nsfw/split_file.py

import os
import sys

def split_file(input_file, n):
    # 檢查檔案是否存在
    if not os.path.isfile(input_file):
        print(f"File {input_file} does not exist.")
        return
    
    # 讀取檔案內容
    with open(input_file, 'r', encoding='utf-8') as f:
        lines = f.readlines()
    
    # 計算需要多少個輸出檔案
    total_lines = len(lines)
    num_files = (total_lines + n - 1) // n  # 向上取整
    
    # 獲取檔案名稱和副檔名
    base_name, ext = os.path.splitext(input_file)
    
    for i in range(num_files):
        # 計算這個檔案應包含的行範圍
        start_line = i * n
        end_line = min(start_line + n, total_lines)
        
        # 創建新的檔案名稱
        output_file = f"{base_name}_{i+1:04d}{ext}"
        
        # 寫入對應行到新檔案
        with open(output_file, 'w', encoding='utf-8') as f:
            f.writelines(lines[start_line:end_line])
        
        print(f"Created {output_file} with lines from {start_line} to {end_line-1}")

if __name__ == "__main__":
    if len(sys.argv) != 3:
        print("Usage: python split_file.py <input_file> <lines_per_file>")
        sys.exit(1)

    input_file = sys.argv[1]
    try:
        lines_per_file = int(sys.argv[2])
    except ValueError:
        print("The number of lines per file must be an integer.")
        sys.exit(1)

    split_file(input_file, lines_per_file)