RagQueen

This is a RAG (Retrieval-Augmented Generation) to retrieve information from documentation written in markdown. It uses LangChain to process the inputs, HuggingFaces' all-MiniLM-L6-v2 as sentence transformer and OllamaModel llama3.1 for the inferance.

Firstly set the User Agent string to identify yourself when scraping the inputs

export USER_AGENT="ragqueen/1.0"

Set

export TOKENIZERS_PARALLELISM=false

To download the documentation and process it into embeddings

uv run process_input.py --config config.yaml

To ask a query

uv run query.py