PurrfectKit’s documentation

PurrfectKit (/purr-fekt-kit/) is a Python library for effortless Retrieval-Augmented Generation (RAG) workflows.

Note

This project is under active development.

PurrfectKit simplifies Retrieval-Augmented Generation (RAG), a technique combining information retrieval and text generation to answer queries using external documents.

The workflow involves five steps, each named after a Thai cat breed for a memorable experience:

1. Suphalak: Extract text from files (e.g., PDFs).

2. Malet: Split text into manageable chunks.

3. WichienMaat: Convert chunks into numerical embeddings for search.

4. KhaoManee: Search for chunks relevant to a query.

5. Kornja: Generate answers using retrieved chunks (under development).

Contents:

• Installation
  • Prerequisites
• Quickstart
• Usage
  • Suphalak
  • Malet
  • WichienMaat
  • KhaoManee
  • Kornja
• PurrfectKit
  • purrfectmeow package
• Contributing