Skip to content

SETUP_GUIDE.md

Latest commit

 

History

History
126 lines (92 loc) · 2.9 KB

SETUP_GUIDE.md

File metadata and controls

126 lines (92 loc) · 2.9 KB

Setup Guide for PDF Analyzer

Prerequisites Setup

1. Install Ollama

  1. Download Ollama:

    • Go to https://ollama.ai
    • Download the Windows installer
    • Run the installer and follow the instructions

  2. Verify Installation:

    ollama --version

  3. Start Ollama Service:

    ollama serve

  4. Pull a Model (in a new terminal):

    ollama pull llama3.2

    Other good models for Arabic content:

    ollama pull qwen2.5
ollama pull mistral

2. Install Python Dependencies

cd "d:\INTERNSHIP\pdf_analyzer"
pip install -r requirements.txt

3. Test the Setup

  1. Test Ollama:

    ollama list

    You should see the models you pulled.

  2. Test with a simple question:

    ollama run llama3.2 "Hello, how are you?"

Running the Application

Option 1: Web Interface (Recommended)

  1. Start the Streamlit app:

    streamlit run pdf_analyzer.py

  2. Open browser to: http://localhost:8501

  3. Upload your PDF and start asking questions!

Option 2: Command Line

python cli.py "1-النظام الأساسي للحكم.pdf"

Option 3: Batch File (Windows)

Double-click run_app.bat to start the web interface.

Tips for Best Results

For Arabic Documents:

  • Use models like qwen2.5 which have better Arabic support
  • Ask questions in Arabic for better understanding
  • The system will maintain the language of the source document

Chunk Size Optimization:

  • Large documents: Use chunk size 800-1000
  • Small documents: Use chunk size 300-500
  • Technical documents: Use smaller chunks (200-400)

Model Selection:

  • General documents: llama3.2, mistral
  • Technical/Legal: qwen2.5
  • Code documentation: codellama

Troubleshooting

"Ollama not found"

  • Make sure Ollama is installed and in your PATH
  • Restart your terminal after installation
  • Try running ollama serve first

"Model not found"

  • Run ollama pull model-name to download the model
  • Check available models with ollama list

"No text extracted from PDF"

  • Ensure the PDF contains selectable text (not just images)
  • Try with a different PDF file
  • Check if the PDF is password protected

Memory Issues

  • Reduce chunk size in the web interface
  • Close other applications to free memory
  • Use a smaller model like llama3.2:8b instead of larger variants

Example Usage

  1. Upload the Arabic PDF: 1-النظام الأساسي للحكم.pdf

  2. Ask questions like:

    • "ما هو موضوع هذه الوثيقة؟"
    • "اذكر أهم النقاط في النظام"
    • "ما هي المواد المتعلقة بالحكم؟"

  3. The system will:

    • Find relevant sections in the PDF
    • Generate answers based only on the PDF content
    • Show you the source pages for verification