TroyHow
/

Helsinki-NLP

Model card Files Files and versions Community

Helsinki-NLP / check_required_files.py

TroyHow's picture

Upload 82 files

77307ef verified 2 months ago

raw history blame contribute delete

No virus

2.99 kB

	import os
	import zipfile

	# Initialize counters
	total_files_count = 0
	contains_all_files_count = 0
	missing_files_count = 0
	missing_files_info = [] # Used to store information about missing files

	def check_zip_contents(zip_file_path):
	global total_files_count
	global contains_all_files_count
	global missing_files_count
	global missing_files_info

	# Increase the total count of ZIP files
	total_files_count += 1

	try:
	# Your existing code for processing the zip file contents
	contains_all_files_count = 0 # Initialize contains_all_files_count here
	with zipfile.ZipFile(zip_file_path, 'r') as zf:
	# Get all files and folders in the ZIP file
	file_list = [os.path.normpath(item) for item in zf.namelist()]

	# Get the ZIP file name (excluding extension)
	zip_file_name = os.path.splitext(os.path.basename(zip_file_path))[0]

	# Build the expected list of files and folders
	expected_folder = os.path.normpath(zip_file_name)
	expected_files = ['config.json', 'generation_config.json', 'pytorch_model.bin', 'source.spm', 'target.spm', 'tokenizer_config.json', 'vocab.json']

	# Check if the folder exists
	if expected_folder not in file_list:
	missing_files_count += 1
	missing_files_info.append(f"{zip_file_name} does not contain the expected folder.\n")
	return

	# Check if individual files exist
	missing_files = []
	for expected_file in expected_files:
	file_path = os.path.join(expected_folder, expected_file)
	if file_path not in file_list:
	missing_files.append(expected_file)

	if not missing_files:
	contains_all_files_count += 1
	else:
	missing_files_count += 1
	missing_files_info.append(f"{zip_file_name} is missing the following files: {', '.join(missing_files)}\n")

	except zipfile.BadZipFile as e:
	print(f"Error: {e}")
	# Additional information about the exception, like file path
	print(f"File path: {zip_file_path}")

	# The rest of your code remains unchanged

	# Use the current working directory as the folder path
	folder_path = os.getcwd()

	# Get all ZIP files in the folder
	zip_files = [f for f in os.listdir(folder_path) if f.endswith('.zip')]

	# Iterate over ZIP files and check their contents
	for zip_file in zip_files:
	zip_file_path = os.path.join(folder_path, zip_file)
	check_zip_contents(zip_file_path)

	# Display the count of ZIP files containing all files and folders
	print(f"\nNumber of ZIP files containing all files and folders: {contains_all_files_count}")

	# Display information about missing files
	for info in missing_files_info:
	print(info)

	# Display the count results
	print(f"\nProcessed {total_files_count} ZIP files")
	print(f"Number of files missing: {missing_files_count}")