Sessions 5-7a: 955 tests, deployment ready

2026-06-08 18:35:13 -04:00
parent 06b82624a2
commit 1fa04dc776
371 changed files with 49366 additions and 955 deletions
@@ -0,0 +1,105 @@
+"""
+seed_historical.py — One-time historical data seeder for VYNDR.
+Run ONCE before launch to backfill coaching and player-out data.
+
+Usage:
+    python scripts/seed_historical.py
+"""
+
+import sys
+import os
+import time
+import logging
+
+# Allow imports from src/services/python
+sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..', 'src', 'services', 'python'))
+
+from coaching_parser import parse_nba_coaching_from_game_id, parse_mlb_coaching_from_game_id
+from player_outs import find_and_log_historical_player_outs
+
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s [%(levelname)s] %(message)s'
+)
+logger = logging.getLogger(__name__)
+
+
+def seed_nba_coaching(season='2024-25'):
+    """Seed NBA coaching data from LeagueGameLog for a full season."""
+    from nba_api.stats.endpoints import LeagueGameLog
+
+    logger.info(f"Fetching NBA game log for season {season}...")
+    game_log = LeagueGameLog(season=season, season_type_all_star='Regular Season')
+    df = game_log.get_data_frames()[0]
+
+    game_ids = df['GAME_ID'].unique()
+    total = len(game_ids)
+    logger.info(f"Found {total} unique NBA games to process.")
+
+    for i, game_id in enumerate(game_ids, start=1):
+        try:
+            parse_nba_coaching_from_game_id(game_id)
+        except Exception as e:
+            logger.error(f"NBA game {game_id} failed: {e}")
+
+        if i % 50 == 0:
+            logger.info(f"NBA coaching progress: {i}/{total} games processed")
+
+        time.sleep(0.6)
+
+    logger.info(f"NBA coaching seed complete. {total} games processed.")
+
+
+def seed_mlb_coaching(season=2024):
+    """Seed MLB coaching data from statsapi schedule for a full season."""
+    import statsapi
+
+    start_date = f'{season}-03-28'
+    end_date = f'{season}-09-29'
+
+    logger.info(f"Fetching MLB schedule for {start_date} to {end_date}...")
+    schedule = statsapi.schedule(start_date=start_date, end_date=end_date)
+
+    game_ids = [g['game_id'] for g in schedule]
+    total = len(game_ids)
+    logger.info(f"Found {total} MLB games to process.")
+
+    for i, game_id in enumerate(game_ids, start=1):
+        try:
+            parse_mlb_coaching_from_game_id(game_id)
+        except Exception as e:
+            logger.error(f"MLB game {game_id} failed: {e}")
+
+        if i % 100 == 0:
+            logger.info(f"MLB coaching progress: {i}/{total} games processed")
+
+        time.sleep(0.3)
+
+    logger.info(f"MLB coaching seed complete. {total} games processed.")
+
+
+def seed_player_out_history(season='2024-25'):
+    """Seed historical player-out data for a given season."""
+    logger.info(f"Seeding player-out history for season {season}...")
+    try:
+        find_and_log_historical_player_outs(season=season)
+        logger.info("Player-out history seed complete.")
+    except Exception as e:
+        logger.error(f"Player-out history seed failed: {e}")
+        raise
+
+
+if __name__ == '__main__':
+    logger.info("=== VYNDR Historical Data Seeder ===")
+    logger.info("This should be run ONCE before launch.\n")
+
+    logger.info("--- Step 1/3: NBA Coaching ---")
+    seed_nba_coaching(season='2024-25')
+
+    logger.info("--- Step 2/3: MLB Coaching ---")
+    seed_mlb_coaching(season=2024)
+
+    logger.info("--- Step 3/3: Player-Out History ---")
+    seed_player_out_history(season='2024-25')
+
+    logger.info("=== All historical seeds complete. ===")