vyndr/nba-service/app/services/stats.py

import time
from datetime import datetime, timedelta, timezone

from nba_api.stats.endpoints import playercareerstats, playergamelog
from nba_api.stats.library.parameters import SeasonAll

from app.utils.cache import cache_get, cache_set
from app.utils.player_map import resolve_player
from app.config import (
    SEASON_AVG_TTL, LAST_N_TTL, SPLITS_TTL, NBA_API_DELAY,
    NBA_API_RETRY_DELAY, NBA_API_TIMEOUT,
)

# Map nba_api columns to our internal stat names
STAT_MAP = {
    "PTS": "points",
    "REB": "rebounds",
    "AST": "assists",
    "FG3M": "threes",
    "BLK": "blocks",
    "STL": "steals",
    "TOV": "turnovers",
    "MIN": "minutes",
    "GP": "games_played",
}


def get_current_season():
    """Return current NBA season string (e.g., '2025-26'). Season starts in October."""
    now = datetime.now(timezone.utc)
    year = now.year if now.month >= 10 else now.year - 1
    return f"{year}-{str(year + 1)[-2:]}"


def _call_nba_api(fn, **kwargs):
    """Call nba_api with rate limiting and retry."""
    time.sleep(NBA_API_DELAY)
    try:
        return fn(**kwargs, timeout=NBA_API_TIMEOUT)
    except Exception:
        time.sleep(NBA_API_RETRY_DELAY)
        return fn(**kwargs, timeout=NBA_API_TIMEOUT)


def _map_stats(row):
    """Convert a single nba_api stats row to our internal format."""
    stats = {}
    for nba_col, our_name in STAT_MAP.items():
        val = row.get(nba_col)
        if val is not None:
            stats[our_name] = round(float(val), 1)
    # Compute PRA
    pts = stats.get("points", 0)
    reb = stats.get("rebounds", 0)
    ast = stats.get("assists", 0)
    stats["pra"] = round(pts + reb + ast, 1)
    return stats


def _extract_team(career_data, season):
    """Extract team abbreviation from career stats for given season."""
    rows = career_data.get_data_frames()[0]
    season_row = rows[rows["SEASON_ID"] == season]
    if not season_row.empty:
        return season_row.iloc[0]["TEAM_ABBREVIATION"]
    if not rows.empty:
        return rows.iloc[-1]["TEAM_ABBREVIATION"]
    return "UNK"


def get_season_avg(player_name, stat_type=None, season=None):
    """Get a player's season averages."""
    player_id, full_name = resolve_player(player_name)
    if player_id is None:
        return None

    if season is None:
        season = get_current_season()

    cache_key = f"nba:season:{player_id}:{season}"
    cached = cache_get(cache_key)
    if cached is not None:
        result = cached
        result["source"] = "cache"
        if stat_type and stat_type in result["stats"]:
            result["stats"] = {stat_type: result["stats"][stat_type]}
        return result

    career = _call_nba_api(playercareerstats.PlayerCareerStats, player_id=player_id)
    df = career.get_data_frames()[0]
    season_row = df[df["SEASON_ID"] == season]

    if season_row.empty:
        return {
            "player": full_name,
            "player_id": player_id,
            "team": "UNK",
            "season": season,
            "source": "live",
            "stats": {},
        }

    row = season_row.iloc[0].to_dict()
    team = row.get("TEAM_ABBREVIATION", "UNK")
    stats = _map_stats(row)

    result = {
        "player": full_name,
        "player_id": player_id,
        "team": team,
        "season": season,
        "source": "live",
        "stats": stats,
    }

    cache_set(cache_key, result, SEASON_AVG_TTL)

    if stat_type and stat_type in result["stats"]:
        result_filtered = dict(result)
        result_filtered["stats"] = {stat_type: result["stats"][stat_type]}
        return result_filtered

    return result


def get_last_n(player_name, n=10, stat_type=None):
    """Get a player's averages over their last N games."""
    player_id, full_name = resolve_player(player_name)
    if player_id is None:
        return None

    n = min(max(n, 1), 30)

    cache_key = f"nba:last:{player_id}:{n}"
    cached = cache_get(cache_key)
    if cached is not None:
        result = cached
        result["source"] = "cache"
        if stat_type and stat_type in result["stats"]:
            result["stats"] = {stat_type: result["stats"][stat_type]}
        return result

    season = get_current_season()
    gamelog = _call_nba_api(
        playergamelog.PlayerGameLog,
        player_id=player_id,
        season=season,
    )
    df = gamelog.get_data_frames()[0]

    if df.empty:
        return {
            "player": full_name,
            "player_id": player_id,
            "team": "UNK",
            "last_n": n,
            "source": "live",
            "stats": {},
        }

    last_n_df = df.head(n)
    team = last_n_df.iloc[0].get("TEAM_ABBREVIATION", "UNK") if not last_n_df.empty else "UNK"

    # Compute averages
    avg_row = {}
    for col in STAT_MAP:
        if col in last_n_df.columns:
            avg_row[col] = last_n_df[col].mean()
    avg_row["GP"] = len(last_n_df)

    stats = _map_stats(avg_row)

    result = {
        "player": full_name,
        "player_id": player_id,
        "team": team,
        "last_n": n,
        "source": "live",
        "stats": stats,
    }

    cache_set(cache_key, result, LAST_N_TTL)

    if stat_type and stat_type in result["stats"]:
        result_filtered = dict(result)
        result_filtered["stats"] = {stat_type: result["stats"][stat_type]}
        return result_filtered

    return result


def get_splits(player_name, stat_type, split_type, opponent=None):
    """Get situational splits for a player."""
    player_id, full_name = resolve_player(player_name)
    if player_id is None:
        return None

    cache_key = f"nba:splits:{player_id}:{stat_type}:{split_type}"
    if opponent:
        cache_key += f":{opponent}"

    cached = cache_get(cache_key)
    if cached is not None:
        cached["source"] = "cache"
        return cached

    season = get_current_season()
    gamelog = _call_nba_api(
        playergamelog.PlayerGameLog,
        player_id=player_id,
        season=season,
    )
    df = gamelog.get_data_frames()[0]

    if df.empty:
        return {
            "player": full_name,
            "player_id": player_id,
            "stat_type": stat_type,
            "split_type": split_type,
            "source": "live",
            "splits": {},
        }

    # Map stat_type to nba_api column
    reverse_map = {v: k for k, v in STAT_MAP.items()}
    if stat_type == "pra":
        nba_cols = ["PTS", "REB", "AST"]
    else:
        nba_col = reverse_map.get(stat_type)
        if nba_col is None or nba_col not in df.columns:
            return None
        nba_cols = [nba_col]

    def avg_stat(subset):
        if subset.empty:
            return 0
        if stat_type == "pra":
            return round((subset["PTS"] + subset["REB"] + subset["AST"]).mean(), 1)
        return round(subset[nba_cols[0]].mean(), 1)

    team = df.iloc[0].get("TEAM_ABBREVIATION", "UNK") if not df.empty else "UNK"

    if split_type == "home_away":
        # MATCHUP contains "vs." for home games, "@" for away
        home = df[df["MATCHUP"].str.contains("vs.", na=False)]
        away = df[df["MATCHUP"].str.contains("@", na=False)]
        splits = {
            "home": {"avg": avg_stat(home), "games": len(home)},
            "away": {"avg": avg_stat(away), "games": len(away)},
        }

    elif split_type == "rest_days":
        df = df.copy()
        df["GAME_DATE_PARSED"] = df["GAME_DATE"].apply(_parse_game_date)
        df = df.sort_values("GAME_DATE_PARSED")

        b2b = []
        one_day = []
        two_plus = []

        dates = df["GAME_DATE_PARSED"].tolist()
        for i, row_idx in enumerate(df.index):
            if i == 0:
                two_plus.append(row_idx)
                continue
            delta = (dates[i] - dates[i - 1]).days
            if delta <= 1:
                b2b.append(row_idx)
            elif delta == 2:
                one_day.append(row_idx)
            else:
                two_plus.append(row_idx)

        splits = {
            "b2b": {"avg": avg_stat(df.loc[b2b]) if b2b else 0, "games": len(b2b)},
            "1_day_rest": {"avg": avg_stat(df.loc[one_day]) if one_day else 0, "games": len(one_day)},
            "2_plus_days_rest": {"avg": avg_stat(df.loc[two_plus]) if two_plus else 0, "games": len(two_plus)},
        }

    elif split_type == "vs_team":
        if not opponent:
            return None
        opponent_upper = opponent.upper()
        vs_opp = df[df["MATCHUP"].str.contains(opponent_upper, na=False)]
        vs_others = df[~df["MATCHUP"].str.contains(opponent_upper, na=False)]
        splits = {
            "vs_opponent": {"avg": avg_stat(vs_opp), "games": len(vs_opp)},
            "vs_all_others": {"avg": avg_stat(vs_others), "games": len(vs_others)},
        }

    else:
        return None

    result = {
        "player": full_name,
        "player_id": player_id,
        "team": team,
        "stat_type": stat_type,
        "split_type": split_type,
        "source": "live",
        "splits": splits,
    }

    if opponent:
        result["opponent"] = opponent

    cache_set(cache_key, result, SPLITS_TTL)
    return result


def _parse_game_date(date_str):
    """Parse game date from nba_api format. Handles 'MAR 21, 2026' and similar."""
    for fmt in ("%b %d, %Y", "%Y-%m-%d", "%m/%d/%Y"):
        try:
            return datetime.strptime(date_str, fmt)
        except (ValueError, TypeError):
            continue
    return datetime.now(timezone.utc)