This demo application lets users ask questions about the Apache Groovy documentation. It is a chatbot powered by the Gemini Large Language Model created by Google. The application is developed using the Groovy programming language, and the Micronaut framework. The frontend is built using Vue.js.
To interact with this chatbot, you can use the this link: Apache Groovy LLM Doc Chatbot
This chatbot application implements the Retrieval Augmented Generation (RAG) approach:
- a document ingestion phase is loading, parsing, chunking the documentation, and is creating vector embeddings to store them in an in-memory vector database
- at query time, the application calculates a vector embedding, and compares it to the other vector embeddings in the vector database to find the most similar excerpts
This projects uses the LangChain4J LLM orchestration library to interact with the PaLM LLM and the in-memory vector database (also provided by LangChain4j).