependencies required to use the API

Procfile:

This file is a requirement from heroku when deploying an app to heroku. It tells heroku how to run the application

Manifest.in:

Contains instructions for what to include or exclude when building the package

runtime.txt:

File to tell heroku what version of language(python) and version was used to create the API

test_requirements.txt:

Contains dependencies required to test the API

tox.ini:

This file contains the settings for using tox for automated test but I didnt use tox to test this API.

Testing:

• To test this API, run the command "pytest" on cmd

How to use this API :

**To run the different files and test the modules, I created a conda environment for this project, installed all the dependencies in the requirements.txt except prophet. and then followed the steps in this link to install prophet. https://stackoverflow.com/questions/53178281/installing-fbprophet-python-on-windows-10 ** Install the packages in requirements.txt file ** If testing or using the API file on your computer, you have to add the directory to your PYTHONPATH, So that python can find it. Search for 'andrei' on https://stackoverflow.com/questions/3402168/permanently-add-a-directory-to-pythonpath . If you don't do this, python will not be able to find the package and so will not be able to run the import statements

Deploying:

This API has been deployed to heroku. (https://cad-usd-forecast-api.herokuapp.com/)