Purpose

I created this project to demonstrate how to setup a FastAPI application to act as a caching server for the GitHub related queries. I took a containerized approach to solving this problem by spinning up one (or more) FastAPI containers and one redis container.

Environment Variables

The GitHub REST API is rate limited heavily if you don't provide a GITHUB_API_TOKEN. In order to ensure that we don't hit the most conservative rate limit, we should pass this API token. This can be done by setting the variable with the command above.

Additionally, you can also set the APP_PORT environment variable to access the application.

GITHUB_API_TOKEN=YOUR_TOKEN_HERE APP_PORT=8000  docker-compose up --build

Getting Started

To run this application, the recommended approach is to run it through Docker. The root of this project contains a docker-compose.yml file that spins up the necessary containers.

To set the port you would like to access the application on, you can edit the docker-compose configuration file to map to a different port. Currently it is being defaulted to port 8000 if it does not pick up the APP_PORT environment variable on startup. You can also hard code the port value into the docker-compose file if you prefer to avoid setting an environment variable.

server:
  container_name: server
  build:
    context: ./server
    dockerfile: Dockerfile
  ports:
    - "${APP_PORT:-8000}:8000"
  environment:
    - GITHUB_API_TOKEN=${GITHUB_API_TOKEN}

To run the application, you can simply run the command below from the root directory

GITHUB_API_TOKEN=YOUR_TOKEN_HERE APP_PORT=8000  docker-compose up --build

Once the containers are running, we can use the docs page for the FastAPI server to see all of our endpoints that have been defined. You can reach this page through your broswer by going to http://localhost:8000/docs if you have not changed the default port. The page should look like the below screenshot.

Running Tests

Tests are provided for this project to test the functionality of our endpoints. This test suite is defined in the api-suite.sh file.

The tests in the suite may have drifted a bit over time so the results for the created views tests may fail since those metrics have changed over time. You can still validate that everything is working by looking at the expected vs response output in the console and see that the response values for this application follow the same pattern.

Development Mode

If you would like to develop additional features, run the docker-compose.dev.yml file instead. This will spin up a FastAPI container in development mode with volumes mounted into the container to allow for hot reloading!

docker-compose -f docker-compose.dev.yml up --build

Interacting with Redis

I have added two endpoints that can be helpful for interacting with the redis cache as well. There is a /clear-cache endpoint that can be used to clear out the contents of the cache. Any subsequent requests after this is called will result in refreshing the cache. There is also a /cached-keys endpoint that returns all of the keys we have in our redis layer. If all keys are cached correctly, the response will look like the below. To the right of the keys, you can find the endpoint from GitHub that is being cached.

[
  "ORG_MEMBERS", # orgs/Netflix/members
  "ORG_REPOS",   # orgs/Netflix/repos
  "ORGS",        # orgs/Netflix
  "BASE",        # /
]