Rag/rag transition 3

[joelpaulkoch]

See A RAG Library for Elixir

We want to turn chatbot_ex into a RAG system that answers question about ecto. Out of the box, chatbot_ex knows nothing about ecto.

In Step one, we ran the generator.
In Step two, we removed the LLM serving as we will use Ollama to generate responses.

Ingesting Ecto

To enable chatbot_ex to answer questions about ecto, we ingest the documentation and source code of ecto into our RAG system.

First, we change the embeddings model to jina-embeddings-v2-base-code.

Then, we adapt the ingestion pipeline to download code and documentation of ecto from hex.pm, chunk the downloaded files, calculate embeddings of the chunks, and store everything in postgres.