amazon-bedrock-genai-chatbot

This repository contains code samples for a chatbot using Amazon Bedrock, LangChain & MongpDB Atlas Vector Search.

Gain Model Access from Amazon Bedrock Console

Visit the Amazon Bedrock documentation for instructions on gaining model access. For Claude access, use the us-east-1 or us-west-2 region.

NOTE: This codebase uses the region us-west-2. Please update the region in the .env file if using another region.

Create MongoDB instance

Create a MongoDB instance by following the instructions in the documentation. Note down the host, username, and password.

⚠️ Pay attention to the network setup. If you are using SageMaker studio to go along with this tutorial, you will need to expose the MongoDB instance to the internet.:

Populate instance with embeddings

1. Create a .env file in the root directory and add the following environment variables:

MDB_HOST="REPLACE_WITH_HOST_NAME.mongodb.net"
MDB_USERNAME="REPLACE_WITH_YOUR_USERNAME"
MDB_PASSWORD="REPLACE_WITH_YOUR_PASSWORD"

2. Follow the notebook shopping-bot.ipynb to download product data and embed and store in MongoDB Vector

3. Update the .env file in the root directory and add the collection and database names as well:

MDB_HOST="REPLACE_WITH_HOST_NAME.mongodb.net"
MDB_USERNAME="REPLACE_WITH_YOUR_USERNAME"
MDB_PASSWORD="REPLACE_WITH_YOUR_PASSWORD"
MDB_COLLECTION="REPLACE_WITH_YOUR_COLLECTION"
MDB_DATABASE="REPLACE_WITH_YOUR_DATABASE"

Run streamlit application

streamlit run chatbot_rag.py

Start chatting

While chatting, check your terminal window to see how the chain is running.

NOTE: Set verbose=False for chain ConversationalRetrievalChain in the file langchain.py if you dont want to see detail output.