scripts for downloading input, converting docs, generating embeddings and chroma insertion

[Gregory-Pereira]

cc @cooktheryan @MichaelClifford @nerdalert

data is public domain so can be included

General progress recap:

• get documents
• convert documents to markdown
• generate QNA pairs (MANUAL)
• run ilab train on this
• embed documents (converted markdown not QNA)
• Insert into vector DB (agnostic, anything but milvus)
• script that connects the tuned LLM to the vector DB and does the RAG retreival
• query the LLM via the script