Check a list of files against a list of openssl checksums (sha1 and md5)

check_sum.py

"""
File: check_sum.py
Author: Brant Faircloth

Created by Brant Faircloth on 23 February 2012 10:02 PST (-0800)
Copyright (c) 2012 Brant C. Faircloth. All rights reserved.

Description: Check a list of files against hashes.

License:  http://www.opensource.org/licenses/BSD-3-Clause

On sending end, generate sha1 hashes like so:

    for i in *.gz; do openssl sha1 $i >> file-with-sha1-checksums.txt; done

Send that file, along with files being shipped to recipient.  On receving end
recipient validates with:

    python check_sum.py file-with-sha1-checksums.txt
        /path/to/directory/to/validate

"""

import os
import sys
import re
import glob
import hashlib
import argparse


def get_args():
    parser = argparse.ArgumentParser(description='''Validate checksums given a
            list''')
    parser.add_argument('checksums', nargs='?', type=argparse.FileType('r'),
            default=sys.stdin)
    parser.add_argument('dir', nargs='?', default=sys.stdout)
    parser.add_argument('--md5', action='store_true', default = False)
    return parser.parse_args()

def hashfile(afile, hasher, blocksize=65536):
    """ slightly modified from @Omnifarious in stackexchange thread:
    
    http://bit.ly/erSQcW
    
    """
    buf = afile.read(blocksize)
    while len(buf) > 0:
        hasher.update(buf)
        buf = afile.read(blocksize)
    return hasher.hexdigest()


def main():
    args = get_args()
    cs = {}
    if not args.md5:
        regex = re.compile("SHA1\((.*)\)=\s(.*)")
    else:
        regex = re.compile("MD5\((.*)\)=\s(.*)")
    for line in args.checksums:
        ls = line.strip()
        match = regex.search(ls)
        cs[match.groups()[0]] = match.groups()[1]
    print "{0:30} {1:4}".format('Filename', 'Status')
    print '-' * 37
    for fname in glob.glob(os.path.join(args.dir, '*.*')):
        if args.checksums.name in fname:
            pass
        else:
            if not args.md5:
                hsh = hashfile(file(fname, 'r'), hashlib.sha1())
            else:
                hsh = hashfile(file(fname, 'r'), hashlib.md5())
            assert cs[os.path.basename(fname)] == hsh, "File sha1 hashes do not match"""
            
            print "{0:30} {1:4}".format(os.path.basename(fname), '[ok]')


if __name__ == '__main__':
    main()