MongoDB VertexAI Groceries Agent

This project provides an AI-powered agent for grocery shopping, leveraging MongoDB for data storage and Google Vertex AI for semantic search and embeddings.

Check out the Medium tutorial for more information.

Features

• Semantic product search using MongoDB Atlas Vector Search and Vertex AI embeddings
• Add products to user carts in MongoDB

Prerequisites

• Python 3.10+
• Access to Google Cloud Gemini API
• Access to a MongoDB Atlas cluster (instructions below)
• Required Python packages (instructions below)
• Google ADK Python installed (instructions below)

Loading the Dataset and Generating Embeddings

1. Create a free MongoDB Atlas cluster

• Go to MongoDB Atlas and sign up for a free account.
• Click "Build a Database" and choose the free tier (Shared, M0).
• Select your preferred cloud provider and region, then click "Create".
• Create a database user with a username and password.
• Add your IP address to the IP Access List (or allow access from anywhere for development).
• Once the cluster is created, click "Connect" and choose "Connect your application" to get your connection string. Use this string for the CONNECTION_STRING environment variable in the next steps.

2. Clone the repository

git clone https://github.com/mongodb-developer/MongoDB-ADK-Agents.git
cd MongoDB-VertexAI-ADK

3. Load the Dataset into MongoDB Atlas

Import the provided dataset into your MongoDB database using the following command (replace placeholders as needed):

mongoimport --uri "$CONNECTION_STRING" --db "$DATABASE_NAME" --collection "$COLLECTION_NAME" --type csv --headerline --file mongodb-groceries-agent/dataset.csv

4. Generate Embeddings for the Inventory

After loading the data, you need to generate vector embeddings for each product. Run the following script:

python mongodb-groceries-agent/create-embeddings.py

This will process all products in the collection and add/update the embedding field required for semantic search.

5. Build a Vector Search Index for the Inventory

Open the Search and Vector Search tab in the left sidebar in Atlas and create a vector search index on the inventory collection with the following definition:

{
  "fields": [
    {
      "numDimensions": 3072,
      "path": "gemini_embedding",
      "similarity": "cosine",
      "type": "vector"
    }
  ]
}

Setup

1. Install the Python dependencies

pip install -r requirements.txt

2. Install the ADK CLI

Follow the official ADK installation instructions or run:

pip install google-adk

3. Set environment variables

Set the following environment variables in a .env file:

GOOGLE_GENAI_USE_VERTEXAI=FALSE

# Follow the guide: https://www.mongodb.com/docs/guides/atlas/connection-string/
CONNECTION_STRING="Your MongoDB connection string"
# Follow the guide: https://cloud.google.com/api-keys/docs/create-manage-api-keys
GOOGLE_API_KEY="Your Google Cloud API key"

5. Run the agent using ADK

Navigate to the mongodb-groceries-agent directory and run:

adk web

6. Open the web server running at http://127.0.0.1:8000 and start using the application!

Usage

• The agent will start and be ready to handle product search and cart operations.
• You can extend the agent with new tools or integrate it into a larger application.

Project Structure

• mongodb-groceries-agent/agent.py: Main agent logic
• mongodb-groceries-agent/create-embeddings.py: Utility for creating embeddings
• mongodb-groceries-agent/dataset.csv: Example dataset

Notes

• Ensure your Google Cloud and MongoDB credentials are valid and have the necessary permissions.
• For local development, you may want to use a virtual environment.
• The ADK CLI is required for running and managing agents.

License

See LICENSE for details.