CONTEXT: This is a follow-up application to Sentiment CLI.
Train our Robots is my attempt at integrating a ML algorithm into a web app.
HOW IT WORKS: Robots need a tool to help them understand the sentiment of human language. Train Our Robots does just that!
Robots can enter a sentence they heard from a human and get an instant sentiment analysis! Human users can help improve the model in real time by first adding sentences to the database, then by clicking "Train the model"!
STEP 1: Clone the project (in the terminal): git clone https://github.com/tjdolan121/train_our_robots.git
STEP 2: Create a new virtual environment: virtualenv venv
STEP 3: Activate the virtual environment: source venv/bin/activate
STEP 4: Navigate to the project directory, then install requirements: pip install -r requirements.txt
STEP 5: Set up flask environment variables: (macOS)export FLASK_APP=sentiment_app.pyexport FLASK_ENV=development
STEP 6: Instantiate a database and run migrations: flask db upgrade
STEP 7: Seed the database with starter data for the ML model: python setup.py
STEP 8: Run server: flask run
STEP 10: Navigate to http://127.0.0.1:5000 in browser, create an account, and play around!
Commits to master are automatically pushed to the staging app, found here: https://intense-depths-98176.herokuapp.com/
To get up and running with Flask, I followed this amazing tutorial: https://blog.miguelgrinberg.com/post/the-flask-mega-tutorial-part-iv-database
The sentiment data came from here: https://archive.ics.uci.edu/ml/datasets/Sentiment+Labelled+Sentences