TheAMM

Searching for duplicate SWFs with Python

Jul 4th, 2014
265
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
Python 0.81 KB | None | 0 0
  1. #!/usr/bin/python
  2. import os
  3. from sys import argv
  4. from glob import glob
  5. from hashlib import md5 as MD5
  6.  
  7. targetDirectory = "." if len(argv) < 2 else argv[1]
  8.  
  9. foundFiles = {}
  10. for f in glob(os.path.join(targetDirectory, "*.swf")):
  11.     if not os.path.isfile(f): continue
  12.  
  13.     fsize = os.path.getsize(f)
  14.    
  15.     if fsize in foundFiles:
  16.         foundFiles[fsize].append(f)
  17.     else:
  18.         foundFiles[fsize] = [f]
  19.  
  20. duplicateFiles = {}
  21. for size, files in foundFiles.items():
  22.     if len(files) > 1:
  23.         for f in files:
  24.             data = open(f, "r").read()
  25.             md5sum = MD5(data).hexdigest()
  26.  
  27.             if md5sum in duplicateFiles:
  28.                 duplicateFiles[md5sum].append(f)
  29.             else:
  30.                 duplicateFiles[md5sum] = [f]
  31.  
  32. for md5sum, files in duplicateFiles.items():
  33.     if len(files) > 1:
  34.         print "Duplicates for MD5:", md5sum
  35.         for f in files:
  36.             print "  ", f
Advertisement
Add Comment
Please, Sign In to add comment