Azure Rambi

Welcome to the Azure Rambi project! This repository contains code and resources for managing and deploying the Rambi Application using Azure services.

Table of Contents

• Introduction
• Getting Started
• Installation
• Usage
• Contributing
• License

Introduction

Azure Rambi is a project designed to simplify the management and deployment of Azure resources. It provides a set of tools and scripts to automate common tasks.

Getting Started

To get started with Azure Rambi, follow these steps:

1. Clone the repository:

   git clone https://github.com/bmoussaud/azure-rambi.git

2. Navigate to the project directory:

   cd azure-rambi

Deployment

Install the required dependencies by running:

az login
azd auth login
azd up           
? Enter a new environment name: dev
? Select an Azure Subscription to use: 1. fake-subscription (00000000-0000-0000-0000-000000000000)
? Select an Azure location to use: 21. (Europe) France Central (francecentral)
? Pick a resource group to use: Create a new resource group
? Enter a name for the new resource group: azrambi-dev

Warning: the provision phases include an APIM and A Redis Cache that can take a veryyyyyyy loooooong time to become available. Patience ! if it fails run the azd up command again.

 (✓) Done: Key Vault: rambikveab45rexk4hhs (689ms)
  (✓) Done: Azure OpenAI: azrambi-openai-eab45rexk4hhs (695ms)
  (✓) Done: Container Registry: azurerambieab45rexk4hhs (719ms)
  (✓) Done: Storage account: azrambieab45rexk4hhs (723ms)
  (✓) Done: Azure AI Services Model Deployment: azrambi-openai-eab45rexk4hhs/gpt-4o (570ms)
  (✓) Done: Azure AI Services Model Deployment: azrambi-openai-eab45rexk4hhs/o1-mini (584ms)
  (✓) Done: Azure AI Services Model Deployment: azrambi-openai-eab45rexk4hhs/dall-e-3 (991ms)
  (✓) Done: Log Analytics workspace: azure-rambi-log-eab45rexk4hhs (580ms)
  (✓) Done: Cache for Redis: azure-rambi-redis-eab45rexk4hhs (1.846s)
  (✓) Done: Application Insights: azure-rambi-appIn-eab45rexk4hhs (1.4s)
  (✓) Done: Container Apps Environment: azure-rambi (917ms)
  (✓) Done: Azure API Management: azure-rambi-apim-eab45rexk4hhs (23.719s)
  (✓) Done: Container App: movie-generator-svc (17.974s)
  (✓) Done: Container App: gui-svc (18.365s)
  (✓) Done: Container App: movie-poster-svc (18.47s)

Deploying services (azd deploy)

  (✓) Done: Deploying service gui
  - Endpoint: https://gui-svc.happyfield-037b39a8.francecentral.azurecontainerapps.io/

  (✓) Done: Deploying service movie_generator_svc
  - Endpoint: https://movie-generator-svc.happyfield-037b39a8.francecentral.azurecontainerapps.io/

  (✓) Done: Deploying service movie_poster_svc
  - Endpoint: https://movie-poster-svc.happyfield-037b39a8.francecentral.azurecontainerapps.io/

To open the application, click on the link below the Done: Deploying service gui

Troobleshooting configuration

az deployment group create --resource-group <target_rg> --template-file infra/user_portal_role.bicep

Configure Github actions

2. gh auth login
3. Execute infra/configure_authenticate_github_actions.sh .azure/dev/.env

Azure Infrastructue

The infra/main.bicep

• CognitiveServices
  • Open AI gpt-4
  • Open AI o1-mini
  • Open AI dall-e-3
• Container Registry
• Azure Container Apps Environment
• Container Apps
  • gui: the frontend
  • movie_poster_svc : service managing the poster description and the poster generation
  • movie_generator_svc: service managing the movie generation.
• API Management
• Application Insight
• Key Vault
• Storag Account (Blob Storage)

The infra/user_portal_role.bicep grants to the current authenticated user the right to browse the content of the keyvault and the blob storage using the Azure portal.

GitHub Credentials

edit the file infra/authenticate_with_Azure_App_Service_for_GitHub.sh with your context and run it to grant a contributor to the resource group and to generate the secrets used by the GitHub action pipeline to deploy the differents components.

• AZURE_CREDENTIALS
• AZURE_SUBSCRIPTION_ID

Front GUI

• Service: Azure Application Service using the python runtime on linux
• CI/CD Pipeline: .github/workflows/main_azure-rambi.yml

API Management

Note: The bicep files come from: https://github.com/microsoft/AzureOpenAI-with-APIM/

TMDB

The OpenAPI json file is available here: https://developer.themoviedb.org/openapi. You'll find the file for v3 and v4 version

Movie Generator Service

The generate_movie method has been moved to a new FastAPI service located in src/movie_generator_svc. This service is configured to run in Azure Container Apps.

Deployment

To deploy the new movie_generator_svc service, follow these steps:

1. Build the Docker image:

   docker build -t movie_generator_svc:latest src/movie_generator_svc

2. Push the Docker image to your container registry:

   docker tag movie_generator_svc:latest <your-container-registry>/movie_generator_svc:latest
   docker push <your-container-registry>/movie_generator_svc:latest

3. Update the Azure Container Apps configuration to use the new image.

Usage

To use the new movie_generator_svc service, make a POST request to the /generate endpoint with the following JSON payload:

{
    "movie1": {
        "title": "Movie 1 Title",
        "plot": "Movie 1 Plot",
        "poster_url": "Movie 1 Poster URL"
    },
    "movie2": {
        "title": "Movie 2 Title",
        "plot": "Movie 2 Plot",
        "poster_url": "Movie 2 Poster URL"
    },
    "genre": "Desired Genre"
}

The service will return a JSON response with the generated movie details.

Clean up

azd down --force --purge

Contributing

We welcome contributions! Please read our contributing guidelines for more details.

License

This project is licensed under the MIT License. See the LICENSE file for more information.

  - name: Build the image in the Azure Containe Registry
    id: acr
    uses: azure/acr-build@v1
    with:
      service_principal: ${{ secrets.AZURE_SERVICE_PRINCIPAL_ID }}
      service_principal_password: ${{ secrets.AZURE_SERVICE_PRINCIPAL_PASSWORD }}
      tenant: ${{ secrets.AZURE_TENANT_ID }}
      registry: ${{ env.ACR_NAME }}
      repository: azure-rambi
      image: movie_poster_svc
      tag: ${{ github.sha }}
      folder: src/movie_poster_svc
      git_access_token: ${{ secrets.GITHUB_TOKEN }}
      branch: main