strat-gameplay-webapp/backend/app/core/state_manager.py
Cal Corum e2f1d6079f CLAUDE: Implement Week 7 Task 6 - PlayResolver Integration with RunnerAdvancement
Major Refactor: Outcome-First Architecture
- PlayResolver now accepts league_id and auto_mode in constructor
- Added core resolve_outcome() method - all resolution logic in one place
- Added resolve_manual_play() wrapper for manual submissions (primary)
- Added resolve_auto_play() wrapper for PD auto mode (rare)
- Removed SimplifiedResultChart (obsolete with new architecture)
- Removed play_resolver singleton

RunnerAdvancement Integration:
- All groundball outcomes (GROUNDBALL_A/B/C) now use RunnerAdvancement
- Proper DP probability calculation with positioning modifiers
- Hit location tracked for all relevant outcomes
- 13 result types fully integrated from advancement charts

Game State Updates:
- Added auto_mode field to GameState (stored per-game)
- Updated state_manager.create_game() to accept auto_mode parameter
- GameEngine now uses state.auto_mode to create appropriate resolver

League Configuration:
- Added supports_auto_mode() to BaseGameConfig
- SbaConfig: returns False (no digitized cards)
- PdConfig: returns True (has digitized ratings)
- PlayResolver validates auto mode support and raises error for SBA

Play Results:
- Added hit_location field to PlayResult
- Groundballs include location from RunnerAdvancement
- Flyouts track hit_location for tag-up logic (future)
- Other outcomes have hit_location=None

Testing:
- Completely rewrote test_play_resolver.py for new architecture
- 9 new tests covering initialization, strikeouts, walks, groundballs, home runs
- All 9 tests passing
- All 180 core tests still passing (1 pre-existing failure unrelated)

Terminal Client:
- No changes needed - defaults to manual mode (auto_mode=False)
- Perfect for human testing of manual submissions

This completes Week 7 Task 6 - the final task of Week 7!
Week 7 is now 100% complete with all 8 tasks done.

🎯 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2025-10-31 08:20:52 -05:00

539 lines
18 KiB
Python

"""
State Manager - In-memory game state management.
Manages active game states in memory for fast gameplay (<500ms response time).
Provides CRUD operations, lineup management, and state recovery from database.
This is the single source of truth for active game states during gameplay.
Author: Claude
Date: 2025-10-22
"""
import asyncio
import logging
from typing import Dict, Optional, Union
from uuid import UUID
import pendulum
from app.models.game_models import GameState, TeamLineupState, DefensiveDecision, OffensiveDecision
from app.database.operations import DatabaseOperations
logger = logging.getLogger(f'{__name__}.StateManager')
class StateManager:
"""
Manages in-memory game states for active games.
Responsibilities:
- Store game states in memory for fast access
- Manage team lineups per game
- Track last access times for eviction
- Recover game states from database on demand
This class uses dictionaries for O(1) lookups of game state by game_id.
"""
def __init__(self):
"""Initialize the state manager with empty storage"""
self._states: Dict[UUID, GameState] = {}
self._lineups: Dict[UUID, Dict[int, TeamLineupState]] = {} # game_id -> {team_id: lineup}
self._last_access: Dict[UUID, pendulum.DateTime] = {}
# Phase 3: Decision queue for async decision awaiting
# Key: (game_id, team_id, decision_type)
self._pending_decisions: Dict[tuple[UUID, int, str], asyncio.Future] = {}
self.db_ops = DatabaseOperations()
logger.info("StateManager initialized")
async def create_game(
self,
game_id: UUID,
league_id: str,
home_team_id: int,
away_team_id: int,
home_team_is_ai: bool = False,
away_team_is_ai: bool = False,
auto_mode: bool = False
) -> GameState:
"""
Create a new game state in memory.
Args:
game_id: Unique game identifier
league_id: League identifier ('sba' or 'pd')
home_team_id: Home team ID
away_team_id: Away team ID
home_team_is_ai: Whether home team is AI-controlled
away_team_is_ai: Whether away team is AI-controlled
auto_mode: True = auto-generate outcomes (PD only), False = manual submissions
Returns:
Newly created GameState
Raises:
ValueError: If game_id already exists
"""
if game_id in self._states:
raise ValueError(f"Game {game_id} already exists in state manager")
logger.info(f"Creating game state for {game_id} ({league_id} league, auto_mode={auto_mode})")
state = GameState(
game_id=game_id,
league_id=league_id,
home_team_id=home_team_id,
away_team_id=away_team_id,
home_team_is_ai=home_team_is_ai,
away_team_is_ai=away_team_is_ai,
auto_mode=auto_mode
)
self._states[game_id] = state
self._lineups[game_id] = {}
self._last_access[game_id] = pendulum.now('UTC')
logger.debug(f"Game {game_id} created in memory")
return state
def get_state(self, game_id: UUID) -> Optional[GameState]:
"""
Get game state by ID.
Updates last access time when accessed.
Args:
game_id: Game identifier
Returns:
GameState if found, None otherwise
"""
if game_id in self._states:
self._last_access[game_id] = pendulum.now('UTC')
return self._states[game_id]
return None
def update_state(self, game_id: UUID, state: GameState) -> None:
"""
Update game state.
Args:
game_id: Game identifier
state: Updated GameState
Raises:
ValueError: If game_id doesn't exist
"""
if game_id not in self._states:
raise ValueError(f"Game {game_id} not found in state manager")
self._states[game_id] = state
self._last_access[game_id] = pendulum.now('UTC')
logger.debug(f"Updated state for game {game_id} (inning {state.inning}, {state.half})")
def set_lineup(self, game_id: UUID, team_id: int, lineup: TeamLineupState) -> None:
"""
Set team lineup for a game.
Args:
game_id: Game identifier
team_id: Team identifier
lineup: Team lineup state
Raises:
ValueError: If game_id doesn't exist
"""
if game_id not in self._states:
raise ValueError(f"Game {game_id} not found in state manager")
if game_id not in self._lineups:
self._lineups[game_id] = {}
self._lineups[game_id][team_id] = lineup
logger.info(f"Set lineup for team {team_id} in game {game_id} ({len(lineup.players)} players)")
def get_lineup(self, game_id: UUID, team_id: int) -> Optional[TeamLineupState]:
"""
Get team lineup for a game.
Args:
game_id: Game identifier
team_id: Team identifier
Returns:
TeamLineupState if found, None otherwise
"""
return self._lineups.get(game_id, {}).get(team_id)
def remove_game(self, game_id: UUID) -> None:
"""
Remove game from memory.
Call this when a game is completed or being archived.
Args:
game_id: Game identifier
"""
removed_parts = []
if game_id in self._states:
self._states.pop(game_id)
removed_parts.append("state")
if game_id in self._lineups:
self._lineups.pop(game_id)
removed_parts.append("lineups")
if game_id in self._last_access:
self._last_access.pop(game_id)
removed_parts.append("access")
if removed_parts:
logger.info(f"Removed game {game_id} from memory ({', '.join(removed_parts)})")
else:
logger.warning(f"Attempted to remove game {game_id} but it was not in memory")
async def recover_game(self, game_id: UUID) -> Optional[GameState]:
"""
Recover game state from database.
This is called when a game needs to be loaded (e.g., after server restart,
or when a game is accessed that's not currently in memory).
Loads game data from database and rebuilds the in-memory state.
Args:
game_id: Game identifier
Returns:
Recovered GameState if found in database, None otherwise
"""
logger.info(f"Recovering game {game_id} from database")
# Load from database
game_data = await self.db_ops.load_game_state(game_id)
if not game_data:
logger.warning(f"Game {game_id} not found in database")
return None
# Rebuild state from loaded data
state = await self._rebuild_state_from_data(game_data)
# Cache in memory
self._states[game_id] = state
self._last_access[game_id] = pendulum.now('UTC')
logger.info(f"Recovered game {game_id} - inning {state.inning}, {state.half}")
return state
async def _rebuild_state_from_data(self, game_data: dict) -> GameState:
"""
Rebuild game state from database data using the last completed play.
This method recovers the complete game state without replaying all plays.
It uses the final positions from the last play to reconstruct runners and
batter indices.
Args:
game_data: Dictionary with 'game', 'lineups', and 'plays' keys
Returns:
Reconstructed GameState
"""
game = game_data['game']
state = GameState(
game_id=game['id'],
league_id=game['league_id'],
home_team_id=game['home_team_id'],
away_team_id=game['away_team_id'],
home_team_is_ai=game.get('home_team_is_ai', False),
away_team_is_ai=game.get('away_team_is_ai', False),
status=game['status'],
inning=game.get('current_inning', 1),
half=game.get('current_half', 'top'),
home_score=game.get('home_score', 0),
away_score=game.get('away_score', 0),
play_count=len(game_data.get('plays', []))
)
# Get last completed play to recover runner state and batter indices
plays = game_data.get('plays', [])
if plays:
# Sort by play_number desc and get last completed play
completed_plays = [p for p in plays if p.get('complete', False)]
if completed_plays:
last_play = max(completed_plays, key=lambda p: p['play_number'])
# Recover runner state from final positions
from app.models.game_models import RunnerState
runners = []
# Check each base for a runner (using *_final fields)
for base_num, final_field in [(1, 'on_first_final'), (2, 'on_second_final'), (3, 'on_third_final')]:
final_base = last_play.get(final_field)
if final_base == base_num: # Runner ended on this base
# Get lineup_id from corresponding on_X_id field
lineup_id = last_play.get(f'on_{["", "first", "second", "third"][base_num]}_id')
if lineup_id:
runners.append(RunnerState(
lineup_id=lineup_id,
card_id=0, # Will be populated when needed
on_base=base_num
))
# Check if batter reached base
batter_final = last_play.get('batter_final')
if batter_final and 1 <= batter_final <= 3:
batter_id = last_play.get('batter_id')
if batter_id:
runners.append(RunnerState(
lineup_id=batter_id,
card_id=0,
on_base=batter_final
))
state.runners = runners
# Recover batter indices from lineups
# We need to find where each team is in their batting order
home_lineup = [l for l in game_data.get('lineups', []) if l['team_id'] == state.home_team_id]
away_lineup = [l for l in game_data.get('lineups', []) if l['team_id'] == state.away_team_id]
# For now, we'll need to be called with _prepare_next_play() after recovery
# to set the proper batter indices and snapshot
# Initialize to 0 - will be corrected by _prepare_next_play()
state.away_team_batter_idx = 0
state.home_team_batter_idx = 0
logger.debug(
f"Recovered state from play {last_play['play_number']}: "
f"{len(runners)} runners on base"
)
else:
logger.debug("No completed plays found - initializing fresh state")
else:
logger.debug("No plays found - initializing fresh state")
# Count runners on base
runners_on_base = len(state.get_all_runners())
logger.info(f"Rebuilt state for game {state.game_id}: {state.play_count} plays, {runners_on_base} runners")
return state
def evict_idle_games(self, idle_minutes: int = 60) -> int:
"""
Remove games that haven't been accessed recently.
This helps manage memory by removing inactive games. Evicted games
can be recovered from database if needed later.
Args:
idle_minutes: Minutes of inactivity before eviction (default 60)
Returns:
Number of games evicted
"""
cutoff = pendulum.now('UTC').subtract(minutes=idle_minutes)
to_evict = [
game_id for game_id, last_access in self._last_access.items()
if last_access < cutoff
]
for game_id in to_evict:
self.remove_game(game_id)
if to_evict:
logger.info(f"Evicted {len(to_evict)} idle games (idle > {idle_minutes}m)")
return len(to_evict)
def get_stats(self) -> dict:
"""
Get state manager statistics.
Returns:
Dictionary with current state statistics:
- active_games: Number of games in memory
- total_lineups: Total lineups across all games
- games_by_league: Count of games per league
- games_by_status: Count of games by status
"""
stats = {
"active_games": len(self._states),
"total_lineups": sum(len(lineups) for lineups in self._lineups.values()),
"games_by_league": {},
"games_by_status": {},
}
# Count by league
for state in self._states.values():
league = state.league_id
stats["games_by_league"][league] = stats["games_by_league"].get(league, 0) + 1
# Count by status
for state in self._states.values():
status = state.status
stats["games_by_status"][status] = stats["games_by_status"].get(status, 0) + 1
return stats
def exists(self, game_id: UUID) -> bool:
"""
Check if game exists in memory.
Args:
game_id: Game identifier
Returns:
True if game is in memory, False otherwise
"""
return game_id in self._states
def get_all_game_ids(self) -> list[UUID]:
"""
Get list of all game IDs currently in memory.
Returns:
List of game UUIDs
"""
return list(self._states.keys())
# ============================================================================
# PHASE 3: DECISION QUEUE MANAGEMENT
# ============================================================================
def set_pending_decision(
self,
game_id: UUID,
team_id: int,
decision_type: str
) -> None:
"""
Mark that a decision is required and create a future for it.
Args:
game_id: Game identifier
team_id: Team that needs to make the decision
decision_type: Type of decision ('defensive' or 'offensive')
"""
key = (game_id, team_id, decision_type)
# Create a new future for this decision
self._pending_decisions[key] = asyncio.Future()
logger.debug(f"Set pending {decision_type} decision for game {game_id}, team {team_id}")
async def await_decision(
self,
game_id: UUID,
team_id: int,
decision_type: str
) -> Union[DefensiveDecision, OffensiveDecision]:
"""
Wait for a decision to be submitted.
This coroutine will block until submit_decision() is called
with matching parameters.
Args:
game_id: Game identifier
team_id: Team making the decision
decision_type: Type of decision expected
Returns:
The submitted decision (DefensiveDecision or OffensiveDecision)
Raises:
ValueError: If no pending decision exists for these parameters
asyncio.TimeoutError: If decision not received within timeout (handled by caller)
"""
key = (game_id, team_id, decision_type)
if key not in self._pending_decisions:
raise ValueError(
f"No pending {decision_type} decision for game {game_id}, team {team_id}"
)
# Await the future (will be resolved by submit_decision)
decision = await self._pending_decisions[key]
logger.debug(f"Received {decision_type} decision for game {game_id}, team {team_id}")
return decision
def submit_decision(
self,
game_id: UUID,
team_id: int,
decision: Union[DefensiveDecision, OffensiveDecision]
) -> None:
"""
Submit a decision (called by WebSocket handler or AI opponent).
This resolves the pending future created by set_pending_decision().
Args:
game_id: Game identifier
team_id: Team making the decision
decision: The decision being submitted
Raises:
ValueError: If no pending decision exists
"""
# Determine decision type from the decision object
from app.models.game_models import DefensiveDecision
decision_type = "defensive" if isinstance(decision, DefensiveDecision) else "offensive"
key = (game_id, team_id, decision_type)
if key not in self._pending_decisions:
raise ValueError(
f"No pending {decision_type} decision for game {game_id}, team {team_id}"
)
future = self._pending_decisions[key]
# Check if already resolved (should not happen)
if future.done():
logger.warning(f"Decision already submitted for {key}")
return
# Resolve the future with the decision
future.set_result(decision)
# Clean up the future
del self._pending_decisions[key]
logger.info(f"Submitted {decision_type} decision for game {game_id}, team {team_id}")
def cancel_pending_decision(
self,
game_id: UUID,
team_id: int,
decision_type: str
) -> None:
"""
Cancel a pending decision (e.g., on timeout or game abort).
Args:
game_id: Game identifier
team_id: Team that was expected to decide
decision_type: Type of decision
"""
key = (game_id, team_id, decision_type)
if key in self._pending_decisions:
future = self._pending_decisions[key]
if not future.done():
future.cancel()
del self._pending_decisions[key]
logger.debug(f"Cancelled pending {decision_type} decision for game {game_id}, team {team_id}")
# Singleton instance for global access
state_manager = StateManager()