Skip to content

LEARNING_SYSTEM.md

Latest commit

 

History

History
353 lines (264 loc) · 8.3 KB

LEARNING_SYSTEM.md

File metadata and controls

353 lines (264 loc) · 8.3 KB

Adaptive Learning System - Implementation Status

✅ Sprint 1: Foundation (COMPLETED)

Core Services Implemented

1. Learning Profile Service (src/learning/learningProfile.ts)

Status: ✅ Complete and integrated

Features:

  • Student profile tracking with level, typing speed, error rate, performance score
  • Session statistics (blocks completed, hints used, avg completion time)
  • Performance scoring algorithm (0-100 scale)
  • Auto-save to VS Code globalState
  • Level transition logic (shouldLevelUp/shouldLevelDown)

Key Methods:

updateTypingSpeed(wpm: number)
recordError(errorType: string)
recordKeystroke(char: string, timeSinceLastKeystroke: number)
recordBlockCompleted(timeTaken: number)
shouldLevelUp(): boolean
shouldLevelDown(): boolean

Storage: Persists to VS Code workspace state as devxLearningProfile

2. Activity Monitor (src/learning/activityMonitor.ts)

Status: ✅ Complete and integrated

Features:

  • Real-time keystroke tracking with timing buffer (last 100 keystrokes)
  • WPM calculation based on keystroke timing
  • Idle detection with level-specific thresholds:
    • Beginner: 15s
    • Intermediate: 30s
    • Pro: 120s
  • Stuck detection (30s on same line with errors)
  • Callback system for idle/stuck events

Key Methods:

recordKeystroke(char: string, location: string)
checkIdleState()
onIdle(callback: (idleTime: number) => void)
onStuck(callback: (location: string) => void)

Integration: Wired to text document change events in extension.ts

3. Level Adapter (src/learning/levelAdapter.ts)

Status: ✅ Complete and integrated

Features:

  • Automatic level suggestions based on performance
  • User-friendly prompts with emojis (🌱🌿🌳)
  • Level descriptions and icons
  • Non-intrusive suggestions

Thresholds:

  • Level up: Performance score ≥ 85
  • Level down: Performance score ≤ 40

User Commands Added

1. devx.changeLevel

Status: ✅ Registered in package.json and extension.ts

Features:

  • Quick pick menu with 3 levels
  • Shows level descriptions
  • Updates status bar immediately
  • Confirmation message

Keybinding: None (Command Palette only)

2. devx.resetProfile

Status: ✅ Registered in package.json and extension.ts

Features:

  • Modal confirmation dialog
  • Clears all learning progress
  • Resets activity monitor
  • Updates status bar

3. devx.viewStats

Status: ✅ Registered in package.json and extension.ts

Features:

  • Shows comprehensive learning statistics
  • Modal dialog with formatted stats:
    • Current level
    • Performance score
    • Typing speed (WPM)
    • Accuracy percentage
    • Average completion time
    • Blocks completed
    • Hints used

UI Elements

Level Status Bar Item

Status: ✅ Implemented and visible

Features:

  • Shows current level with icon (🌱/🌿/🌳)
  • Click to change level
  • Always visible in status bar
  • Updates in real-time

Position: Left side of status bar

Activity Monitoring Integration

Text Document Change Listener

Status: ✅ Integrated in extension.ts

Features:

  • Captures every text change
  • Records keystroke with location
  • Updates learning profile
  • Checks for idle/stuck states

Idle Detection Callbacks

Status: ✅ Wired in extension.ts

Behavior:

  • Shows notification after idle threshold
  • Offers help options
  • Can trigger guidance command
  • Non-blocking

Stuck Detection Callbacks

Status: ✅ Wired in extension.ts

Behavior:

  • Detects prolonged stay on same line with errors
  • Shows logic guidance prompt
  • Links to guidance command
  • Encourages problem-solving

📊 Current Statistics Tracked

  1. Performance Metrics:

    • Typing speed (WPM)
    • Error rate (%)
    • Performance score (0-100)
    • Average completion time

  2. Session Data:

    • Blocks completed
    • Hints used
    • Errors encountered
    • Total keystrokes

  3. Behavioral Patterns:

    • Idle frequency
    • Stuck frequency
    • Error types distribution

🔄 Next Steps (Sprint 2-5)

Sprint 2: Mode System

Status: ⏳ Not started

Tasks:

  • Create src/learning/modes/beginnerMode.ts
  • Create src/learning/modes/intermediateMode.ts
  • Create src/learning/modes/proMode.ts
  • Implement mode-specific code generation logic
  • Update TeacherSlate to use mode system

Sprint 3: Comment Hints System

Status: ⏳ Not started

Tasks:

  • Create src/learning/commentHints.ts
  • Implement Intermediate mode hint generation
  • Add context-aware hint logic
  • Wire to document change events

Sprint 4: Code Blocks System

Status: ⏳ Not started

Tasks:

  • Create src/learning/codeBlocks.ts
  • Implement block generation from lessons
  • Add block completion tracking
  • Integrate with performance scoring

Sprint 5: Sensei Mentor (Logic Guidance)

Status: ⏳ Not started

Tasks:

  • Create src/learning/senseiMentor.ts
  • Implement logic-focused guidance (NOT syntax)
  • Add architectural insights
  • Wire to idle/stuck callbacks
  • Create context analysis system

Important: Sensei focuses on LOGIC and ARCHITECTURE, not syntax help!

🛠️ Technical Details

Dependencies

  • VS Code Extension API ^1.95.0
  • @google/generative-ai ^0.24.1 (for AI guidance)
  • TypeScript ^5.9.3

Build Status

✅ Compiles successfully with webpack ✅ No runtime errors ⚠️ Minor TypeScript warnings (NodeJS.Timeout, setTimeout types - safe to ignore)

Storage

  • Learning profiles: context.globalState.get('devxLearningProfile')
  • API keys: VS Code secrets API
  • Settings: VS Code workspace configuration

🚀 Testing the Learning System

Manual Testing Steps

  1. Install/Reload Extension

    • Press F5 in VS Code
    • Or package with vsce package and install .vsix

  2. Check Status Bar

    • Should see "🌱 Beginner" in status bar
    • Click it to test level change menu

  3. Test Commands

    • Open Command Palette (Ctrl+Shift+P)
    • Try DevX: Change Learning Level
    • Try DevX: View Learning Statistics
    • Try DevX: Reset Learning Profile

  4. Test Activity Monitoring

    • Open a code file
    • Start typing code
    • Wait 15 seconds (idle detection should trigger)
    • Stay on one line with errors for 30s (stuck detection)

  5. Check Profile Persistence

    • Type some code
    • Check stats (should show keystrokes, WPM)
    • Reload VS Code window
    • Check stats again (should persist)

📝 Configuration Options

Currently using defaults. Future sprints will add:

  • Custom idle thresholds
  • Performance score weights
  • Level transition thresholds
  • AI provider preferences for guidance

🐛 Known Issues

  1. TypeScript warnings for NodeJS.Timeout and setTimeout

    • Impact: None (types work at runtime)
    • Fix: Add @types/node if needed

  2. Sensei guidance command placeholder

    • Status: Uses existing askSensei command
    • Fix: Implement requestGuidance in Sprint 5

📚 Architecture

src/
├── extension.ts                 ✅ Main entry point (integrated)
├── learning/
│   ├── learningProfile.ts      ✅ Profile tracking (complete)
│   ├── activityMonitor.ts      ✅ Behavior monitoring (complete)
│   ├── levelAdapter.ts         ✅ Auto-adaptation (complete)
│   ├── modes/                  ⏳ To be created (Sprint 2)
│   │   ├── beginnerMode.ts
│   │   ├── intermediateMode.ts
│   │   └── proMode.ts
│   ├── commentHints.ts         ⏳ To be created (Sprint 3)
│   ├── codeBlocks.ts           ⏳ To be created (Sprint 4)
│   └── senseiMentor.ts         ⏳ To be created (Sprint 5)
└── ... (existing files)

✨ Features Summary

Implemented ✅

  • Student profile with 5+ metrics
  • Real-time activity monitoring
  • Idle/stuck detection
  • Manual level selection
  • Statistics dashboard
  • Profile persistence
  • Status bar integration
  • Auto-level suggestions

Planned ⏳

  • 3-mode code generation system
  • Comment-based hints
  • Block-based learning
  • Logic-focused mentoring
  • Context-aware guidance
  • Architecture insights
  • Custom thresholds

Last Updated: Sprint 1 Completion Version: 1.1.0 Status: Foundation Complete, Ready for Sprint 2