AI Study Planner

Syllabus → topics → week-by-week study plan + curated videos. A Flask app built by a 4-person team; I owned the backend.

Team: Tushar (backend) · Manipal (backend) · Poras (frontend) · Swayam (frontend)

1. Upload a syllabus (PDF or TXT).
2. Extract topics (syllabus_processor.py — heading/regex parsing).
3. Schedule topics across weeks (scheduler.py — deadline-aware spreading).
4. Recommend YouTube videos per topic (video_recommender.py).
5. Render dashboard + result pages.

upload → pdfplumber (fallback: PyPDF2) → extract_topics() → session
      → dashboard.html (topics + video links)
      → result.html (generated schedule)

app.py                    # Flask app (all routes)
syllabus_processor.py     # topic extraction
scheduler.py              # schedule generation
video_recommender.py      # video suggestions
index.html upload.HTML    # ┐
dashboard.html result.html│ ├ must live in templates/ (see step 1)
style.css                 # ┘ must live in static/
README.md
extras: project video (mp4), report (pdf), slides (pptx)

Important: Flask only serves templates from templates/. The HTML files currently sit at the repo root — reorganize once after cloning (nothing is deleted, just moved into place):

git clone https://github.com/Tusharkapoor-oop/tushar_cse-AI-and-ML-A_AI-STUDY-PLANNER.git
cd tushar_cse-AI-and-ML-A_AI-STUDY-PLANNER

# one-time: put templates/static where Flask expects them
mkdir -p templates static
mv index.html dashboard.html result.html templates/
mv upload.HTML templates/upload.html     # note: lowercase .html for Linux
mv style.css static/

# dependencies
pip install flask pdfplumber PyPDF2 werkzeug flask_sqlalchemy flask_login

# run
python app.py
# → http://127.0.0.1:5000

Test with any text/PDF syllabus — the app deletes your upload immediately after parsing.

• Encoding-agnostic text reads: TXT uploads are tried as UTF-8 → Latin-1 → UTF-16.
• PDF fallback chain: pdfplumber first (tables + layout), PyPDF2 if that yields nothing.
• Session-scoped state: topics live in the Flask session — no database required for the core flow.
• Upload hygiene: secure_filename + extension allow-list + immediate os.remove after parse.
• Limit: 16 MB upload cap (MAX_CONTENT_LENGTH).

• Topic extraction is rule-based (regex/heading heuristics), not an LLM — precision varies with syllabus formatting.
• app.run(debug=True) is the documented dev mode; put a real WSGI server in front for anything shared.
• No automated tests yet — planned: pytest fixtures around extract_topics() with 3 sample syllabi.
• The walkthrough video (16.7 MB project video final (2) (1) (1).mp4) will move to a GitHub Release to keep the clone small.

Project walkthrough (Google Drive)

No license file yet — MIT intended (to be added by the repository owner).