#!/usr/bin/env python3
"""
检查转换后的音素数据集的结构
"""

import pickle

def check_data_structure(pkl_path):
    """检查PKL文件的数据结构"""
    with open(pkl_path, 'rb') as f:
        data = pickle.load(f)

    print("=== 数据结构分析 ===")
    print(f"数据类型: {type(data)}")
    print(f"顶层键: {list(data.keys())}")

    # 检查phoneme_data
    if 'phoneme_data' in data:
        phoneme_data = data['phoneme_data']
        print(f"\nphoneme_data类型: {type(phoneme_data)}")
        print(f"音素数量: {len(phoneme_data)}")
        print(f"音素列表: {list(phoneme_data.keys())[:10]}...")

        # 检查第一个音素的结构
        first_phoneme = list(phoneme_data.keys())[0]
        segments = phoneme_data[first_phoneme]
        print(f"\n第一个音素 '{first_phoneme}':")
        print(f"  segments类型: {type(segments)}")
        print(f"  segments数量: {len(segments)}")

        if len(segments) > 0:
            first_segment = segments[0]
            print(f"  第一个segment类型: {type(first_segment)}")
            print(f"  第一个segment键: {list(first_segment.keys())}")

            # 显示segment的详细内容
            print(f"  第一个segment内容:")
            for key, value in first_segment.items():
                if key == 'original_timestamps':
                    print(f"    {key}: {type(value)}")
                    if isinstance(value, dict):
                        for ts_key, ts_value in value.items():
                            print(f"      {ts_key}: {ts_value}")
                else:
                    print(f"    {key}: {value}")

    # 检查conversion_info
    if 'conversion_info' in data:
        conversion_info = data['conversion_info']
        print(f"\nconversion_info:")
        for key, value in conversion_info.items():
            print(f"  {key}: {value}")

if __name__ == "__main__":
    pkl_path = "../phoneme_segmented_data/phoneme_dataset_20251009_202457_with_original_timestamps.pkl"
    check_data_structure(pkl_path)