PDF Keyword Extractor Web

A web-based application to extract keywords and summaries from PDF documents, with OCR support for scanned PDFs. The app uses Flask for the web interface, PyMuPDF for PDF text extraction, and Tesseract OCR for scanned PDFs. Keyword extraction is available via Frequency and TF-IDF methods, and automatic text summarization is provided.

Features

• Extract text from PDFs (supports OCR for scanned PDFs)
• Keyword extraction using Frequency or TF-IDF methods
• Automatic text summarization
• Custom stopwords support (via data/stopwords.txt)
• User-friendly web interface

Installation

1. Install System Dependencies

• Tesseract OCR and Poppler are required for OCR and PDF-to-image conversion.
  • Windows:
    • Tesseract
    • Poppler
    • Ensure Tesseract is installed in C:\Program Files\Tesseract-OCR\tesseract.exe or C:\Program Files (x86)\Tesseract-OCR\tesseract.exe, or update the path in your system environment variables.
  • macOS:
    • brew install tesseract poppler
  • Linux:
    • sudo apt install tesseract-ocr poppler-utils

2. Clone the Repository

git clone https://github.com/rajdeep13-coder/PDF_Keyword_extractor.git
cd PDF_Keyword_extractor

3. Install Python Dependencies

It is recommended to use a virtual environment.

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt

Usage

1. Start the Application:

   python app.py

   The app will run on http://127.0.0.1:5000/ by default.

2. Upload a PDF:

   • Go to the web interface.
   • Upload a PDF file.
   • Choose the keyword extraction method (Frequency or TF-IDF).
   • The app will display extracted keywords and a summary.

3. OCR Support:

   • If the PDF is scanned or contains images, OCR will be automatically used if text extraction fails or is insufficient.
   • Ensure Tesseract and Poppler are installed and accessible.

Project Structure

PDF_Keyword_extractor/
├── app.py                  # Main Flask application
├── extractor/
│   ├── keyword_finder.py   # Keyword extraction and summarization logic
│   └── pdf_reader.py       # PDF text extraction with OCR support
├── data/
│   └── stopwords.txt       # Custom stopwords (optional)
├── static/
│   └── styles.css          # CSS styles
├── templates/
│   └── index.html          # Web interface template
├── requirements.txt        # Python dependencies
└── README.md               # Project documentation

Configuration

• Secret Key: The Flask app uses a secret key from the SECRET_KEY environment variable, or defaults to 'your_secret_key_here'.
• Stopwords: You can customize stopwords by editing data/stopwords.txt.

Dependencies

• Flask
• nltk
• scikit-learn
• gensim
• scipy
• PyMuPDF
• Pillow
• gunicorn (for production deployment)
• pdf2image
• pytesseract

Notes

• For best OCR results, ensure Tesseract and Poppler are correctly installed and configured in your system PATH.
• The app will automatically use OCR if the extracted text is too short or if explicitly requested in the code.
• Summarization uses a TF-IDF-based approach for sentence ranking.

License

MIT License