dirunify.py

dirunify.py

# Author: Pierce Brooks

import os
import sys
import shutil
import hashlib
import logging
import traceback

def hashify(target):
    sha = hashlib.sha256()
    sha.update(target)
    return str(sha.hexdigest())

directories = sys.argv[1:]
hashes = {}
for directory in directories:
    if not (os.path.exists(directory)):
        continue
    for root, folders, files in os.walk(directory):
        names = list(sorted(files))
        for name in names:
            try:
                path = os.path.join(root, name)
                descriptor = open(path, "rb")
                salt = hashify(descriptor.read())
                descriptor.close()
                if not (salt in hashes):
                    hashes[salt] = path
                else:
                    if (sys.flags.debug):
                        print(path)
            except:
                logging.error(traceback.format_exc())
for key in hashes:
    source = hashes[key]
    destination = os.path.join(os.getcwd(), os.path.basename(source))
    if not (os.path.exists(destination)):
        try:
            shutil.copy2(source, destination)
        except:
            logging.error(traceback.format_exc())