A multi-agent system implemented with CrewAI created to support coding activities

Agent generated with googleCloudPlatform/agent-starter-pack

This project is organized as follows:

study-helper-agent/
├── app/                 # Core application code
│   ├── agent.py         # Main agent logic
│   ├── agent_engine_app.py # Agent Engine application logic
│   └── utils/           # Utility functions and helpers
├── deployment/          # Infrastructure and deployment scripts
├── notebooks/           # Jupyter notebooks for prototyping and evaluation
├── tests/               # Unit, integration, and load tests
├── Makefile             # Makefile for common commands
└── pyproject.toml       # Project dependencies and configuration

Before you begin, ensure you have:

• uv: Python package manager - Install
• Google Cloud SDK: For GCP services - Install
• Terraform: For infrastructure deployment - Install
• make: Build automation tool - Install (pre-installed on most Unix-based systems)

Install required packages using uv:

make install

If not done during the initialization, set your default Google Cloud project and Location:

export PROJECT_ID="qwiklabs-gcp-02-662fdf28d3fa"
export LOCATION="us-central1"
gcloud config set project $PROJECT_ID
gcloud auth application-default login
gcloud auth application-default set-quota-project $PROJECT_ID

For full command options and usage, refer to the Makefile.

1. Prototype: Build your Generative AI Agent using the intro notebooks in notebooks/ for guidance. Use Vertex AI Evaluation to assess performance.
2. Integrate: Import your chain into the app by editing app/agent.py.
3. Test: Explore your chain's functionality using the Streamlit playground with make playground. The playground offers features like chat history, user feedback, and various input types, and automatically reloads your agent on code changes.
4. Deploy: Configure and trigger the CI/CD pipelines, editing tests if needed. See the deployment section for details.
5. Monitor: Track performance and gather insights using Cloud Logging, Tracing, and the Looker Studio dashboard to iterate on your application.

You can test deployment towards a Dev Environment using the following command:

gcloud config set project <your-dev-project-id>
make backend

The repository includes a Terraform configuration for the setup of the Dev Google Cloud project. See deployment/README.md for instructions.

The repository includes a Terraform configuration for the setup of a production Google Cloud project. Refer to deployment/README.md for detailed instructions on how to deploy the infrastructure and application.

You can use this Looker Studio dashboard template for visualizing events being logged in BigQuery. See the "Setup Instructions" tab to getting started.

The application uses OpenTelemetry for comprehensive observability with all events being sent to Google Cloud Trace and Logging for monitoring and to BigQuery for long term storage.