Sentiment analysis through song valence + lyrics

• Upgrade webpack-dev-server (+ webpack) to 4.x compatibility
  • https://www.npmjs.com/advisories/725

• dockerize before heroku (really, just run deploys fully through docker)

• setup Cassandra (tables, models, saving spotify data)
• setup user endpoints (TODO - swagger setup)
• swap base python to anaconda env
• add toolkits (keras, nltk, ..)
• investigate parallel requests (asyncio / tornado)

• H-S diagonal (pleasantness (valence) vs arousal (energy))

• Dataset collection (base dataset?) + (grab recents for user)
• Dimensionality Reduction using PCA
• Anomaly detection / pruning?

• Lyric scrape (genius api) (RUN ASNYC)
• keyword extraction + cleanup (contractions etc..)

• some fancypants ui idk

• Use Celery -> automated fetch for registered accounts
• Listen (tone/key, bpm) + week reports

• python 3.x

• yarn install
• pip3 install -r requirements.txt
• npm run start
• python manage.py sync_cassandra
• python manage.py seed # TODO? seed from pickles.
• python manage.py runserver
• Connect to localhost:8000

• ./manage.py sync_cassandra

• turn off all running Docker containers

  • docker-compose down

• delete any persistent data

  • rm -rf data/

• rebuild the images

  • docker-compose build

• start Cassandra

  • docker-compose up cassandra

• view cluster status

  • docker-compose run nodetool status

• create schema

  • docker-compose run cqlsh -f /schema.cql

• confirm schema

  • docker-compose run cqlsh -e "DESCRIBE SCHEMA;"

docker ps docker system prune docker exec -it 426c3e50d2b0 ip addr