Rag-System

Rag-System is a Python Retrieval-Augmented Generation framework with persistent Chroma storage, incremental ingestion, citations, and multi-provider LLM support.

Install

pip install Rag-System

Quick start

from Rag_System import RAG

rag = RAG()
rag.ingest("documents")

response = rag.ask("What is Retrieval-Augmented Generation?")
print(response.answer)
for citation in response.citations:
    print(citation.filename, citation.page_number)

ingest accepts a single file, a folder, a recursive glob, or a list of paths. Supported formats include PDF, TXT, Markdown, DOCX, CSV, Excel, JSON, HTML, and XML. Unchanged files are skipped; changed files are re-indexed and missing files are removed.

Public API

rag.ingest("paper.pdf")
rag.ingest("documents/**/*.pdf")
rag.ingest(["paper.pdf", "manual.md"])
rag.delete("paper.pdf")
rag.update("paper.pdf")
rag.stats()
rag.reset()
rag.compare("Compare these documents")
rag.summarize("Summarize the documents")

RAGResponse exposes answer, citations, sources, and retrieval statistics.

Configuration

Constructor options include provider, model, embedding_model, data_dir, and verbose. Environment variables can configure the selected provider, model, chunking, retrieval, and storage settings. For example:

RAG_PROVIDER=groq
GROQ_MODEL=llama-3.3-70b-versatile
GROQ_API_KEY=gsk_...

The default client is lazy: construction does not download models or ingest files. Call ingest before ask.

CLI

rag ingest
rag ask "Which colleges offer MCA?"
rag stats
rag documents