Autocomplete Solution

This includes example code enough to run an autocompletion example on GKE.

I don't believe this is asked any more by candidates, so I've made it public. It's not best practise or good code, I'd re-write this if I were to do it again.

It is primarily written in Spring Boot and comprises of three microservices:

  1. Autocomplete Map Generator - autocomplete-map/ - generates autocomplete mappings and stores them in Google Cloud Storage - triggers on a pub/sub notification from Cloud Storage
  2. Autocomplete Microservice - autocomplete/ - builds an in-memory map of autocomplete mappings and serves them - replaces the in-memory database when alerted by cloud storage
  3. Presentation Layer - presentation/ - provides the Javascript and so on for the app

Compiling / Building Docker images

To get up and running customise each app's src/main/resources/application.yml file for your needs and then run the following for each microservice:

# Replace this with your docker prefix
DOCKER_PREFIX=mattsday

pushd autocomplete;
(
	mvn clean package dockerfile:build -DskipTests
	docker push ${DOCKER_PREFIX}/autocomplete
)& wait
popd
pushd autocomplete-map
(
	mvn clean package dockerfile:build -DskipTests
	docker push ${DOCKER_PREFIX}/autocomplete-map
)& wait
popd
pushd presentation
(
	mvn clean package dockerfile:build -DskipTests
	docker push ${DOCKER_PREFIX}/presentation
)& wait
popd

Running in k8s

kubectl create -f interview.yml