BookWorm is a full stack application allowing users to query bibliometric data, sourced from OpenAlex's data snapshot, and provides the option of viewing the resulting data in a graph-based data visualization.

On opening, users are presented with a blank canvas.

In the top left corner is a button to open the "Query Panel", which utilizes GraphiQL to provide an embedded GraphQL interface for creating queries, as well as documentation for the structure of the graph.

If the query is valid and successful, the API will return a JSON object that mirrors the structure of the request.

On closing the query panel, users can view and interact with the generated graph visualization.

This uses a GraphQL API for the query structure.

Requires Docker engine and docker compose.

Quickstart:

• Make a copy of .env.example and rename it to .env
• In the /rails folder, run bin/rails credentials:edit to generate a Rails master key. Copy the resulting key string to RAILS_MASTER_KEY in your .env file.
• Set POSTGRES_PASSWORD in the .env to a password that will be used for the database.
• Generate an SSL key and certificate (stackoverflow discussion here) and place both files in /nginx/certs. They must be named key.pem and certificate.pem.
• In the main directory, run docker compose build to build the rails and react docker images.
• In the main directory, run docker compose up to start the containers. Once the NGINX container has started, the application should be visible on [LOCALHOST-OR-YOUR-IP-ADDRESS]:80

To run a local development setup, you can run the rails development server and the react npm development server independently.

Requires a working installation of Postgres, npm, and bundler, and Ruby 3.2.2.

• If you do not have a ruby version manager, I recommend rbenv.
• If you do not have bundler installed, run gem install bundler
• For Postgres, follow the instructions in config/database.yml if you see an error related to pg on running bundle install

In the /rails folder, run npm i to install development dependencies. This repository uses prettier and the ruby-plugin for code formatting. I recommend activating format-on-save (https://www.digitalocean.com/community/tutorials/how-to-format-code-with-prettier-in-visual-studio-code).

In the /rails folder:

• Run bundle install to install gems.
• Run rails db:create to create the development and test databases.
• Run rails db:migrate to migrate the development and test databases.

To start the development server, run rails s. It should automatically start in development mode, but you can run rails s -e development if not.

The graphiql user interface should now be available at localhost:3001/graphiql.

In the /react folder:

• Run npm install to install node packages.
• Run npm run dev to launch the Vite/React frontend in development mode.

This repository uses RSpec for testing the rails application. (These need to be reworked as they haven't been maintained in recent changes).

In the /rails folder:

• Run rspec to run all of the rspec tests, or rspec spec/path/to/*_spec.rb to run a specific test.

Coverage reports are generated using simplecov, and will be in /rails/coverage.

The test suite is configured to use vcr for tests that interact with any external APIs, primarily the OpenAlex API. If you encounter errors related to vcr while running tests, try deleting the recorded 'cassette' files in spec/fixtures/vcr_cassettes/ and running rspec again to re-record them.

Planning on trying Jest for testing front end. WIP.

Data is sourced from OpenAlex's data snapshot, which also provides information about the structure of the source data and some basic instructions for uploading to a relational database, such as the Postgres database used in BookWorm.

There are rake tasks in /rails/lib/tasks/ for reading from CSV files to the database - this process could be optimized by running rake tasks in parallel or making adjustments to the insertion batch sizes in each file.

The data visualization is generated using a combination of React Flow for rendering nodes and edges, and D3 for calculating node positions.