📚 Book Reader

Turning textbooks into study decks — automatically.

Book Reader is a Streamlit application that converts textbook-style PDF or TXT documents into interactive, chapter-aware study decks. It extracts document structure, segments content by chapter and heading, generates concise Markdown slides via OpenRouter, supports contextual chatbot Q&A on the active section, and exports the final deck as a downloadable PowerPoint file.

This project was completed as a B.Tech Computer Science and Engineering major project at Delhi Technical Campus, GGSIPU.

🔗 Live Demo: srijan-bookreader.streamlit.app 🔗 GitHub: github.com/Srijan1683/BookReader

---

🎓 Academic Evaluation

This project was submitted and graded as the final-year major project.

Evaluation Component	Score
Dissertation & Viva Voce	93
Project Progress Evaluation	97
Overall Percentage	95%
Final Semester CGPA	10

Institution: Delhi Technical Campus, GGSIPU (Guru Gobind Singh Indraprastha University)

---

✨ Features

• Upload PDF or TXT documents through a clean Streamlit web interface
• Extract PDF table-of-contents data with PyMuPDF
• Segment documents into chapter-level and heading-level study chunks
• Generate 2–4 concise study slides per heading using OpenRouter-hosted LLMs
• Local heuristic fallback when API access is unavailable — the app never fully breaks
• Browse the deck slide-by-slide with previous/next controls
• Jump directly to any chapter or slide via dropdown navigation
• Continuous, globally numbered slides across the whole deck
• Ask a contextual chatbot questions about the currently active section
• Export the full generated deck as a downloadable .pptx file

🛠️ Tech Stack

Layer	Tools
UI	Streamlit
LLM	OpenRouter (direct HTTP client — no SDK dependency)
PDF Parsing	PyMuPDF
Slide Export	python-pptx
Config	python-dotenv
Testing	pytest

🏗️ Architecture

Application flow:

1. webapp.py accepts the uploaded document and starts the generation flow
2. src/utils.py extracts pages and table-of-contents entries
3. src/segmentor.py groups the document into chapter and heading chunks
4. src/slide_generator.py generates Markdown slide content via OpenRouter, with a local fallback path when the API is unavailable
5. src/deck.py normalizes slide titles and assigns global slide numbers across the deck
6. webapp.py renders the slide browser, chatbot, chapter/slide jump controls, and PowerPoint export

Key design decisions:

• OpenRouter is called directly through Python's standard HTTP library (src/openrouter_client.py) rather than the OpenAI SDK, avoiding SDK/proxy compatibility issues
• The model is fully configurable via the OPENROUTER_MODEL environment variable — no hard-coded model IDs
• A local heuristic slide generator acts as a graceful fallback (ALLOW_LOCAL_SLIDE_FALLBACK) if the API is down or unreachable
• ChromaDB-backed retrieval (rag/) exists as an optional legacy path and is not part of the main production flow

📁 Project Structure

Book-Reader-m/
├── webapp.py                         # Streamlit app and UI
├── main.py                           # Document-to-slides pipeline
├── requirements.txt                  # Python dependencies
├── README.md                         # Project documentation
├── scripts/
│   └── generate_test_pdf.py          # Test PDF generator
├── src/
│   ├── deck.py                       # Slide normalization and deck flattening
│   ├── models.py                     # Shared data models
│   ├── openrouter_client.py          # OpenRouter chat-completion client
│   ├── segmentor.py                  # Chapter and heading segmentation
│   ├── slide_generator.py            # Slide generation plus local fallback
│   ├── utils.py                      # PDF loading and TOC helpers
│   ├── llm.py                        # Prompt helper
│   ├── prompts.yaml                  # Prompt templates
│   └── vector_db.py                  # Vector search helper
├── rag/                               # Optional legacy retrieval-augmented chat
│   ├── chat_engine.py
│   ├── embed.py
│   └── retriever.py
├── test_assets/
│   └── book_reader_test.pdf          # Stable test input
└── tests/
    ├── test_deck.py
    ├── test_toc.py
    └── test_chapter_processor.py

🚀 Getting Started

1. Clone and install dependencies:

git clone https://github.com/Srijan1683/BookReader.git
cd BookReader
pip install -r requirements.txt

2. Configure environment variables (create a .env file):

OPENROUTER_API_KEY=your_key_here
OPENROUTER_MODEL=your_model_id
ALLOW_LOCAL_SLIDE_FALLBACK=true

3. Run the app:

python -m streamlit run webapp.py

🧪 Testing

pytest -q

Run a quick syntax check across core modules:

python -m py_compile webapp.py main.py src/openrouter_client.py src/slide_generator.py src/segmentor.py src/deck.py

---

👤 Author

Srijan Banerjee B.Tech, Computer Science and Engineering — Delhi Technical Campus, GGSIPU

• Portfolio: srijan-banerjee.netlify.app
• GitHub: github.com/Srijan1683
• LinkedIn: linkedin.com/in/banerjee-srijan