Graphql version of a tiny part of a cinephiles' app
The app is a social media platform where people can book watch parties for movies, review and share them.
So the app has different home tab screens:
- Movies List
- Movie Details Page
- Reviews Section
- My Bookings
Parts of different models render on different screens as well. Fetching the apt content for different screens fits a good use case of using GraphQL.
- Build the docker image:
./docker_build.sh - Run the container
docker run -d -p 5000:5000 --name=binge-graphql binge-graphql
Note: The DB would be ephemeral just like the container as it'd a sqlite DB within the container. If we want to make it persistent we can mount a volume on to the container and adjust the db path in app accordingly.
Access GraphiQL interface on host's http://127.0.0.1:5000/graphql
You can play around with the data using some of the queries mentioned in test_queries dir.
- GraphQL - Awesome way to give more control to client and have a clear schema
- HTTP verbs have a reduced meaning in GraphQL compared to REST APIs. Almost all requests (both Query and Mutations) goes through POST
- Relay: Nodes/edges concept for pagination, sorting, filtering etc.
- interfaces, unions, fragments
- sqlalchemy - I've used Django models extensively, but this is more lower level and relationships like FK are handled differently (also, composite keys)
- graphene - Python library for supporting GraphQL. It has all the flexibility for further extensions and stuff
- graphene_sqlalchemy - A wrapper to provide out-of-the-box G-QL objects if we have sqlalchemy models
- User based serialized fields in response? Is there a better way than using Unions?
- Different mutations for Create, Update and Delete?