LangChain AI Agent with YugabyteDB

Local YugabyteDB deployment

1. Create a directory to store data locally.

   rm -rf ~/yb_docker_data
   mkdir ~/yb_docker_data
   

2. Create a Docker network.

   docker network create custom-network
   

3. Deploy a 3-node cluster to this network.

   docker run -d --name yugabytedb-node1 --net custom-network \
       -p 15433:15433 -p 7001:7000 -p 9001:9000 -p 5433:5433 \
       -v ~/yb_docker_data/node1:/home/yugabyte/yb_data --restart unless-stopped \
       yugabytedb/yugabyte:2.20.0.2-b1-aarch64 \
       bin/yugabyted start \
       --cloud_location=aws.us-east-1.us-east-1a \
       --fault_tolerance=region \
       --base_dir=/home/yugabyte/yb_data --background=false
   
   docker run -d --name yugabytedb-node2 --net custom-network \
       -p 15434:15433 -p 7002:7000 -p 9002:9000 -p 5434:5433 \
       -v ~/yb_docker_data/node2:/home/yugabyte/yb_data --restart unless-stopped \
       yugabytedb/yugabyte:2.20.0.2-b1-aarch64 \
       bin/yugabyted start --join=yugabytedb-node1 \
       --cloud_location=aws.us-central-1.us-central-1a \
       --fault_tolerance=region \
       --base_dir=/home/yugabyte/yb_data --background=false
   
   docker run -d --name yugabytedb-node3 --net custom-network \
       -p 15435:15433 -p 7003:7000 -p 9003:9000 -p 5435:5433 \
       -v ~/yb_docker_data/node3:/home/yugabyte/yb_data --restart unless-stopped \
       yugabytedb/yugabyte:2.20.0.2-b1-aarch64 \
       bin/yugabyted start --join=yugabytedb-node1 \
       --cloud_location=aws.us-west-1.us-west-1a \
       --fault_tolerance=region \
       --base_dir=/home/yugabyte/yb_data --background=false
   

Adding data

1. Copy SQL files to Docker container.

   docker cp {project-root-directory}/sql/. yugabytedb-node1:/home/sql/
   

2. Load the schema and seed data with schema.sql and data.sql.

   docker exec -it yugabytedb-node1 bin/ysqlsh -h yugabytedb-node1 -f /home/sql/schema.sql
   docker exec -it yugabytedb-node1 bin/ysqlsh -h yugabytedb-node1 -f /home/sql/data.sql
   
   

3. Load the Airbnb dataset to the cluster (note, it can take a minute to load the data):

   docker exec -it yugabytedb-node1 bin/ysqlsh -h yugabytedb-node1 \
       -c "\copy airbnb_listings from /home/sql/airbnb_listings_with_embeddings.csv with DELIMITER '^' CSV"
   

Running Backend Services

The backend consists of 2 Flask servers, one (app.py) for accepting chat messages from the UI to interact with an A.I. agent, and another (api.py) for communication betweeen the agent and the database.

1. Install application dependencies in a virtual environment.

cd {project-root-directory}/python-server/
python3 -m venv agent-application-env
source agent-application-env/bin/activate
pip install -r requirements.txt
# NOTE: Users with M1 Mac machines should use requirements-m1.txt instead:
# pip install -r requirements-m1.txt

2. Create an OpenAI API Key and store it's value in a secure location. This will be used to connect the application to the LLM to generate SQL queries and an appropriate response from the database.

3. (Optional) Create a Tavily API Key if you'd like to include search engine functionality in the A.I. agent.

4. Configure the application environment variables in {project_directory/.env.example} and include them in the same directory in a file named .env.

5. Run the application services in seperate terminal windows.

python app.py
* Running on http://127.0.0.1:3000

python api.py
* Running on http://127.0.0.1:8000

Running UI

Install project dependencies and run the UI:

cd {project-root-directory}/langchain-agent-ui
npm install
npm run dev

By default, the UI will run on http://localhost:5173/ and proxies HTTP requests to http://localhost:3000 in vite.config.js.