🧠 PDF Report Summarizer

This is a Streamlit-based application that allows users to upload PDF reports and generate concise, structured summaries using the LLaMA3 model through Ollama.

🔧 Features

• 📄 Upload and OCR-based text extraction from PDFs
• 🧩 Chunked summarization and final summary synthesis
• 🎯 Selectable report types: MPR, Board Report, Financial, etc.
• 📊 Page count estimation and time-to-completion display
• 🎛️ Custom or auto-limit on final summary page size
• 🟢 Progress bar with real-time chunk status
• ⛔ Stop processing at any point
• 🖨️ Copy/print support and styled HTML preview/download
• 🧠 Final summary always retained unless reset manually
• 🎛️ Final summaries are generated using user-selected prompt templates.
• 🧩 You can toggle chunk-level results and constrain length using a page selector.
• ⛔ Automatically prevents system sleep during long runs.
• 🎛️ Summary remains until a new file is uploaded or reset.

---

🧠 Model Used

• Model: llama3:8b-instruct-q4_K_M
• Manager: Ollama
• ollama pull llama3:8b-instruct-q4_K_M
• Auto-launched via subprocess if not already running

📦 Requirements

• Python 3.8+
• Tesseract-OCR (Installed and added to PATH)
• Poppler for PDF-to-image conversion (on Windows)

⚠️ Make sure you manually configure:

• pytesseract.pytesseract.tesseract_cmd in summarizer_pipeline.py
• poppler_bin path in summarizer_pipeline.py

✅ Dependencies (requirements.txt)

UI & Web App

• streamlit
• markdown

PDF Processing

• PyMuPDF==1.23.21 # used as fitz to extract text from PDFs
• pdf2image==1.16.3 # converts PDF pages to images for OCR

OCR

• pytesseract==0.3.10 # OCR engine (must install Tesseract binary separately)

HTTP Requests

• requests

pip install -r requirements.txt

🔧 External Dependencies

These tools must be installed separately and added to your system PATH.

• 📌 1. Tesseract OCR Used for extracting text from scanned PDF images.

• Tesseract-OCR (Installed and added to PATH)

• 📌 2. Poppler for Windows Used by pdf2image to convert PDF pages into images.

• Poppler for PDF-to-image conversion

---

🏁 How to Run

1. Clone the Repository

   git clone https://github.com/rychrr/pdf_summarizer_offline.git   
   

2. Install Python Dependencies

     pip install -r requirements.txt
   

3. Launch Streamlit App

    streamlit run summarizers.py
   
   

📁 Project Structure

pdf-summarizer/ ├── summarizers.py # Main Streamlit frontend app ├── summarizer_pipeline.py # Backend summarization pipeline ├── prompts/ # Prompt templates │ ├── mpr_prompt.txt │ ├── final_mpr_prompt.txt │ ├── board_prompt.txt │ └── final_board_prompt.txt ├── cache/ # OCR cache directory ├── img/ │ └── yourimg.png # Optional logo for styled summary ├── requirements.txt └── README.md

👨‍💻 Author - Ejike Ozonkwo

• LinkedIn
• Github

🤝 Contributing

Pull requests welcome. Open issues for enhancements or bugs.

🧾 License

MIT License