simulate_with_plot.py

"""
抽卡模拟器 —— 图表输出版
用法: python simulate_with_plot.py [垫抽数] [-n 次数] [--pool 卡池] [--guaranteed] [-o 文件]
"""
import argparse
import platform
import sys
import time

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.ticker as mticker

from core import run_simulation, POOL_CONFIGS

# ── 中文字体设置 ────────────────────────────────────────────
_OS = platform.system()
if _OS == 'Darwin':
    plt.rcParams['font.sans-serif'] = ['PingFang HK', 'PingFang SC', 'Arial Unicode MS', 'Heiti TC']
elif _OS == 'Windows':
    plt.rcParams['font.sans-serif'] = ['SimHei', 'Microsoft YaHei']
else:
    plt.rcParams['font.sans-serif'] = ['WenQuanYi Micro Hei', 'Droid Sans Fallback', 'Noto Sans CJK SC']
plt.rcParams['axes.unicode_minus'] = False


def plot_results(stats: dict, output_file: str) -> None:
    cfg      = stats['cfg']
    n        = stats['num_simulations']
    pity     = stats['initial_pity']
    pulls    = stats['pulls']
    max_pull = stats['max_pull']
    spark    = cfg['spark']

    counts  = np.bincount(pulls, minlength=max_pull + 1)
    cum_pct = np.cumsum(counts) / n * 100
    x       = np.arange(1, max_pull + 1)

    guaranteed_note = '  ⚑ 翻转保证已激活' if stats['initial_guaranteed'] else ''
    fig, axs = plt.subplots(2, 2, figsize=(14, 11))
    fig.suptitle(
        f'{cfg["display_name"]}  抽卡模拟结果'
        f'  (垫抽: {pity}  |  次数: {n:,}{guaranteed_note})',
        fontsize=13, fontweight='bold', y=0.99,
    )
    ax1, ax2 = axs[0, 0], axs[0, 1]
    ax3, ax4 = axs[1, 0], axs[1, 1]

    # ── 图1：频率分布直方图 ───────────────────────────────────
    ax1.bar(x, counts[1:], color='steelblue', edgecolor='steelblue', alpha=0.7, width=1.0)
    ax1.axvline(stats['mean_pulls'],   color='red',       linestyle='--', linewidth=1.5,
                label=f"均值: {stats['mean_pulls']:.1f}")
    ax1.axvline(stats['median_pulls'], color='limegreen', linestyle='--', linewidth=1.5,
                label=f"中位数: {stats['median_pulls']:.0f}")
    ax1.set_title('抽数频率分布')
    ax1.set_xlabel('所需抽数')
    ax1.set_ylabel('模拟次数')
    ax1.set_xlim(0, max_pull + 2)
    ax1.legend(fontsize=9)
    ax1.yaxis.set_major_formatter(mticker.FuncFormatter(lambda v, _: f'{int(v):,}'))

    # ── 图2：累积概率曲线 (CDF) ───────────────────────────────
    ax2.plot(x, cum_pct[1:], color='darkorange', linewidth=2)
    ax2.set_title('累积出货概率 (CDF)')
    ax2.set_xlabel('所需抽数')
    ax2.set_ylabel('累积概率 (%)')
    ax2.set_xlim(0, max_pull + 2)
    ax2.set_ylim(0, 105)
    ax2.grid(True, linestyle='--', alpha=0.5)

    for p, v in stats['percentiles'].items():
        if p in (25, 50, 75, 90):
            ax2.plot(v, p, 'ro', markersize=5, zorder=5)
            offset_y = -8 if p > 75 else 5
            ax2.annotate(f'{p}%: {v}抽', xy=(v, p),
                         xytext=(v + 2, p + offset_y), fontsize=8,
                         arrowprops=dict(arrowstyle='->', color='black', lw=0.8))

    if spark > 0 and spark <= len(cum_pct):
        spark_pct_val = 100 - cum_pct[spark - 1]
        ax2.text(max_pull * 0.65, 8,
                 f'大保底 ({spark}抽)\n触发率: {spark_pct_val:.2f}%',
                 fontsize=8, bbox=dict(facecolor='tomato', alpha=0.15, boxstyle='round'))

    # ── 图3：各6★出货时机区间图 ──────────────────────────────
    timing_stats = stats['timing_stats']
    n_positions  = len(timing_stats)
    bar_colors   = ['#4fc3f7', '#81c784', '#ffb74d', '#e57373', '#ce93d8', '#80cbc4', '#ffcc80']

    for i, (n_star, ts) in enumerate(timing_stats.items()):
        color = bar_colors[i % len(bar_colors)]
        y = n_positions - 1 - i

        ax3.barh(y, ts['p90'] - ts['p10'], left=ts['p10'],
                 height=0.35, color=color, alpha=0.3)
        ax3.barh(y, ts['p75'] - ts['p25'], left=ts['p25'],
                 height=0.55, color=color, alpha=0.85)
        ax3.plot([ts['mean'], ts['mean']], [y - 0.35, y + 0.35],
                 color='white', linewidth=2.5, zorder=5)
        ax3.plot([ts['mean'], ts['mean']], [y - 0.35, y + 0.35],
                 color='#c62828', linewidth=1.5, zorder=6)
        label_x = min(ts['p90'] + 2, max_pull + 2)
        ax3.text(label_x, y,
                 f" 均值 {ts['mean']:.0f}抽\n 常见 {ts['p25']:.0f}–{ts['p75']:.0f}抽",
                 va='center', fontsize=8, color='#333333')

    ax3.set_yticks(list(range(n_positions - 1, -1, -1)))
    ax3.set_yticklabels([f'第{n}个6★' for n in timing_stats.keys()], fontsize=9)
    ax3.set_xlabel('当前卡池已抽数')
    ax3.set_title('各6★出货时机分布\n(深色=25%-75%常见区间  浅色=10%-90%  竖线=均值)')
    ax3.set_xlim(0, max_pull + 30)
    ax3.set_ylim(-0.8, n_positions - 0.2)
    ax3.grid(True, axis='x', linestyle='--', alpha=0.4)

    if spark > 0:
        ax3.set_xticks(sorted({0, spark // 4, spark // 2, spark * 3 // 4, spark}))
        ax3.axvline(spark, color='tomato', linestyle=':', linewidth=1, alpha=0.6)
    else:
        step = cfg['hard_pity'] // 2
        ax3.set_xticks(list(range(0, max_pull + step, step)))

    # ── 图4：数据摘要文本 ─────────────────────────────────────
    ax4.axis('off')

    spark_line = (f'  大保底触发率  : {stats["spark_rate"]*100:.2f}%\n'
                  if spark > 0 else '')
    spook_line = (f'  平均歪往期限定 : {stats["mean_limited"]:.2f} 次\n'
                  if cfg['limited_spook_rate'] > 0 else '')
    gua_line   = ('  ⚑ 翻转保证已激活\n'
                  if stats['initial_guaranteed'] else '')

    timing_lines = '\n'.join(
        f"  第{pos}个6★  均值{ts['mean']:.0f}抽  "
        f"常见 {ts['p25']:.0f}-{ts['p75']:.0f}抽  "
        f"(10-90%: {ts['p10']:.0f}-{ts['p90']:.0f}抽)"
        for pos, ts in timing_stats.items()
    )

    summary = (
        f"模拟参数\n"
        f"  卡池       : {cfg['display_name']}\n"
        f"  模拟次数   : {n:,}\n"
        f"  初始垫抽   : {pity}\n"
        f"{gua_line}"
        f"\n"
        f"整体统计\n"
        f"  均值抽数   : {stats['mean_pulls']:.2f}\n"
        f"  中位数抽数 : {stats['median_pulls']:.0f}\n"
        f"  标准差     : {stats['stdev_pulls']:.2f}\n"
        f"  平均6星数  : {stats['mean_stars']:.2f}\n"
        f"{spook_line}"
        f"{spark_line}"
        f"\n"
        f"各6★出货时机（均值 / 常见区间）\n"
        f"{timing_lines}"
    )
    ax4.text(0.05, 0.95, summary, fontsize=9.5, va='top',
             transform=ax4.transAxes,
             bbox=dict(facecolor='lightyellow', alpha=0.6, boxstyle='round'))

    plt.tight_layout(rect=[0, 0, 1, 0.97])
    plt.savefig(output_file, dpi=150, bbox_inches='tight')
    print(f"图表已保存: {output_file}")
    plt.close(fig)


def main() -> None:
    pool_choices = list(POOL_CONFIGS.keys())

    parser = argparse.ArgumentParser(
        description='游戏抽卡概率模拟器（图表输出）',
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog=(
            '示例:\n'
            '  python simulate_with_plot.py 30\n'
            '  python simulate_with_plot.py 0 -n 200000 --pool arknights_limited\n'
            '  python simulate_with_plot.py 0 --pool arknights_limited --guaranteed -o result.png\n'
        ),
    )
    parser.add_argument(
        'pity', nargs='?', type=int, default=None,
        help='上一个6星后已垫的抽数（默认交互输入）',
    )
    parser.add_argument('-n', '--simulations', type=int, default=100_000, metavar='N',
                        help='蒙特卡洛模拟次数（默认: 100000）')
    parser.add_argument('--pool', choices=pool_choices, default='default',
                        help=f'卡池选择（默认: default）')
    parser.add_argument('--guaranteed', action='store_true',
                        help='翻转保证已激活（上次6星歪了），仅对 arknights_limited 有效')
    parser.add_argument('-o', '--output', type=str, default=None, metavar='FILE',
                        help='输出图片文件名（默认自动生成）')
    args = parser.parse_args()

    cfg      = POOL_CONFIGS[args.pool]
    max_pity = cfg['hard_pity'] - 1

    pity = args.pity
    if pity is None:
        try:
            raw  = input(f"请输入已垫抽数 (0–{max_pity}，默认为0): ").strip()
            pity = int(raw) if raw else 0
        except (ValueError, EOFError):
            print("输入无效，使用默认值 0")
            pity = 0

    if not (0 <= pity <= max_pity):
        print(f"错误: 垫抽数应在 0 到 {max_pity}之间，当前输入: {pity}", file=sys.stderr)
        sys.exit(1)

    gua_tag     = '_guaranteed' if args.guaranteed else ''
    output_file = args.output or f'result_{args.pool}_pity{pity}{gua_tag}.png'

    print(f"卡池: {cfg['display_name']}  |  模拟次数: {args.simulations:,}")
    t0    = time.perf_counter()
    stats = run_simulation(pity, args.simulations, args.pool, args.guaranteed)
    elapsed = time.perf_counter() - t0
    print(f"模拟完成，耗时 {elapsed:.2f}s")

    plot_results(stats, output_file)

    print(f"\n====== 核心数据 ======")
    print(f"均值: {stats['mean_pulls']:.2f}  中位数: {stats['median_pulls']:.0f}  标准差: {stats['stdev_pulls']:.2f}")
    for p, v in stats['percentiles'].items():
        print(f"  {p:>3}% 的玩家在 {v:>4} 抽内出货")
    if cfg['spark'] > 0:
        print(f"大保底触发率: {stats['spark_rate']*100:.2f}%")


if __name__ == '__main__':
    main()