PDF Assistant Demo

A Streamlit-based AI assistant that:

• Reads and processes PDF documents.
• Stores document chunks in a Chroma vector database.
• Uses NoSQL-style metadata for document context.
• Embeds and retrieves relevant content to answer user questions.

This project demonstrates:

• Vector embeddings with nomic-embed-text
• Chroma vector database for semantic search
• NoSQL-style metadata handling
• PDF ingestion, text splitting, and contextual retrieval

---

Features

• Upload one or more PDFs.
• Get AI-generated answers that include:
  • Red Flags / Risks
  • Positive Aspects
  • Recommendations
• View retrieved document chunks for transparency.
• Persistent local vector DB so embeddings aren't recalculated unnecessarily.

---

Requirements

• Python 3.10+
• Ollama installed and running locally.
• The following Ollama models available:
  • llama3.2
  • nomic-embed-text

---

Installation

1. Clone the repository

   git clone https://github.com/Bredicus/pdf-assistant-demo.git
   cd pdf-assistant-demo

2. Create and activate a virtual environment

   python -m venv venv

3. Install dependencies

   pip install -r requirements.txt

4. Ensure Ollama is running

   ollama serve

5. Pull required models

   ollama pull llama3.2
   ollama pull nomic-embed-text

---

Running the App

streamlit run app.py

The app will open in your browser at http://localhost:8501.

---

Project Structure

pdf-assistant-demo/
│
├── app.py              # Main Streamlit application
├── requirements.txt    # Python dependencies
├── README.md           # Project documentation
└── chroma_db/          # Persistent Chroma vector database (auto-created)

---

Example Workflow

1. Upload one or more PDF files.
2. The app extracts text, splits it into chunks, and stores embeddings in Chroma.
3. Ask a question about the content.
4. The assistant retrieves relevant chunks and provides a structured answer:
   • Risks
   • Positives
   • Recommendations