This folder contains notebooks that demonstrate how to integrate popular OpenAI services with Elasticsearch.

The following notebooks are available:

• OpenAI embeddings and retrieval augmented generation (RAG)

In the openai-KNN-RAG.ipynb notebook you'll learn how to:

• Index the OpenAI Wikipedia vector dataset into Elasticsearch.
• Embed a question with the OpenAI embeddings endpoint.
• Perform semantic search on the Elasticsearch index using the encoded question.
• Send the top search results to the OpenAI Chat Completions API endpoint for retrieval augmented generation (RAG).
• Use OpenAI's model to generate a response for a given conversation.
• Set a system message which defines the assistant's role and how user messages are processed.