/acolyte

API and web app that allows business users to ask data questions in plain English

Primary LanguageVueMIT LicenseMIT

Acolyte

Demo / Website

Self-service data analytics powered by GPT-3

Screenshot

About

Acolyte is self-service business analytics. Users can ask questions like "how many customers do we have?" or "who placed the most orders last week?" in plain English and get the right answer from their data warehouse.

Deployment

There are two separate applications to deploy (the api and client). The easiest way to deploy is on Heroku via the following steps:

  1. Create a new API Heroku app and a new Client heroku app.
  2. In the API app (via the Heroku command line or the Heroku web UI):
    1. add the heroku/python buildpack.
    2. Add a DATABASE_URL environment variable with the full URI to your database (i.e. postgres://USER:PASSWORD@HOSTNAME:PORT/DATABASE)
    3. Add an OPENAI_API_KEY environment variable with your OpenAI API key.
  3. In the Client app (via the Heroku command line or the Heroku web UI):
    1. Add the heroku/nodejs buildpack and the https://github.com/heroku/heroku-buildpack-static buildpack.
    2. Add a VUE_APP_API_ENDPOINT with the url to your API (e.g. https://acolyte-api.herokuapp.com/api)
    3. Optionally, if you want to track usage using Heap, add a VUE_APP_HEAP_ID environment variable with your Heap environment ID.
  4. In this directory, via the Heroku command line, run the following:
heroku git:remote --remote heroku-client -a YOUR_CLIENT_HEROKU_APP_NAME
heroku git:remote --remote heroku-server -a YOUR_API_HEROKU_APP_NAME
git subtree push --prefix client heroku-client master
git subtree push --prefix api heroku-server master