drinksfinder

• The api_server directory contains the source code and Dockerfile for the API server.
• The data_access_layer directory contains the source code and Dockerfile for the "database". Note, this is just for PoC purposes; in real life this would have multiple layers including a cache (simulated in the code by some pre-populated tables), an abstraction layer and a back-end database.
• The terraform directory contains Terraform code to deploy an AKS cluster in Azure, and adds a namespace and an optional Google Maps API key to the cluster.
• The kubernetes directory contains Kubernetes manifests to be applied to the AKS cluster.

To deploy:

• From the terraform directory, run terraform init and terraform apply. Enter variables when prompted. If you don't have a Google Maps API key, leave this blank but note that the /api/drinksfinder/v1/pubs/near/postcode/<postcode> path will be unavailable.
• Once the cluster is deployed, using the kubeconfig settings for the new cluster, deploy the manifests in kubernetes (i.e. kubectl apply -f kubernetes/)

To access the API:

• Find the endpoint IP address with kubectl -n drinksfinder get service apiserver -o jsonpath='{.status.loadBalancer.ingress[0].ip}'}
• Attempt the following requests:
  • curl http://<IP>/api/drinksfinder/v1/pubs - list of pubs
  • curl http://<IP>/api/drinksfinder/v1/pubs?start=3&limit=5 - pagination
  • curl http://<IP>/api/drinksfinder/v1/pubs?order_by=beer - sorting (supports "beer", "atmosphere", "amenities" and "value")
  • curl http://<IP>/api/drinksfinder/v1/pubs?tag=karaoke&tag="free wifi" - must match all tags
  • curl http://<IP>/api/drinksfinder/v1/pubs/near - orders by distance from "primary location" (if you're reading this, you probably know where that is)
  • curl http://<IP>/api/drinksfinder/v1/pubs/near/postcode/<postcode> - orders by distance from postcode. This requires a Google Maps API key to be provided to the terraform apply.

Pagination and tags work with all searches. Ordering and location-based searches are mutually exclusive. If-None-Match/ETag caching is supported.