#!/usr/bin/env python3
"""扫描 training_data 目录下所有 CSV，按 bot 类型统计胜率"""
import sys, os, glob, re
from collections import Counter, defaultdict

def scan(csv_dir='training_data'):
    csvs = sorted(glob.glob(os.path.join(csv_dir, 'game_*.csv')))
    if not csvs:
        print("No CSV files found")
        return
    
    total = 0
    per_bot = defaultdict(lambda: {'wins': 0, 'games': 0})
    per_game = []  # (game_id, winner_pos, bots_str)
    
    for fp in csvs:
        with open(fp) as f:
            for line in f:
                if line.startswith('#RESULT'):
                    parts = line.strip().split(',')
                    winner = int(parts[1].split('=')[1])
                    bots = parts[2].split('=')[1].split(',')
                    game_id = os.path.splitext(os.path.basename(fp))[0]
                    per_game.append((game_id, winner, line.strip()))
                    total += 1
                    
                    # winner_pos is 1-indexed, bots[winner-1]
                    for i, bt in enumerate(bots):
                        per_bot[bt]['games'] += 1
                    winner_bt = bots[winner - 1] if 0 < winner <= len(bots) else '?'
                    per_bot[winner_bt]['wins'] += 1
                    break
    
    if total == 0:
        print("No #RESULT lines found. Run with updated hjha-console.")
        return
    
    print(f"{total} games scanned")
    print(f"{'Bot':<12} {'Wins':>6} {'Rate':>8} {'Games':>6}")
    print("-" * 35)
    for bt in sorted(per_bot.keys()):
        d = per_bot[bt]
        rate = d['wins'] / d['games'] * 100 if d['games'] > 0 else 0
        print(f"{bt:<12} {d['wins']:>6} {rate:>7.1f}% {d['games']:>6}")
    
    # 也统计按 bot 组合
    combo_stats = defaultdict(lambda: {'total': 0, 'winners': Counter()})
    for gid, winner, line in per_game:
        m = re.search(r'bots=([^,]+(?:,[^,]+)*)', line)
        if m:
            key = m.group(1)
            combo_stats[key]['total'] += 1
            combo_stats[key]['winners'][winner] += 1
    
    if len(combo_stats) > 0:
        print(f"\nBy bot composition:")
        for combo in sorted(combo_stats.keys()):
            d = combo_stats[combo]
            bots = combo.split(',')
            print(f"  [{combo}] — {d['total']} games:")
            for pos, count in sorted(d['winners'].items()):
                bt = bots[pos-1] if 0 < pos <= len(bots) else '?'
                print(f"    pos{pos}({bt}): {count}/{d['total']} ({count/d['total']*100:.0f}%)")

if __name__ == '__main__':
    d = sys.argv[1] if len(sys.argv) > 1 else 'training_data'
    # 如果指定了目录不存在, 可能是编号的round目录
    if not os.path.isdir(d):
        # 尝试 training_rounds 下最新
        rounds = sorted(glob.glob('training_rounds/r*'))
        if rounds:
            d = os.path.join(rounds[-1], 'training_data')
            if not os.path.isdir(d):
                print(f"No training_data in {rounds[-1]}")
                sys.exit(1)
        else:
            print("No data found")
            sys.exit(1)
    scan(d)