Recursively compare files in two folders

compare-folders.py

#Recursively compare files in two folders
#by Tim Sheerman-Chase, 2014
#Released under the CC0 license

import sys, os, hashlib

def CheckFolderPair(fo1, fo2, stats):
	fiList1 = os.listdir(fo1)

	for fina in fiList1:
		finaFull1 = fo1+"/"+fina
		finaFull2 = fo2+"/"+fina

		#Ignore certain types of system folders on NAS box
		if fina in ['#recycle', '@eaDir']:
			continue

		print finaFull1, stats
		if os.path.isfile(finaFull1):
			if os.path.isfile(finaFull2):
				size1 = os.stat(finaFull1).st_size
				size2 = os.stat(finaFull2).st_size
				
				if size1 != size2:
					stats['wrong-size'] += 1
				else:
					#Compare hashes
					#print fina
					ha1 = hashlib.md5()
					ha1.update(open(finaFull1, "r").read())

					ha2 = hashlib.md5()
					ha2.update(open(finaFull2, "r").read())

					if ha1.digest() != ha2.digest():
						print [ord(ch) for ch in ha1.digest()]
						print [ord(ch) for ch in ha2.digest()]
						stats['wrong-hash'] += 1
					
			else:
				stats['missing-file2'] += 1	

		if os.path.isdir(finaFull1):
			if os.path.isdir(finaFull2):
				CheckFolderPair(finaFull1, finaFull2, stats)
			else:
				stats['missing-dir2'] += 1
			

if __name__ == "__main__":
	if len(sys.argv) < 3:
		print "Usage: {0} folder1 folder2".format(sys.argv[0])
		exit(0)

	stats = {}
	stats['missing-file2'] = 0
	stats['missing-dir2'] = 0
	stats['wrong-size'] = 0
	stats['wrong-hash'] = 0

	CheckFolderPair(sys.argv[1], sys.argv[2], stats)