OpenAI Assistant

Run

1. rename the .env.example file to .env and fill in the required values.
2. Select F5 to start the bot in debugger mode.
3. From your browswer, navigate to http://0.0.0.0/sales to start the bot.

[Optional] Get a Literal AI API key

Note

Literal AI is an all in one observability, evaluation and analytics platform for building LLM apps.

Go to Literal AI, create a project and go to Settings to get your API key.

Deploying on Azure

Review the Deploy a containerized Flask or FastAPI web app on Azure App Service

1. Clone the repository

   git clone https://github.com/gloveboxes/contoso-sales-azure-openai-assistant.git

2. Change directory to the project folder

   cd contoso-sales-azure-openai-assistant

3. Open the project in Visual Studio Code

   code .

4. You'll be prompted to open in a Dev Container. Select Reopen in Container

5. Open the terminal in VS Code

6. Export the following environment variables

   Set these deployment variables in your terminal

   export RESOURCE_GROUP_NAME=<value>
   export RESOURCE_GROUP_LOCATION=<value>      # Suggest a location near the model location, use az account list-locations --query "[].name" --output tsv
   export APP_SERVICE_PLAN_NAME=<value>
   export WEB_APP_NAME=<value>
   export CONTAINER_REGISTRY_NAME=<value>
   export CONTAINER_IMAGE_NAME=<value>
   export SUBSCRIPTION_ID=$(az account show --query id --output tsv)

   Set these Azure OpenAI environment variables in your terminal

   export AZURE_OPENAI_RESOURCE_NAME=<value>   # This should be the Azure AI Proxy endpoint for the event
   export AZURE_OPENAI_LOCATION=<value>        # eg swedencentral
   export AZURE_OPENAI_DEPLOYMENT=<value>      # eg gpt-4o
   export AZURE_OPENAI_API_VERSION=<value>     # eg 2024-05-01-preview

   Set the following application specific environment variables in your terminal

   export AZURE_OPENAI_ENDPOINT=<value>        # This should be the Azure AI Proxy endpoint for the event
   export LITERAL_API_KEY=<value>              # Optional, go to Literal AI https://cloud.getliteral.ai/, create a project and go to Settings to get your API key.
   export CHAINLIT_AUTH_SECRET=$(chainlit create-secret | cut -d'"' -f2)
   export AZURE_OPENAI_ASSISTANT_ID=<value>

Create the Azure resources

1. Create a resource group

   az group create --name $RESOURCE_GROUP_NAME --location $RESOURCE_GROUP_LOCATION

2. Create a container registry

   az acr create --resource-group $RESOURCE_GROUP_NAME --name $CONTAINER_REGISTRY_NAME --sku Basic

3. Create an App Service plan

   az appservice plan create --name $YOUR_APP_SERVICE_PLAN_NAME --resource-group $RESOURCE_GROUP_NAME --sku B1 --is-linux

Deploy the Azure OpenAI resource

1. Deploy an Azure OpenAI resource

   az cognitiveservices account create \
   --name $AZURE_OPENAI_RESOURCE_NAME \
   --resource-group $RESOURCE_GROUP_NAME \
   --kind OpenAI \
   --sku S0 \
   --location $AZURE_OPENAI_LOCATION

2. Deploy an Azure OpenAI Model

   az cognitiveservices account deployment create \
   --resource-group $RESOURCE_GROUP_NAME \
   --name $AZURE_OPENAI_RESOURCE_NAME \
   --deployment-name $AZURE_OPENAI_DEPLOYMENT \
   --model $AZURE_OPENAI_DEPLOYMENT \
   --model-version latest

Deploy the web app

1. Build the container image

   From VS Code terminal, run the following command to build the container image. This will build the container image in Azure and push to your Azure Container Registry.

   az acr build --resource-group $RESOURCE_GROUP_NAME --registry $CONTAINER_REGISTRY_NAME --image $CONTAINER_IMAGE_NAME .

2. Create the web app

   az webapp create --resource-group $RESOURCE_GROUP_NAME \
   --plan $YOUR_APP_SERVICE_PLAN_NAME \
   --name $WEB_APP_NAME \
   --assign-identity [system] \
   --role AcrPull \
   --scope /subscriptions/$SUBSCRIPTION_ID/resourceGroups/$RESOURCE_GROUP_NAME \
   --acr-use-identity --acr-identity [system] \
   --container-image-name $CONTAINER_IMAGE_NAME

3. Set environment variables

   az webapp config appsettings set --resource-group $RESOURCE_GROUP_NAME \
   --name contoso-sales \
   --settings \
       AZURE_OPENAI_ENDPOINT=$AZURE_OPENAI_ENDPOINT \
       AZURE_OPENAI_API_VERSION=$AZURE_OPENAI_API_VERSION \
       AZURE_OPENAI_DEPLOYMENT=$AZURE_OPENAI_DEPLOYMENT \
       LITERAL_API_KEY=$LITERAL_API_KEY \
       CHAINLIT_AUTH_SECRET=$CHAINLIT_AUTH_SECRET \
       AZURE_OPENAI_ASSISTANT_ID=$AZURE_OPENAI_ASSISTANT_ID