Ready Tensor RAG Assistant

A retrieval-augmented generation (RAG) chatbot that helps users explore Ready Tensor publications through natural language queries. Built with LangChain, FAISS/Chroma, and an LLM backend.

---

🚀 Features

• Custom Knowledge Base: 20–50 Ready Tensor publications ingested.
• Vector Search: Semantic embeddings for accurate context retrieval.
• Conversational Interface: Query the knowledge base with natural questions.
• Extendable: Add memory, reasoning chains (ReAct, CoT), or logging.
• Minimal UX: CLI / notebook interface for fast prototyping.

---

🏗️ Architecture

1. Document Ingestion

   • Load publications dataset (PDFs, text, or scraped content).
   • Embed with all-MiniLM-L6-v2.

2. Vector Store

   • Store embeddings in FAISS or Chroma.

3. Retriever + LLM

   • LangChain pipeline connects retriever → LLM → response.
   • Example backend: OpenAI GPT, Anthropic, or Hugging Face LLM.

4. Interface

   • Notebook or CLI for testing queries.

---

⚙️ Installation

# clone repository
git clone https://github.com/Paul-masingah/readytensor-rag-assistant.git
cd readytensor-rag-assistant

# create venv
python -m venv venv
source venv/bin/activate  # (Linux/Mac)
venv\Scripts\activate     # (Windows)

# install dependencies
pip install -r requirements.txt

---

▶️ Usage

# run notebook
jupyter notebook notebooks/demo.ipynb

# or run CLI chatbot
python src/chatbot.py

Example queries:

• "What is this publication about?"
• "Which models or tools were used?"
• "Any limitations or assumptions?"

---

📂 Project Structure

readytensor-rag-assistant/
│
├── data/                # sample Ready Tensor publications
├── notebooks/           # demo notebooks
│   └── demo.ipynb
├── src/
│   ├── ingest.py        # document ingestion + embedding
│   ├── chatbot.py       # main conversational bot
│   └── utils.py
├── requirements.txt
└── README.md

---

📊 Datasets

• Ready Tensor sample publications (Google Drive)
• MiniLM sentence transformer (Hugging Face)
• Optional: Kaggle datasets
• Optional: TensorFlow Hub models

---

🔗 References & Code Links

• Main repo: [https://github.com/Paul-masingah/smart-rag-assistant/releases
• LangChain
• Chroma
• FAISS

---

📜 License

MIT License — free for personal and commercials use