feat: initialize agentics hackathon configuration #14
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Add .hackathon.json with project setup for TV5 Hackathon: - Project name: hackathon-tv5 - Team name: agentics - Configured available tool options
Complete Samsung TV integration for the Agentics TV5 Hackathon: Features: - Device discovery via SSDP/UPnP on local network - WebSocket-based TV control (port 8002) - Wake-on-LAN power on support - Remote key commands (power, volume, navigation, playback) - App management (list, launch streaming apps) - MCP server integration for AI agent access (STDIO + SSE) - CLI interface for direct TV control Supported streaming apps: - YouTube, Netflix, Prime Video, Disney+, Spotify - Apple TV, HBO Max, Hulu, Plex, Twitch MCP Tools (13): - samsung_tv_discover, samsung_tv_list, samsung_tv_connect - samsung_tv_power, samsung_tv_volume, samsung_tv_navigate - samsung_tv_key, samsung_tv_apps, samsung_tv_launch_app - samsung_tv_home, samsung_tv_status, samsung_tv_set_default - samsung_tv_remove Tech stack: - TypeScript, Node.js 18+ - samsung-tv-control, node-ssdp, wake_on_lan - Zod for schema validation - Vitest for testing (27 tests passing)
- Add Q-Learning preference learning module with experience replay - Implement WASM-optimized cosine similarity for content embeddings - Create content embedding generation with genre/type/rating features - Add ReasoningBank-style pattern storage for successful viewing patterns - Build SmartTVClient with automatic session tracking and learning - Create 13 MCP learning tools for AI agent integration - Add LearningPersistence for file-based model storage - Include IndexedDB persistence for browser/WASM environments - Integrate learning tools with existing MCP server - Add comprehensive test suite (23 tests) for learning system Learning system features: - Epsilon-greedy action selection - Temporal difference Q-value updates - Experience replay for better sample efficiency - User preference learning from viewing behavior - Content similarity using 64-dimension embeddings - Time-of-day and contextual recommendations
- Add TMDb API client with caching and full endpoint coverage - Support search, trending, popular, top-rated, discover endpoints - Map TMDb genres to learning system genres - Include streaming provider detection for deep linking - Convert TMDb content to ContentMetadata format Content Discovery MCP Tools (12 new tools): - content_search: Search movies and TV shows - content_trending: Get trending content - content_popular: Get popular content - content_top_rated: Get top-rated content - content_discover: Filter by genre, rating, year - content_details: Get detailed info with cast - content_similar: Find similar content - content_recommendations: TMDb recommendations - content_now_playing: Movies in theaters - content_upcoming: Upcoming releases - content_personalized: Learning-based recommendations - content_for_mood: Mood-based suggestions Also adds: - posterUrl and backdropUrl to ContentMetadata - Integration with learning system - 21 new tests (71 total)
- Rewrite main README.md focused on Samsung TV integration - Add docs/user-guide/ with complete usage guide - Add docs/developer/ with architecture and API reference - Move VERCEL_SETUP.md and WORKFLOWS.md to docs/developer/ Documentation now includes: - 38 MCP tools reference - TV control, learning system, content discovery guides - Q-Learning algorithm explanation - Code examples and type definitions - Troubleshooting section
- Add scripts/train-benchmark.ts with Q-Learning training simulation - Include 20 sample content items (movies, TV shows, documentaries) - Simulate 5 user profiles with different viewing preferences - Benchmark embedding generation (135K ops/sec) - Benchmark cosine similarity (1.3M ops/sec WASM-optimized) - Benchmark batch search (81K ops/sec) - Benchmark cache performance (99.6% hit rate) - Train over 500 episodes with experience replay - Track reward improvement and top actions Results: - Embedding: 135,448 ops/sec - Similarity: 1,285,875 ops/sec - Training: 0.18s for 500 episodes - Patterns learned: 609
- Add problem statement (45 min decision time) - Add solution overview with 4 key benefits - Include demo conversation example - Add architecture diagram - Include benchmark results table - Add Claude Desktop integration guide - List all 38 MCP tools - Add tech stack table - Include roadmap section
…chmarks - Add detailed Q-Learning algorithm explanation with states, actions, rewards - Add content embedding explanation with 64-dim vector breakdown - Add WASM-optimized similarity calculation code example - Add complete feature tables for all 38 MCP tools with examples - Add step-by-step setup tutorial (prerequisites, TMDb key, installation) - Add architecture diagram and file structure explanation - Add technology stack with rationale for each choice - Add real-world timing benchmarks - Fix ESM/CJS import issue in discovery.ts for node-ssdp
- Validates all 38 MCP tools (13 TV, 13 learning, 12 content) - Tests Q-Learning system configuration and state management - Tests 64-dim embedding generation and similarity search - Validates 4 Zod schemas (device, app, content, session) - Tests TMDb client initialization with mood mapping - Runs performance benchmarks (128K embeddings/sec, 1.1M similarity/sec) - Verifies all CLI entry points exist
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Add .hackathon.json with project setup for TV5 Hackathon: