strat-gameplay-webapp/CLAUDE.md
Cal Corum d60b7a2d60 CLAUDE: Store team display info in DB and fix lineup auto-start
Backend:
- Add game_metadata to create_game() and quick_create_game() endpoints
- Fetch team display info (lname, sname, abbrev, color, thumbnail) from
  SBA API at game creation time and store in DB
- Populate GameState with team display fields from game_metadata
- Fix submit_team_lineup to cache lineup in state_manager after DB write
  so auto-start correctly detects both teams ready

Frontend:
- Read team colors/names/thumbnails from gameState instead of useState
- Remove useState approach that failed across SSR navigation
- Fix create.vue redirect from legacy /games/lineup/[id] to /games/[id]
- Update game.vue header to show team names from gameState

Docs:
- Update CLAUDE.md to note dev mode has broken auth, always use prod

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
2026-01-17 08:43:26 -06:00

285 lines
9.5 KiB
Markdown

# 🚨 CRITICAL: @ MENTION HANDLING 🚨
When ANY file is mentioned with @ syntax, you MUST IMMEDIATELY call Read tool on that file BEFORE responding.
You will see automatic loads of any @ mentioned filed, this is NOT ENOUGH, it only loads the file contents.
You MUST perform Read tool calls on the files directly, even if they were @ included.
This is NOT optional - it loads required CLAUDE.md context. along the file path.
See @./.claude/force-claude-reads.md for details.
---
# Paper Dynasty Real-Time Game Engine - Development Guide
## Project Overview
Web-based real-time multiplayer baseball simulation platform replacing legacy Google Sheets system. Consists of:
- **Shared Backend**: FastAPI (Python 3.11+) with WebSocket support, PostgreSQL persistence
- **Dual Frontends**: Separate Vue 3/Nuxt 3 apps per league (SBA and PD) with shared component library
**Critical Business Driver**: Legacy Google Sheets being deprecated - this is mission-critical replacement.
## Architecture Principles
### Backend Philosophy
- **Hybrid State Model**: In-memory game state for performance + PostgreSQL for persistence/recovery
- **League-Agnostic Core**: Polymorphic player models, config-driven league variations
- **Async-First**: All I/O operations use async/await patterns
- **Type Safety**: Pydantic models for validation, SQLAlchemy for ORM
### Frontend Philosophy
- **Mobile-First**: Primary design target is mobile portrait mode
- **Real-Time Updates**: WebSocket (Socket.io) for all game state changes
- **Shared Components**: Maximize reuse between league frontends
- **Type Safety**: TypeScript with strict mode
## Technology Stack
### Backend
- FastAPI + Socket.io (WebSocket)
- PostgreSQL 14+ with SQLAlchemy 2.0
- Pydantic for data validation
- pytest for testing
### Frontend (Per League)
- Vue 3 Composition API + Nuxt 3
- TypeScript (strict mode)
- Tailwind CSS
- Pinia for state management
- Socket.io-client
- Discord OAuth via HttpOnly cookies (see `COOKIE_AUTH_IMPLEMENTATION.md`)
## Key Technical Patterns
### Polymorphic Player Architecture
Use factory pattern for league-specific player types:
- `BasePlayer` (abstract base)
- `SbaPlayer` (simple model)
- `PdPlayer` (detailed scouting data)
- `Lineup.from_api_data(config, data)` factory method
### WebSocket Event Flow
1. Player action → WebSocket → Backend
2. Validate against in-memory state
3. Process + resolve outcome
4. Update in-memory state
5. Async write to PostgreSQL
6. Broadcast state update to all clients
### Game State Recovery
On reconnect: Load plays from DB → Replay to rebuild state → Send current state
## Project Structure
```
strat-gameplay-webapp/
├── backend/ # FastAPI game engine
│ ├── app/
│ │ ├── core/ # Game engine, dice, state management
│ │ ├── config/ # League configs and result charts
│ │ ├── websocket/ # Socket.io handlers
│ │ ├── models/ # Pydantic + SQLAlchemy models
│ │ └── api/ # REST endpoints
│ └── tests/
├── frontend-sba/ # SBA League Nuxt app
├── frontend-pd/ # PD League Nuxt app (disabled)
├── docker-compose.yml # Base service configuration
├── docker-compose.dev.yml # Development overrides (hot-reload)
├── docker-compose.prod.yml # Production overrides (optimized)
├── start.sh # Single-command startup script
└── scripts/env-switch.sh # Environment profile switcher
```
## Quick Start (All-Docker Workflow)
The entire stack runs in Docker with a single command. No local Python or Node.js required.
> **⚠️ ALWAYS USE PROD MODE**: Discord OAuth does not work in dev mode due to cookie/CORS configuration. Since this system isn't live yet, always build and run with `prod` mode for testing.
```bash
# Production (optimized build) - USE THIS
./start.sh prod
# Development - DO NOT USE (auth broken)
# ./start.sh dev
# Stop all services
./start.sh stop
# View logs
./start.sh logs
# Check status
./start.sh status
# Force rebuild
./start.sh rebuild [dev|prod]
```
### What Each Mode Does
| Mode | Backend | Frontend | Use Case |
|------|---------|----------|----------|
| `prod` | Production build | SSR optimized build | **Always use this** - auth works correctly |
| `dev` | Hot-reload (uvicorn --reload) | Hot-reload (nuxt dev) | ❌ Auth broken - do not use |
### Service URLs
| Service | Dev Mode | Prod Mode |
|---------|----------|-----------|
| Frontend | http://localhost:3000 | https://gameplay-demo.manticorum.com |
| Backend API | http://localhost:8000 | https://gameplay-demo.manticorum.com/api |
| API Docs | http://localhost:8000/docs | https://gameplay-demo.manticorum.com/api/docs |
### Docker Compose Architecture
The stack uses layered compose files:
- **`docker-compose.yml`** - Base services (Redis, Backend, Frontend-SBA)
- **`docker-compose.dev.yml`** - Development overrides (volume mounts, hot-reload)
- **`docker-compose.prod.yml`** - Production overrides (optimized builds, restart policies)
The `start.sh` script handles composing these correctly.
## Environment Configuration
### Environment Profiles
| Profile | Backend URL | Frontend URL |
|---------|-------------|--------------|
| `dev` | `http://localhost:8000` | `http://localhost:3000` |
| `prod` | `https://gameplay-demo.manticorum.com` | `https://gameplay-demo.manticorum.com` |
### Environment Files
| File | Purpose |
|------|---------|
| `backend/.env` | Active backend config (gitignored) |
| `backend/.env.dev` | Local development settings |
| `backend/.env.prod` | Production settings |
| `frontend-sba/.env` | Active frontend config (gitignored) |
| `frontend-sba/.env.dev` | Local development settings |
| `frontend-sba/.env.prod` | Production settings |
### Key Environment Variables
**Backend** (in `backend/.env`):
- `DATABASE_URL` - PostgreSQL connection string
- `DISCORD_CLIENT_ID/SECRET` - OAuth credentials
- `DISCORD_SERVER_REDIRECT_URI` - Server-side OAuth callback
- `FRONTEND_URL` - Frontend base URL for redirects
- `CORS_ORIGINS` - Allowed origins (JSON array)
- `ALLOWED_DISCORD_IDS` - User whitelist (comma-separated, empty = all)
**Frontend** (in `frontend-*/.env`):
- `NUXT_PUBLIC_API_URL` - Backend API URL (public, client-side)
- `NUXT_API_URL_INTERNAL` - Backend URL for SSR (Docker internal: `http://backend:8000`)
- `NUXT_PUBLIC_WS_URL` - WebSocket URL
- `NUXT_PUBLIC_DISCORD_CLIENT_ID` - OAuth client ID (public)
- `NUXT_PUBLIC_DISCORD_REDIRECT_URI` - OAuth callback URL
### Discord OAuth Setup
Both environments require redirect URIs in [Discord Developer Portal](https://discord.com/developers/applications):
| Environment | Redirect URI |
|-------------|--------------|
| dev | `http://localhost:8000/api/auth/discord/callback/server` |
| prod | `https://gameplay-demo.manticorum.com/api/auth/discord/callback/server` |
### Manual Environment Switching
If you need to switch between dev/prod configs without Docker:
```bash
./scripts/env-switch.sh dev # Copy .env.dev → .env
./scripts/env-switch.sh prod # Copy .env.prod → .env
```
The `start.sh` script handles this automatically based on mode.
## Development Guidelines
### Code Quality
- **Python**: Dataclasses preferred, rotating loggers with `f'{__name__}.<className>'`
- **Error Handling**: "Raise or Return" pattern - no Optional unless required
- **Testing**: Run tests freely without asking permission
- **Imports**: Always verify imports during code review to prevent NameErrors
- **Git Commits**: Prefix with "CLAUDE: "
### Performance Targets
- Action response: < 500ms
- WebSocket delivery: < 200ms
- DB writes: < 100ms (async)
- State recovery: < 2 seconds
### Security Requirements
- Discord OAuth for authentication
- Server-side game logic only (zero client trust)
- Cryptographically secure dice rolls
- All rules enforced server-side
## Phase 1 MVP Scope (Weeks 1-13)
**Core Deliverables**:
1. Authentication (Discord OAuth)
2. Game creation & lobby
3. Complete turn-based gameplay with all strategic decisions
4. Real-time WebSocket updates
5. Game persistence & recovery
6. Spectator mode
7. Mobile-optimized UI
8. AI opponent support
**Explicitly Out of Scope for MVP**:
- Roster management
- Marketplace
- Tournaments
- Advanced analytics
## Critical References
- **Full PRD**: `/mnt/NV2/Development/strat-gameplay-webapp/prd-web-scorecard-1.1.md`
- **Player Model Architecture**: PRD lines 378-551
- **Database Schema**: PRD lines 553-628
- **WebSocket Events**: PRD lines 630-669
- **League Config System**: PRD lines 780-846
## League Differences
### SBA League
- Minimal player data (id, name, image)
- Simpler rules configuration
### PD League
- Detailed scouting data on players
- Complex probability calculations
- Additional analytics requirements
## Success Metrics
- 90% player migration within 1 month
- 99.5% uptime
- < 500ms average action latency
- 60%+ mobile usage
- Zero data corruption
---
## Current Implementation Status
**Phase 3E-Final**: **COMPLETE** (2025-01-10)
Backend is production-ready for frontend integration:
- All 15 WebSocket event handlers implemented
- Strategic decisions (defensive/offensive)
- Manual outcome workflow (dice rolling + card reading)
- Player substitutions (3 types)
- Box score statistics (materialized views)
- Position ratings integration (PD league)
- 730/731 tests passing (99.9%)
**Next Phase**: Vue 3 + Nuxt 3 frontend implementation with Socket.io client
---
**Note**: Subdirectories have their own CLAUDE.md files with implementation-specific details to minimize context usage.