BigQuery Chat Agent

A simple web service using FastAPI and LangChain to interact with an LLM agent connected to Google BigQuery, with a Streamlit frontend.

Setup

1. Clone the repository:

   git clone <your-repo-url>
   cd <your-repo-directory>

2. Create a virtual environment and activate it:

   python3 -m venv venv
   source venv/bin/activate  # On Windows use `venv\Scripts\activate`

3. Install dependencies:

   pip install -r requirements.txt

4. Set up Google Cloud authentication: Ensure you have authenticated with Google Cloud. The easiest way is often using the gcloud CLI:

   gcloud auth application-default login

   Alternatively, set the GOOGLE_APPLICATION_CREDENTIALS environment variable to the path of your service account key file.

5. Configure environment variables: Create a .env file in the root directory and add your Google Cloud Project ID:

   GOOGLE_CLOUD_PROJECT=your-gcp-project-id
   # Optional: Specify BIGQUERY_DATASET if you want to limit the agent to a specific dataset
   # BIGQUERY_DATASET=your_dataset_name
   # Optional: Specify backend PORT if needed (default is 8080)
   # PORT=8000
   # Optional: Specify backend URL if running elsewhere
   # BACKEND_URL=http://your-backend-host:port/chat
   

Running the Service

1. Start the Backend (FastAPI): Open a terminal, navigate to the project directory, and run:

   python3 main.py
   # or using uvicorn for development with auto-reload
   # uvicorn main:app --reload --port 8080

   Keep this terminal running.

2. Start the Frontend (Streamlit): Open a second terminal, navigate to the project directory, and run:

   streamlit run frontend.py

   Streamlit will open the chat interface in your web browser.

Usage

Interact with the chat interface that opens in your browser. Ask questions about the data in your BigQuery project/dataset.

Alternatively, you can still send POST requests directly to the backend /chat endpoint:

curl -X POST "http://localhost:8080/chat" \
     -H "Content-Type: application/json" \
     -d '{
           "message": "Your question for the agent, e.g., How many tables are in the dataset?"
        }'

Replace "Your question for the agent..." with your actual query.