This project extracts structured headings (H1, H2, H3) from .pdf documents using Python and PyMuPDF.
It is containerized using Docker, following Adobe's constraints:

• Processes all PDFs inside /app/input/
• Outputs structured .json to /app/output/
• Runs without network access

.
├── extractor.py          # Main Python script
├── requirements.txt      # Dependencies
├── Dockerfile            # Docker container setup
├── input/                # Put your PDFs here
│   └── sample.pdf
├── output/               # JSON files appear here
│   └── sample.json

docker build --platform=linux/amd64 -t adobe-solution:harshi123 .

docker run --rm -v "${PWD}/input:/app/input" -v "${PWD}/output:/app/output" --network none adobe-solution:harshi123

• On Windows CMD, use %cd% instead of ${PWD}
• On macOS/Linux, ${PWD} works fine

Place all .pdf files you want to process into the /input folder.

Each PDF is converted into a .json file in /output/, e.g.:

{
  "title": "Abstract:",
  "outline": [
    { "level": "H1", "text": "Abstract:", "page": 2 },
    { "level": "H1", "text": "Existing System:", "page": 5 }
    // ...
  ]
}

• No external network usage (--network none)
• Input/output handled via volume mounts only
• Lightweight base image (python:3.10-slim)
• Batch processes all .pdf files in /input/

• Harshini Nadendla
• Spurthi Inturu
• Chakrish Vejendla

This project was developed as part of Adobe's Hackathon Challenge – Round 1A.