Large Language Models - Demo Project

This repository supports my sessions on large language models and their integration into Domino apps. You can find the slides for the sessions below:

• Experimenting with Integrating Large Language Models into Domino Apps" at Engage 2024.
• Integrating Large Language Models into Domino Applications: A Primer at OpenNTF Webinars
  • Youtube

Prerequisites

• Project Metadata Database: Some of the code requires a project metadata database, which is publicly available as an OpenNTF project.
• Langchain4j-Domino Integration: The project utilizes langchain4j-domino, which will be made publicly available in the near future. Experimental version is here.
• LLM Model Container: The ingestion code leverages the Ollama LLM management tool, preferably installed locally.
• Vector Database: The ingestion target is a Chroma vector database, preferably installed locally.
• OpenAI API: The AiService demos require access to the OpenAI API, along with a valid API key.
  • Copy the .env.example file to your user directory, rename it to .env, and add your OpenAI API key.
• Other XPages plugins:
  • OpenNTF Domino API
  • Domino JNX - XPages
  • Jakarta EE Project

Demos

Module: langchain4j-demos-cli

• Prompts and Fewshots Demos: Demonstrations of the basic functionality of the AiService.
• ingestProjectMetadata: Demonstrates how project metadata can be ingested into a vector database.

Module: nsf/odp.langchain4j-xsp-demo

This XPages application highlights the capabilities of langchain4j through three distinct demonstrations:

1.	**Few Shots Example**: A simple use case of the AiService.
1.	**Semantic Search**: Demonstrates two variations of semantic search using OLLAMA and OpenAI.
1.	**Helpful Assistant**: An advanced example featuring a chat service enhanced with Retrieval-Augmented Generation (RAG) and tooling addons.

To configure the demos:

1. Create an NSF Database: Generate an NSF database from the provided ODP file.

2. Deploy to Domino Server: Place the NSF database on a Domino server and sign it.

3. Prepare Metadata Dataset: Download the OpenNTF Projects Dataset, deploy it to the same server, sign it, and create a full-text index.

4. Set Up Server Properties: Open the xsp.properties file in your server’s data directory and define the following variables:

PMT_METADATA_DB=demos/pmt_metadata.nsf
OLLAMA_URI=http://<ollama-server>:11434
OLLAMA_MODEL=mxbai-embed-large:latest
CHROMA_URI=http://<chroma-server>:8000
langchain4j.OPENAI_API_KEY=<Openai-api-key>

Compatibility

Java

This project has been tested on Java 17 or higher.

Apple Silicon

If you’re running Apple Silicon computers, use a x86_64 JVM (e.g. IBM Semeru OpenSDK 17) to run the demo databases.