# written by Stephen Vermeulen, http://vermeulen.ca/one-time-pad.html
# placed in the public domain in 2006
#
# make a one time pad out of the contents of a bunch of
# files.  Process is:
#
# 1. do a directory listing (of files greater than 8k)
# 2. select 128 files at random from at least 256 files
# 3. from each file select a 2k chunk of data, starting at
#    a random position in the file
# 4. concat all the chunks of data into a list
# 5. shuffle the list
# 6. MD5 digest the list (initializing the digest with some
#    user entered random text), take 32 bytes off the list and
#    feed it to MD5, producing 16 bytes of output and save that
#    repeating until 128k of output is made

import random
import md5
import os
import os.path
import sys
import binascii

def dirlist(path, flist):
    d = os.listdir(path)
    for x in d:
        p = os.path.join(path, x)
        if os.path.isfile(p):
            if os.path.getsize(p) > 8196:
                flist.append(p)
        elif os.path.isdir(p):
            # recurse into this directory
            dirlist(p, flist)

files = []
if len(sys.argv[1:]) > 0:
    for d in sys.argv[1:]:
        dirlist(d, files)
else:
    print "Make a one time pad"
    print "Syntax is:"
    print "  makeotp.py dir1 [dir2 ...]"
    print "where dir1 (etc.) are directories of files with at least 256 files"
    print "that are 8K or larger.  If successful will create a file called 'otp.txt'"
    print "in the current directory"
    sys.exit(0)

if len(files) < 256:
    print "Did not find enough suitable files, only found:", len(files)
    print "please run again with a different list of directories"
    sys.exit(0)

salt = ""
while len(salt) < 10:
    salt += raw_input( "Enter some random text and hit return (10 char min): " )
    
print "Total eligible files in the directories:", len(files)
sfiles = random.sample(files, 128)
data = []
for f in sfiles:
    n = os.path.getsize(f) - 2048
    fin = open(f, "rb")
    fin.seek(random.randint(0, n))
    d = fin.read(2048)
    fin.close()
    data = data + list(d)

random.shuffle(data)

m = md5.new()
m.update(salt)
fout = open("otp.txt", "w")
i = 0
while i < len(data):
    for j in range(32):
        m.update(data[i + j])
    z1 = m.digest()
    for j in range(32):
        m.update(data[i + j])
    z2 = m.digest()
    fout.write("%04d: %s%s\n" % (i/64, binascii.hexlify(z1), binascii.hexlify(z2)))
    i += 64

fout.close()
print "Wrote output to: otp.txt"