README.md

Elastic Chatbot RAG App

This is a sample app that combines Elasticsearch, Langchain and a number of different LLMs to create a chatbot experience with ELSER with your own private data.

Requires at least 8.11.0 of Elasticsearch.

Download the Project

Download the project from Github and extract the chatbot-rag-app folder.

curl https://codeload.github.com/elastic/elasticsearch-labs/tar.gz/main | \
tar -xz --strip=2 elasticsearch-labs-main/example-apps/chatbot-rag-app

Make your .env file

Copy env.example to .env and fill in values noted inside.

Installing and connecting to Elasticsearch

There are a number of ways to install Elasticsearch. Cloud is best for most use-cases. Visit the Install Elasticsearch for more information.

Once you decided your approach, edit your .env file accordingly.

Running your own Elastic Stack with Docker

If you'd like to start Elastic locally, you can use the provided docker-compose-elastic.yml file. This starts Elasticsearch, Kibana, and APM Server and only requires Docker installed.

Use docker compose to run Elastic stack in the background:

docker compose -f docker-compose-elastic.yml up --force-recreate -d

Then, you can view Kibana at http://localhost:5601/app/home#/

If asked for a username and password, use username: elastic and password: elastic.

Clean up when finished, like this:

docker compose -f docker-compose-elastic.yml down

Connecting to LLM

We support several LLM providers, but only one is used at runtime, and selected by the LLM_TYPE entry in your .env file. Edit that file to choose an LLM, and configure its templated connection settings:

• azure: Azure OpenAI Service
• bedrock: Amazon Bedrock
• openai: OpenAI Platform and services compatible with its API.
• vertex: Google Vertex AI
• mistral: Mistral AI
• cohere: Cohere

Running the App

There are two ways to run the app: via Docker or locally. Docker is advised for ease while locally is advised if you are making changes to the application.

Run with docker

Docker compose is the easiest way, as you get one-step to:

• build the frontend
• ingest data into elasticsearch
• run the app, which listens on http://localhost:4000

Double-check you have a .env file with all your variables set first!

docker compose up --build --force-recreate

Note: First time creating the index can fail on timeout. Wait a few minutes and retry.

Clean up when finished, like this:

docker compose down

Run locally

If you want to run this example with Python and Node.js, you need to do a few things listed in the Dockerfile. The below uses the same production mode as used in Docker to avoid problems in debug mode.

Double-check you have a .env file with all your variables set first!

Build the frontend

The web assets are in the frontend directory, and built with yarn.

# Install and use a recent node, if you don't have one.
nvm install --lts
nvm use --lts
# Build the frontend web assets
(cd frontend; yarn install; REACT_APP_API_HOST=/api yarn build)

Configure your python environment

Before we can run the app, we need a working Python environment with the correct packages installed:

python3 -m venv .venv
source .venv/bin/activate
# Install dotenv which is a portable way to load environment variables.
pip install "python-dotenv[cli]"
pip install -r requirements.txt

Run the ingest command

First, ingest the data into elasticsearch:

FLASK_APP=api/app.py dotenv run -- flask create-index

Note: First time creating the index can fail on timeout. Wait a few minutes and retry.

Run the app

Now, run the app, which listens on http://localhost:4000

dotenv run -- python api/app.py

Advanced

Updating package versions

To update package versions, recreate requirements.txt and reinstall like this. Once checked in, any commands above will use updates.

rm -rf .venv
python3 -m venv .venv
source .venv/bin/activate
# Install dev requirements for pip-compile
pip install pip-tools
# Recreate requirements.txt
pip-compile
# Install main dependencies
pip install -r requirements.txt

Elasticsearch index and chat_history index

By default, the app will use the workplace-app-docs index and the chat history index will be workplace-app-docs-chat-history. If you want to change these, edit ES_INDEX and ES_INDEX_CHAT_HISTORY entries in your .env file.

Indexing your own data

The ingesting logic is stored in data/index_data.py. This is a simple script that uses Langchain to index data into Elasticsearch, using RecursiveCharacterTextSplitter to split the large JSON documents into passages. Modify this script to index your own data.

See Langchain documentation for more ways to load documents.

---