Looker Agent with ADK and MCP Toolbox

This repository contains the necessary steps to implement and deploy a Looker agent using the ADK (Agent Development Kit) and MCP (Model Context Protocol) Toolbox.

---

MCP Toolbox for Looker Implementation

First, you need to set up the MCP Toolbox server.

1. Navigate to the mcp-toolbox directory and replace the contents in the .env file with your variables:

   BASE_URL=your_looker_instance_url
   LOOKER_CLIENT_ID=your_client_id
   LOOKER_CLIENT_SECRET=your_client_secret

2. Load your variables from the .env file into your terminal session as environment variables.:

   export $(grep -v '^#' .env | xargs)

3. Start the server:

   ./toolbox --tools-file "tools.yaml"

4. To test it locally with an interactive UI:

   ./toolbox --ui

---

MCP Toolbox with ADK

Next, create the agent and connect it to your tools.

1. Navigate to the my-agents directory, set up a virtual environment, and activate it:

   cd my-agents
   python -m venv .venv
   source .venv/bin/activate

2. Replace the contents in the .env file with your variables:

   GOOGLE_GENAI_USE_VERTEXAI=1
   GOOGLE_CLOUD_PROJECT=your_google_cloud_project_id
   GOOGLE_CLOUD_LOCATION=us-central1
   LOOKER_CLIENT_ID=your_looker_client_id
   LOOKER_CLIENT_SECRET=your_looker_client_secret
   LOOKER_BASE_URL=your_looker_instance_url

3. Install the ADK and MCP Toolbox packages:

   pip install google-adk toolbox-core

4. Test your agent locally. This command will return a URL for testing.

   adk web

   Note: The MCP server must be running for the agent to access the tools.

---

Deploying the Agent to Cloud Run

Finally, deploy the MCP Toolbox and the agent to Cloud Run.

Deploy the MCP Toolbox

1. Navigate to your mcp-toolbox directory, define your PROJECT_ID, and enable the necessary services:

   export PROJECT_ID="YOUR_GOOGLE_CLOUD_PROJECT_ID"
   gcloud services enable run.googleapis.com \
   cloudbuild.googleapis.com \
   artifactregistry.googleapis.com \
   secretmanager.googleapis.com

2. Create a service account:

   gcloud iam service-accounts create toolbox-identity
   gcloud projects add-iam-policy-binding $PROJECT_ID \
   --member="serviceAccount:toolbox-identity@$PROJECT_ID.iam.gserviceaccount.com" \
   --role="roles/secretmanager.secretAccessor"
   gcloud projects add-iam-policy-binding $PROJECT_ID \
   --member="serviceAccount:toolbox-identity@$PROJECT_ID.iam.gserviceaccount.com" \
   --role="roles/cloudsql.client"

3. Upload the tools.yaml file as a secret and set the image variable:

   gcloud secrets create tools --data-file tools.yaml \
   --replication-policy="user-managed" --locations="us-central1"
   export IMAGE=us-central1-docker.pkg.dev/database-toolbox/toolbox/toolbox:latest

4. Deploy to Cloud Run:

   gcloud run deploy toolbox \
   --image $IMAGE \
   --service-account toolbox-identity \
   --region us-central1 \
   --set-secrets "/app/tools.yaml=tools:latest" \
   --args="--tools_file=/app/tools.yaml","--address=0.0.0.0","--port=8080" \
   --allow-unauthenticated

Deploy the Agent Application

1. In your my-agents/looker-agent/agent.py file, update the ToolboxSyncClient URL to your new Cloud Run service URL.

2. Navigate to the my-agents directory and define the following variables:

   export GOOGLE_CLOUD_PROJECT="YOUR PROJECT ID"
   export GOOGLE_CLOUD_LOCATION="YOUR REGION"
   export AGENT_PATH="looker-agent/"
   export SERVICE_NAME="looker-service"
   export APP_NAME="looker-app"
   export GOOGLE_GENAI_USE_VERTEXAI=True

3. Deploy the agent application to Cloud Run:

   adk deploy cloud_run \
   --project=$GOOGLE_CLOUD_PROJECT \
   --region=$GOOGLE_CLOUD_LOCATION \
   --service_name=$SERVICE_NAME \
   --app_name=$APP_NAME \
   --with_ui \
   $AGENT_PATH

Mission Accomplished! 🚀 You have successfully created and deployed your agent. You can now use it anytime from the Cloud Run URL provided by the final deployment command.