MultiViewer

# CLAUDE.md This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository. ## Project Overview This is `mvf1` - a Python package, command line interface, and MCP (Model Context Protocol) server for controlling MultiViewer video players. MultiViewer is a desktop application for watching Formula 1 and other motorsports with multiple synchronized video streams. ## Key Architecture ### Core Components - **`mvf1/mvf1.py`**: Main library with `MultiViewerForF1` class and `Player` class - `MultiViewerForF1`: Primary interface for controlling MultiViewer via GraphQL API (localhost:10101) - `Player`: Represents individual video players with methods for control (pause, mute, seek, etc.) - Communicates with MultiViewer desktop app via GraphQL at `http://localhost:10101/api/graphql` - **`mvf1/mcp.py`**: FastMCP server implementation exposing MultiViewer controls as MCP tools - Provides AI agents with tools to control video players - Tools include: `f1_live_timing_state()`, `players()`, `player_create()`, `player_delete()`, etc. - **`mvf1/cmdline.py`**: Click-based CLI with two main command groups - `players`: Control all players (list, close, mute, pause, sync) - `player`: Control specific player by ID or title - **`mvf1/mvf1_schema.py`**: Auto-generated GraphQL schema definitions using sgqlc ### Key Features - **Player Management**: Create, delete, control individual video players - **Live Timing Integration**: Access F1 and WEC live timing data - **Synchronization**: Sync all players to commentary or specific player timestamp - **MCP Integration**: Expose functionality to AI agents via Model Context Protocol ## Development Commands ### Testing ```bash # Run all tests across Python versions tox # Run tests with coverage pytest --cov=./ ``` ### Installation ```bash # Install in development mode pip install -e ./ ``` ### CLI Usage ```bash # List all active players mvf1-cli players ls # Run MCP server with default URL mvf1-cli mcp # Run MCP server with custom URL mvf1-cli mcp --url http://192.168.1.100:10101/api/graphql ``` ## Dependencies - **sgqlc**: GraphQL client for communicating with MultiViewer - **click**: Command line interface framework - **fastmcp**: MCP server implementation ## Testing Notes - Tests are in `tests/` directory - Uses pytest with asyncio support - Tox configuration tests Python 3.10-3.13 - Test data includes `players.json` fixture