🧠 Mind Palace - AI-Powered Learning Platform

An interactive Streamlit app that turns PDFs into full learning notebooks with AI summaries, topic-aware flashcards, quizzes, study plans, mnemonics, progress tracking, and a Socratic tutor.

🌟 Features

• PDF → Notebook pipeline: upload once, store PDF/text/summary/topics in MongoDB
• AI Summary: Groq-hosted openai/gpt-oss-120b (default) over full document text
• Topic Extraction: AI prompt + heuristic fallback; topic-aware text slicing via local ONNX embeddings
• Learning tools: flashcards, quizzes (structured output), study scheduler, progress tracker, memory aids (acronym/song/phrase/story)
• RAG-style chat: Socratic “Talk to Doc” with topic-focused context
• Gamification: task points, achievements, and sidebar “today’s tasks”

🚀 Getting Started

Prerequisites

• Python 3.10+
• MongoDB running locally (mongodb://127.0.0.1:27017/mind_palace by default)
• Groq API key (for groq client); optional custom model name
• ONNX embedding file at onnx/model_int8.onnx (nomic-embed-text-v1.5, int8)

Installation & Run

cd "c:\Users\vishn\PROJECT\STEP BY STEP DEC 7"
pip install -r requirements.txt

Create .env:

GROQ_API_KEY=your_api_key
GROQ_MODELS=openai/gpt-oss-120b    # optional, defaults to this
GROQ_MODEL=openai/gpt-oss-120b     # used on Talk to Duck page
MONGODB_URI=mongodb://127.0.0.1:27017/mind_palace

Run:

streamlit run app.py

Open http://localhost:8501.

Download the ONNX embedding model (one-time setup):

powershell -ExecutionPolicy Bypass -File .\scripts\download_nomic_model.ps1 -InstallDeps

📁 Project Structure

STEP BY STEP DEC 7/
├── app.py                      # Home: upload, notebook list, sidebar tasks
├── pages/
│   ├── 1_📄_Summary.py         # Summary + topics + stats
│   ├── 2_📖_PDF_Viewer.py      # Inline viewer + download
│   ├── 3_🎴_Flashcards.py      # Topic-aware generation + review
│   ├── 4_📝_Quiz.py            # Structured MCQ quizzes + scoring
│   ├── 5_💬_Talk_to_Duck.py     # Socratic tutor with topic context
│   ├── 6_📅_Study_Scheduler.py # Day-by-day plan + task completion
│   ├── 7_🎯_Progress_Tracker.py# Points, achievements, completion %
│   └── 8_🧠_Acronym_Generator.py# Mnemonics (acronym/song/phrase/story/all)
├── utils/
│   ├── db.py                   # Mongo CRUD, embedded flashcards/quizzes/acronyms/progress
│   ├── helpers.py              # Groq client, PDF text extraction, prompt loader
│   ├── text_extraction.py      # Topic-aware extraction via ONNX + fallback
│   ├── onnx_embedder.py        # Nomic embed ONNX wrapper
│   └── sidebar_utils.py        # Shared sidebar renderer
├── prompts/                    # JSON prompt configs (summary, topics, flashcards, scheduler, quiz, mnemonics)
├── scripts/
│   └── download_nomic_model.ps1 # One-command Nomic ONNX model downloader
├── onnx/model_int8.onnx        # Local embedding model (int8)
├── requirements.txt
└── .env                         # Runtime config (Groq + MongoDB)

🎯 Usage Workflow

1. Upload a PDF → notebook is created with full-text summary + topics.
2. Review Summary, then PDF Viewer for the source.
3. Generate Flashcards or Quiz per topic (uses topic-aware text extraction).
4. Build a Study Schedule and mark tasks done; Progress Tracker updates points/achievements.
5. Use Memory Aid Generator for acronyms/songs/phrases/stories or Talk to Doc for Socratic Q&A.

🔧 Technical Stack

• Frontend: Streamlit multipage
• LLM: Groq (groq client, default openai/gpt-oss-120b)
• Embeddings: Local ONNX nomic-embed-text-v1.5 (int8) for topic-aware slicing
• Database: MongoDB (single notebooks collection with embedded progress/flashcards/quizzes/acronyms)
• PDF/Text: PyPDF2 extraction; Base64 PDF storage
• Infra: python-dotenv, pydantic for structured quiz responses

📝 Prompt Management

Prompts live in prompts/*.json with system_instruction + user_instruction. helpers.call_gemini formats placeholders ({topic}, {text}, {target_text}, etc.) and sends context text plus prompt to Groq.

🎮 Gamification

• Points: schedule tasks (10–20 pts each) + quiz scores added to total.
• Achievements: 🌱50, 🔥100, ⭐250, 🎓500 points; ✅10 tasks; 🎯50% completion; 💯100% completion.
• Sidebar shows “today’s tasks” when a notebook is selected.

🐛 Troubleshooting

• MongoDB: ensure mongod is running; check MONGODB_URI.
• Groq auth: verify GROQ_API_KEY; adjust model via GROQ_MODELS/GROQ_MODEL.
• Embeddings: confirm onnx/model_int8.onnx exists; install onnxruntime, transformers, numpy.
• PDF text: encrypted/scanned PDFs may extract empty text; provide selectable text PDFs.

---

Happy Learning! 🎓