🎓 RFAI (Routine Focus AI)

Your Intelligent Time-Block Learning System with Real Content & Progress Tracking

RFAI is an AI-powered learning companion that structures your day into focused learning blocks, fetches actual educational content, and tracks your progress through quizzes and analytics.

🎉 What's New (v2.1)

✅ Real Content Fetching

• YouTube Videos: Actual videos from YouTube API (not just topics!)
• Research Papers: Real papers from ArXiv based on your interests
• Movies: Curated recommendations from IMDB
• Auto-Display: Content shows automatically on dashboard per time block

✅ AI-Powered Study Plans

• Perplexity Integration: Get personalized recommendations from your study plan
• Smart Discovery: AI finds best videos, papers, books for your goals
• Context-Aware: Recommendations adapt to your learning progress

✅ Progress Testing & Quizzes

• Quiz Generation: Create tests for any topic (Quantum Computing, ML, etc.)
• Auto-Grading: Instant feedback with explanations
• Progress Tracking: View scores, weak areas, improvement trends
• Subject Analytics: Track progress per topic

✨ Core Features

• ⏰ Time-Block System - 3 daily blocks: Science (3h), Self-Help (1h), Movies (1.5h)
• 🔒 Soft-Lock Access Control - Content restricted during active blocks for focus
• 📹 Multimodal Attention - Camera, mic, keyboard, mouse tracking
• 🎨 Dynamic Themes - Visual themes change per time block
• 📊 Real-Time Dashboard - Shows content, attention score, progress
• 💾 Session Management - Track learning sessions with analytics
• 🔒 Privacy-First - All data local, camera/mic optional

🚀 Quick Start (2 Minutes)

1. Configure API Keys

# Copy example configuration
cp .env.example .env

# Edit and add your API keys
nano .env

Get Free API Keys:

• YouTube API - For video recommendations
• Perplexity API - For web search
• OMDb API - For movie data

2. Customize Your Interests

Edit interests.json with your learning preferences:

• YouTube topics (Quantum Mechanics, Philosophy, etc.)
• Research paper fields and ArXiv categories
• Movie preferences (directors, genres, style)

3. Run the Application

macOS / Linux:

./run.sh

Windows:

run.bat

4. Open Dashboard

Navigate to: http://localhost:5000/dashboard

📖 How It Works

Your 3-Hour Daily Learning Plan

RFAI implements a structured learning schedule based on daily_3hr_plan.md:

Time Block	Duration	Content	Purpose
YouTube	1 hour	Educational entertainment	Animated history, philosophy, self-help
Papers	1 hour	Academic research	ArXiv papers on your topics
Movies	1-2 hours	Artistic films	Film-school quality cinema

Content Intelligence

YouTube Filtering:

• ✅ Educational - Pure learning (lectures, tutorials)
• ✅ Learning-Borderline - What you want (educational entertainment)
• ❌ Entertainment - Excluded (vlogs, reactions, gaming)

Movie Classification:

• 🎬 Artistic - Film school worthy (Tarkovsky, Kubrick, etc.)
• ⭐ Good Generic - Well-rated commercial films
• 🚫 Pure Entertainment - Excluded

After each movie, the system prompts you to reflect:

• What was the central theme?
• How did cinematography enhance the story?
• What techniques stood out?
• What did you learn?

Dashboard Features

• 📊 Time Allocation Chart - Visual breakdown of daily hours
• 📈 Focus Over Time - Track your attention levels
• ✅ Progress Tracker - See completion towards 3-hour goal
• 🎯 Current Position - Know your week and day in the plan
• 📝 Content Recommendations - Direct links to videos, papers, films

📚 Documentation

• Complete Setup Guide - Detailed configuration and troubleshooting
• API Documentation - All available endpoints
• Configuration Reference - Full .env and interests.json options

🔌 API Endpoints

Schedule & Content

• GET /api/schedule/daily - Get today's 3-hour schedule
• GET /api/schedule/current-block - Get active content block
• POST /api/content/rate - Rate content (1-5 stars)
• POST /api/movie/post-review - Submit movie review

Learning Plans

• GET /api/plans - List all plans
• POST /api/plans/generate - Generate new plan
• GET /api/plans/<id>/current-day - Get current day

Activity & Stats

• GET /api/stats/daily - Daily statistics
• GET /api/activity/today - Activity logs
• GET /api/focus/current - Current focus state

🎯 Configuration

Environment Variables (.env)

# Required
YOUTUBE_API_KEY=your_key
PERPLEXITY_API_KEY=your_key
OMDB_API_KEY=your_key

# Optional
NOTION_API_KEY=your_key
GEMINI_API_KEY=your_key
OLLAMA_BASE_URL=http://localhost:11434

Learning Interests (interests.json)

{
  "youtube_interests": {
    "topics": ["Quantum Mechanics", "Philosophy"],
    "content_style": "learning_borderline_entertainment"
  },
  "research_paper_interests": {
    "fields": ["Physics", "Neuroscience"],
    "arxiv_categories": ["quant-ph", "q-bio.NC"]
  },
  "movie_interests": {
    "style": "artistic_film_school",
    "directors": ["Tarkovsky", "Kubrick"],
    "post_viewing_review": true
  }
}

🛠️ Advanced Options

Disable Background Daemons

NO_DAEMONS=1 ./run.sh

Custom Database Location

export RFAI_DB_PATH="/path/to/database.db"
./run.sh

Enable Debug Logging

DEBUG=1 ./run.sh

Export Focus Data for ML Analysis

# Via API
curl http://localhost:5000/api/activity/today > focus_data.json

# Or click "Export Focus Data" in dashboard

📊 Data Collection

RFAI collects data locally for your analysis:

• Activity logs (apps, websites, duration)
• Focus states (when daemons enabled)
• Content ratings and reviews
• Time allocation breakdown

All data stays on your machine. Export anytime for your own ML experiments.

🐛 Troubleshooting

Problem	Solution
API keys not working	Check .env file, restart server
Learning plan shows 0 hours	Verify daily_3hr_plan.md format
No content recommendations	Test API keys, check interests.json
Focus data empty	Enable daemons, check platform support

Full troubleshooting guide: SETUP_GUIDE.md

🔒 Privacy & Security

• ✅ API keys never logged or printed
• ✅ .env in .gitignore - safe from commits
• ✅ Camera/screen tracking is opt-in
• ✅ All data stored locally
• ✅ No cloud uploads without explicit config

📦 Dependencies

Core dependencies are listed in requirements.txt:

• Flask - API server
• ArXiv - Paper search
• Chart.js - Dashboard visualizations
• ChromaDB - Vector storage (for advanced features)
• Sentence Transformers - Content embeddings

🗂️ Project Structure

Learning_AI/
├── rfai/
│   ├── api/           # API server
│   ├── ai/            # AI components (scheduler, parser, RL)
│   ├── integrations/  # YouTube, ArXiv, EdX, IMDb
│   ├── ui/            # Dashboard UI
│   └── config/        # Configuration loaders
├── database/          # SQLite schema and init
├── .env.example       # API key template
├── interests.json     # User preferences
├── daily_3hr_plan.md  # 52-week learning plan
└── SETUP_GUIDE.md     # Complete setup docs

🎓 Legacy Version

The original Learning_AI demo is preserved in legacy/learning_ai_basic/ for reference.

🤝 Contributing

This is a personal learning system, but suggestions are welcome! Please ensure:

• API keys remain secure
• Privacy features stay opt-in
• Documentation is updated

📄 License

MIT License - See LICENSE file

🆘 Need Help?

1. Read SETUP_GUIDE.md
2. Check rfai.log for errors
3. Test API endpoints: http://localhost:5000/api/status
4. Verify API keys are valid

---

Made with 🧠 for structured, intentional learning