RAG AI Assistant Flask Application

This Flask application provides an AI assistant that can extract and query information from either Wikipedia articles or PDF documents using a Retrieval-Augmented Generation (RAG) pipeline.

Documentation :

https://medium.com/@abhayemani8/building-a-rag-based-document-q-a-application-with-flask-and-langchain-5b8db9c04cb6

Features

• Upload PDF documents or provide Wikipedia URLs for knowledge extraction
• Ask questions about the content using natural language
• Dark-themed user interface with responsive design
• Session management to maintain context between questions

Prerequisites

• Python 3.8 or higher
• Ollama running locally with the llama3.2 model installed
  • Install from ollama.ai
  • Run ollama pull llama3.2 to download the model
• Required Python packages (see requirements.txt)

Installation

1. Clone this repository:

git clone https://github.com/MONARCH1108/AI_Rag_Assist
cd AI_Rag_Assist

2. Create a virtual environment and activate it:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install the required packages:

pip install -r requirements.txt

Running the Application

1. Make sure Ollama is running in the background with the llama3.2 model installed.

2. Start the Flask application:

python Flask_app.py

3. Open your web browser and navigate to http://127.0.0.1:5000/

4. App with out Flask Interface

python app.py

Usage

1. Wikipedia Mode:

   • Enter a Wikipedia URL (e.g., https://en.wikipedia.org/wiki/Artificial_intelligence)
   • Click "Load Article" and wait for processing to complete

2. PDF Mode:

   • Click the "PDF" tab
   • Select a PDF file from your computer
   • Click "Upload PDF" and wait for processing to complete

3. Asking Questions:

   • Once content is loaded, the chat input will be enabled
   • Type your question and press Enter or click the send button
   • The AI will respond based on the content of the loaded document

4. Starting Over:

   • Click "Clear Session" to reset the application and load new content

Technical Details

This application uses:

• LangChain for document processing and the RAG pipeline
• HuggingFace embeddings (sentence-transformers/paraphrase-MiniLM-L6-v2)
• Ollama's local LLM (llama3.2) for answering questions
• ChromaDB for vector storage
• Flask for the web server

Notes

• Processing large documents may take time, especially on systems with limited resources
• The quality of answers depends on both the content provided and the capabilities of the llama3.2 model
• For best results, ask specific questions related to the content of the loaded document