Image Caption Generation

See extra documentation here

Build and deploy cycle

The parts from 7 to 10 can be skipped if you don't want update the container images

1. Update the source code

git pull

2. Make image-caption-generation the working namespace

kubectl config get-context --current --namespace=image-caption-generation

3. Remove any previous deployment

kubectl delete namespace image-caption-generation

4. Recreate an empty namespace for the new deployment

kubectl create namespace image-caption-generation

5. Install RabbitMQ. Notice that you need to install the bitnami repository before.

helm install -f rabbitmq.yaml -n image-caption-generation img-rabbitmq bitnami/rabbitmq --version 12.12.0

6. Install Infinispan. Notice that you need to install openshift repository before.

helm install -f infinispan.yaml -n image-caption-generation infinispan openshift/infinispan-infinispan --version 0.3.2

7. Optionally Compile the Java/Quarkus gallery service:

mvn -f ./gallery/pom.xml clean package

8. Optionally Compile the Java/Quarkus user service:

mvn -f ./users/pom.xml clean package

9. Optionally Build the Docker images locally:

docker-compose build

10. Optionally Push the Docker images to the Docker remote repository:

docker-compose push

11. Deploy the application, Kubernetes will pull the images from the remote repository:

kubectl apply -f kubernetes.yaml

12. See all the pods starting...

kubectl get pods -w