Coffee Word Search

a RESTful api to find a word in a file or a string. Fancy-sauce hello world, amiright?

a.k.a a demo app on Spring Boot and kubernetes (MiniKube). Uses Spring Boot, Java 8, CircleCI, Gradle 5.0, Docker, Kubernetes and bring-your-own-cloud. Docker images are built in CI, pulled/tagged for deployment.

Check out the slides here.

Getting Started

Install all the dependencies

• Install Java8 + Gradle
• Install Docker + kubectl+ kops) + minikube and a hypervisor like (VirtualBox) for local development

Documentation

Read up on the design of this project in the CLASSES.md and CONFIG.md files. User docs for how to call the endpoints and what to expect are found below in the API Calls section and as the docs.html page.

Make a Change

Application changes should be written in Java 8, and separated by concern - site, api, or backend logic (txt and any potential siblings).

Before or during the change, add unit tests in the business logic package unit tests directory. New endpoints/pages should get controller and integration tests as well.

Local Testing

(1) Test manually by running as a jar. the api will be available on localhost:8080. Test plans consist of:

• curling each of the endpoints (/api, /api/string, /api/file), with appropriate data or parameters that will error (e.g. nonexistent files, missing param)
• visit each page in a browser (/, /docs, /resources)

(1) Test the containerized application locally by building the jar, building a dev docker container, and running locally. The site will be available at localhost:8080

$ gradle clean build
$ docker build -t acnagy/coffee-word-search:local .
$ docker run -p 8080:8080 acnagy/coffee-word-search:local

(2) Deploy the docker container a local kubernetes cluster with Minikube for local development and use $ minikube start to create a cluster. For production, create a separate cluster somewhere (e.g. with kopsor a hosted kubernetes service).

$ kubectl create -f deployment.yaml
$ kubectl create -f service.yaml  

The address of the cluster will either be the name of the load balancer created (in the cloud), or it will be the external IP that minikube service coffee-word-search knows about.

Build and Deploy

After making a change, run tests and build locally:

$ gradle clean build

to make sure everything works as it should. Then, commit to a branch and push.

CircleCI will automatically pick up the build, and following the .circle/config.yaml it will:

• run tests
• build a jar
• build a Docker image with that jar, tagging it with the version and build number
• push the Docker image to Docker Hub (acnagy/coffee-word-search)

To deploy, grab the image and re-tag with the minor version referenced in deployment.yaml

$ docker pull acnagy/coffee-word-search:BUILD_TAG
$ docker tag IMAGE_NAME acnagy/coffee-word-search:TAG
$ docker push acnagy/coffee-word-search:TAG

then, deploy to the kubernetes cluster with:

$ kubectl apply -f deployment.yaml

show the status of the roll-out and get more information via:

$ kubectl get deployment coffee-word-search
# or
$ kubectl describe deployment coffee-word-search

API Calls

The API supports three endpoints.

GET: URL/api/ params: none

$ curl -s URL/api/
{
  "term": null,
  "count": 1,
  "input": "Hi! Welcome to the app :)"
}

POST: URL/api/string/ params:

• term: string
• string: string (to search)

$ curl -s -X POST \
-d term="something" \
-d string="something something else" \
URL/api/string | jq

{
  "term": "something",
  "count": 2,
  "input": "something something else"
}

POST: URL/api/file/ params:

• term: string
• file: file path

$ curl -s -X POST \
-F term="something" \
-F file="@fileName" \
URL/api/file | jq

 {
   "term": "somethin",
   "count": 2,
   "input": "test-file3.txt"
 }