Overview

This is a demo application written in Go to showcase Azure Cognitive Services running in Docker Containers. It is a websocket chat web application. It parses all messages using Azure Cognitive Services Sentimnet Analysis. If the sentiment is below a set threshold (ie. the message is too mean) then the AI Chat Moderator will broadcast to everyone to be nice.

This demo uses the Azure Cognitive Services container deployment. This means that all analysis is done locally and not in Azure. A Cognitive Services deployment in Azure is still required for billing purposes but all the analysis is done locally.

Azure Setup

Prerequisites

• AKS Cluster Created. See kubernetes-cluster-setup
• Azure Storage Account for Terraform state storage
• Azure Container Repository for Helm Chart
• Docker Hub Repository for the docker container (could re-use Azur Container Repository)

GitHub Actions

• Excute .github/workflows/infrastructure.yaml

Manual

    cd infrastructure
    terraform init -backend=true -backend-config="access_key=WY4ruq........" -backend-config="key=app.terraform.tfstate"
    terraform plan -out="uat.plan" -var "resource_group_name=DevSub02_ChatApp_RG" -var-file="uat.tfvars"
    terraform apply "uat.plan"

Code Build

Local

    #Set your GOPATH 
    go get github.com/gorilla/websocket
    go get github.com/gin-contrib/cors
    go get github.com/gin-gonic/gin
    go get github.com/gin-gonic/contrib/static
    go get gopkg.in/resty.v1
    cd ./src
    docker build -t chatw
    cd ./chart
    helm package .
    az acr helm push -n ${ACR_NAME} sentimentchat-0.2.${VERAION}.tgz --force

GitHub Actions

• Excute .github/workflows/build.yaml

Deploy to Kubernetes

GitOps

1. The repository is configured to deploy this application via flux