Description

Lightweight RAG Framework: Simple and Scalable Framework with Efficient Embeddings. Leverage: FAISS, ChromaDB, and Ollama

• LLMs
• Vector Database or Similarity Search engine
• Embeddings
• Chunking techniques
• Data Collection
  Each of these objects are managed by the framework and provides the same IN/OUT to enable interoperability and facilitates changes (as this area is moving super fast !).

This project can totally runs locally (meaning on a single laptop without GPUs) and leverages:

• Ollama (https://ollama.com/) for running locally LLMs
• AWS Bedrock LLMs provided through AWS bedrock
• Sentence-transformers (https://pypi.org/project/sentence-transformers/) for the embeddings management
• potentially llamaParse from llamaindex (https://docs.llamaindex.ai/) can be used.
• langchain (https://www.langchain.com/) for chunking (semantic or character)
• Meta FAISS for similarity search (also enable storing and loading indexes)
• ChromaDB for storing and searching into the vector store
• PyMuPDF by default to read and convert PDF content

Framework description

Several objects are provided to manage the main RAG features and characteristics:

• rag: is the main interface for managing all needed request.
• IDocument: manages the document reading and loading (pdf or direct content)
• IChunks: manages the chunks list
• IEmbeddings: Manages the vector and data embeddings
• INearest: Manages the k nearest neighbors retreived by the similarity search engine
• IPrompt: Manages Prompt templating and simple prompt

LLMs supported:

• Ollama
• AWS Claude
• Hugging Face

Document reading methods are supported:

• PDF via PyMuPDF
• PDF via Llamaparse
• HTML stream

Chunking methods are supported:

• Character chunkink (langchain)
• Semantic chunking (langchain)

Vectors stores are currently supported:

• FAISS: search + load and store indexes
• ChromaDB

Embeddings methods are supported:

• via HF Sentence Transformer (the model can be changed)
• via Ollama Embeddings Models (the model can be changed)

Installation

Python Framework installation

1. Download and Install Python,
2. install ragfmk by using pip

pip install [--force-reinstall] wheel file (see the /dist folder)

Environment variables

Some environment variables may need to be set:

• If you need to use llamaParse, the llamaindex token (generated on the web site: https://cloud.llamaindex.ai/login) must be filled out to LLAMAINDEX_API_KEY
• If you need to use hugging face, the hugging face env. token must be filled out into HUGGINGFACE_API_KEY
• For AWS please set the following variables:
  • AWS_ACCESS_KEY_ID=...
  • AWS_SECRET_ACCESS_KEY=...

Installation/Preparation for Ollama

1. Install ollama (https://ollama.com/)
2. Run ollama in the command line and pull at least one model. tinydolphin for example is a good choice as it is a very small model and can then run on a simple laptop without a big latency.

Example of use

See the tests folder.