m2rads/AskAnyBook

Setup

1. Create and fill in .env using .env.example as an example.

2. Install required gems

bundle install

Install required yarn packages

yarn install

3. Turn your PDF into embeddings for GPT-3:

ruby ./lib/scripts/pdf_to_pages_embeddings.rb --pdf /your/path/to/book.pdf

Note

This script uses pdftotext gem to read pdf files in ruby. In order to use this gem you need to install Poppler on your machine.

To install Poppler using Homebrew:

brew install poppler

For more information on installing Poppler refer to: https://www.viget.com/articles/extract-embedded-text-from-pdfs-with-poppler-in-ruby/

4. Set up database tables and initialize seed data:

rails db:migrate
rails db:seed

Run locally

To run the rails server over port 3000 for this app open a terminal window and type:

rails s

And to run webpack dev server open another terminal window and run

./bin/webpack-dev-server

Deploy to Heroku

1. Create a new heroku app

heroku create my-app

2. Set heruko as remote repo

git push herku main

3. Last step is to setup database

heroku db:migrate
heroku db:seed

Note that you would need to create your embeddings and place them in ./lib/assets/book.pdf

Happy hacking