# import pandas as pd
# train_file = '/home/bingxing2/ailab/group/ai4bio/public/multi-omics/multi-omics/downstream/EnhancerPromoter/train.csv'
# val_file = '/home/bingxing2/ailab/group/ai4bio/public/multi-omics/multi-omics/downstream/EnhancerPromoter/val.csv'
# test_file = '/home/bingxing2/ailab/group/ai4bio/public/multi-omics/multi-omics/downstream/EnhancerPromoter/test.csv'

# # File paths
# file_paths = [train_file, val_file, test_file]


# def calculate_combined_statistics(file_paths, column_name):
#     """ Calculate max and mean lengths of the specified column across multiple CSV files """
#     combined_df = pd.concat([pd.read_csv(file_path) for file_path in file_paths], ignore_index=True)
#     # Drop rows with NaN values in the specified column
#     combined_df = combined_df[combined_df[column_name].notna()]
#     # Compute lengths of the sequences
#     lengths = combined_df[column_name].apply(len)
#     return lengths.max(), lengths.mean()


# # Calculate statistics for siRNA_sense_seq
# print("enhancer statistics:")
# max_len, mean_len = calculate_combined_statistics(file_paths, 'enhancer')
# print(f"Max length = {max_len}, Mean length = {mean_len:.2f}")

# # Calculate statistics for gene_target_seq
# print("\npromoter statistics:")
# max_len, mean_len = calculate_combined_statistics(file_paths, 'promoter')
# print(f"Max length = {max_len}, Mean length = {mean_len:.2f}")


# import matplotlib.pyplot as plt

# def calculate_lengths(file_paths, column_name):
#     """ Calculate lengths of the specified column across multiple CSV files """
#     lengths = []
#     for file_path in file_paths:
#         df = pd.read_csv(file_path)
#         # Drop rows with NaN values in the specified column
#         df = df[df[column_name].notna()]
#         # Compute lengths of the sequences
#         lengths.extend(df[column_name].apply(len))
#     return lengths

# # File paths


# # Calculate lengths for siRNA_sense_seq and gene_target_seq
# siRNA_lengths = calculate_lengths(file_paths, 'enhancer')
# gene_target_lengths = calculate_lengths(file_paths, 'promoter')

# # Create histograms
# plt.figure(figsize=(14, 7))

# # Plot for siRNA_sense_seq lengths
# plt.subplot(1, 2, 1)
# plt.hist(siRNA_lengths, bins=30, color='skyblue', edgecolor='black')
# plt.title('enhancer Length Distribution')
# plt.xlabel('Length')
# plt.ylabel('Frequency')

# # Plot for gene_target_seq lengths
# plt.subplot(1, 2, 2)
# plt.hist(gene_target_lengths, bins=30, color='lightgreen', edgecolor='black')
# plt.title('promoter Length Distribution')
# plt.xlabel('Length')
# plt.ylabel('Frequency')

# # Save the figure
# plt.tight_layout()
# plt.savefig('length_distributions.png')



import pandas as pd

# 定义 CSV 文件路径
csv_files = ['train.csv',
 'val.csv', 
 'test.csv']

# 初始化统计字典
label_stats = {'train': {0: 0, 1: 0}, 'val': {0: 0, 1: 0}, 'test': {0: 0, 1: 0}}

# 统计每个文件中的 label 数量
for file in csv_files:
    # 读取 CSV 文件
    df = pd.read_csv(file)
    
    # 提取文件名（去掉 .csv 后缀）
    file_name = file.split('.')[0]
    
    # 统计 label 列中的 0 和 1 的数量
    label_counts = df['label'].value_counts()
    
    # 更新统计字典
    for label in label_counts.index:
        label_stats[file_name][label] = label_counts[label]

# 打印统计结果
for dataset, counts in label_stats.items():
    print(f"{dataset}.csv")
    print(f"Label 0: {counts[0]}")
    print(f"Label 1: {counts[1]}")
    print()