"""
Author: Yagnik Poshiya
GitHub: @yagnikposhiya
Charotar University of Science and Technology
"""

import os
import shutil
import math

def split_dataset(input_dir, output_dir, files_per_directory):
    if not os.path.exists(output_dir):
        os.makedirs(output_dir)

    files = os.listdir(input_dir)
    num_files = len(files)
    num_directories = math.ceil(num_files / files_per_directory)

    for i in range(num_directories):
        start_idx = i * files_per_directory
        end_idx = min((i + 1) * files_per_directory, num_files)
        subset_files = files[start_idx:end_idx]

        directory_name = f"clips_{i}"
        directory_path = os.path.join(output_dir, directory_name)

        os.makedirs(directory_path)
        for file_name in subset_files:
            src_file_path = os.path.join(input_dir, file_name)
            dest_file_path = os.path.join(directory_path, file_name)
            shutil.copy(src_file_path, dest_file_path)

if __name__ == "__main__":
    input_directory = "/home/yagnikposhiya/WorkStation/Datasets/SIH/CommonVoiceCorpusHindi15/clips"  # Change this to your dataset directory
    output_directory = "/home/yagnikposhiya/WorkStation/Datasets/SIH/CommonVoiceCorpusHindi15/clips2"  # Change this to your desired output directory
    files_per_directory = 10000  # Maximum files per directory

    split_dataset(input_directory, output_directory, files_per_directory)