From fa3112ae58fb10af894e7a5e84e22bafb3801268 Mon Sep 17 00:00:00 2001 From: TKE Date: Tue, 22 Dec 2020 11:51:52 +0100 Subject: [PATCH] Add filetype to md5it --- md5it.py | 25 +++++++++++++++++++++---- 1 file changed, 21 insertions(+), 4 deletions(-) diff --git a/md5it.py b/md5it.py index d7093d5..d904b6b 100755 --- a/md5it.py +++ b/md5it.py @@ -1,13 +1,30 @@ #!/usr/bin/env python3 import os import hashlib -BLOCKSIZE=65536 -files = [f for f in os.listdir('.') if os.path.isfile(f)] -for f in files: +import magic + +extensions={ + "message_rfc822":"eml", + "text_html":"html" + } + +def get_md5(filename): md5=hashlib.md5() with open(f,'rb') as fd: fb=fd.read(BLOCKSIZE) while len(fb)>0: md5.update(fb) fb=fd.read(BLOCKSIZE) - os.rename(f,md5.hexdigest()) + return md5.hexdigest() + +def get_magic(filename): + magic_string=magic.from_file(filename, mime=True) + magic_string=magic_string.replace("/","_") + if magic_string in extensions: + magic_string=extensions[magic_string] + return magic_string + +BLOCKSIZE=65536 +files = [f for f in os.listdir('.') if os.path.isfile(f)] +for f in files: + os.rename(f,get_md5(f)+"."+get_magic(f))