Giglist creates Spotify playlists containing the 10 ten tracks by artists you're going to see live.
Giglist has an Angular frontend hosted as a static site in an S3 bucket behind a CloudFront distribution. Gigs are stored in a DynamoDB table. Gig CRUD operations are performed via an API Gateway REST API, which triggers a Lambda function. The addition of a new gig triggers a Lambda function via DynamoDB streams. This Lambda adds songs to a user's upcoming gigs playlist using the Spotify API and schedules the deletion of these songs by setting up an EventBridge Scheduler schedule to trigger another Lambda function the day after the gig.
The backend Python source directories are functions (contains Lambda function code, organised by function) and libs
(contains Python code that is shared across > 1 Lambda). Set these as source directories in PyCharm to remove false
import errors.
The frontend Angular source is in frontend. It can be run locally using the Angular CLI:
cd frontend
npm install
ng serve --configuration {env}
FE deployment is currently manual. Run ng build --configuration {env} and copy the outputs of dist/frontend/browser into the static
site S3 bucket. Non-production sites will require a password for access. This is checked by the frontend_auth Cloud
Front function.
Tests are organised in the tests directory.
tests
├── integration
├── resources
└── unit
Tests are run with pytest. Running pytest in the root directory of the project will run all integration and unit tests.
CI/CD is performed using GitHub Actions. Workflows are located in .github/workflows.
The backend currently uses Python 3.12.
It can be deployed using the AWS SAM CLI:
cd infra
./scripts/populate-sam-template.py -c ../functions/frontend_auth/app.js
sam validate --lint -t template.inlined.yaml
sam build -t template.inlined.yaml
sam deploy --config-env dev
Manually created sources are:
- ACM certificates (because of DNS verification "chicken and egg" paradox in SAM)