RAG Scraper

A Python-based RAG (Retrieval Augmented Generation) application that allows you to ask questions about web content or PDF documents using Ollama's local LLMs.

---

Youtube Video Guide

Watch Video Here

---

GUI Software Guideline

Download this repository as a zipped file and inside build folder, you'll find the compiled executable file. No need to install anything except Ollama and the required models.

---

Software Preview

---

Prerequisites

1. Install Ollama

Windows

1. Download Ollama from the official website: Ollama Downloads
2. Run the installer.
3. After installation, Ollama will run as a service in the background.

Linux

curl -fsSL https://ollama.ai/install.sh | sh

macOS

curl -fsSL https://ollama.ai/install.sh | sh

---

2. Install Required Ollama Models

After installing Ollama, open a terminal/command prompt and run:

# Install the base LLM model
ollama pull llama3.2

# Install the embeddings model
ollama pull all-minilm

---

Installation

1. Clone this repository:

   git clone https://github.com/TufayelLUS/RAG-Scraper-AI-GUI.git
   cd RAG-Scraper

2. Create a virtual environment:

   python -m venv venv

3. Activate the virtual environment:

   Windows:

   venv\Scripts\activate

   Linux/MacOS:

   source venv/bin/activate

4. Install the required packages:

   pip install -r requirements.txt

---

Usage

1. Make sure Ollama is running in the background.
2. Run the application:

   python "RAG Scraper.py"

Using the Interface:

a. Select Loader Type:

• Web Loader: For scraping web content.
• PDF Loader: For reading PDF files.

b. Select Models:

• Base Model: Choose the LLM for answering questions (e.g., llama3.2).
• Embeddings Model: Choose the model for text embeddings (e.g., all-minilm).

c. For Web Loader:

• Enter URLs (one per line).
• Enter the CSS class name of the content area to narrow down the context. Leave empty if not applicable.
  • Example: For a blog post, you might use "article-content" or "post-content".

d. For PDF Loader:

• Click "Browse" to select a PDF file.

e. Ask Questions:

• Enter your question in the question field.
• Click "ASK AI" to get answers.

---

Features

• Support for both web pages and PDF documents.
• Dynamic model selection from available Ollama models.
• Persistent settings saved in an INI file.
• Customizable content extraction for web pages.
• Error handling and user-friendly warnings.

---

Configuration

The application saves your preferences in settings.ini:

• Last used loader type
• Selected base model
• Selected embeddings model

---

Troubleshooting

• "Could not connect to Ollama API" error:

  • Ensure Ollama is running.
  • Check if the Ollama service is active.
  • Restart Ollama if necessary.

• Models not showing up:

  • Make sure you've pulled the models using ollama pull [model-name].
  • Check Ollama's status.
  • Restart the application.

• Web scraping not working:

  • Verify the URL is accessible.
  • Check if the CSS class name is correct.
  • Some websites may block scraping.

• PDF loading issues:

  • Ensure the PDF is not password-protected.
  • Check if the file is accessible.
  • Image based PDFs may not work.
  • Verify the PDF file is not corrupted.

---

Notes

• The application uses local LLMs through Ollama, ensuring privacy.
• Processing time depends on your hardware capabilities.
• Model performance varies based on the selected models.
• First-time model downloads may take time depending on your internet connection.

---

Requirements

• Python 3.8 or higher
• Ollama installed and running
• Sufficient disk space for models
• Internet connection for web scraping
• Adequate RAM (16GB minimum recommended)

---

Looking for a Software Engneer?

Reach out to me on Fiverr.