ttfFromDocx

# extract .ttf files from .docx
# [email protected]

import base64
import sys
import os, zipfile
from BeautifulSoup import BeautifulStoneSoup as soup

def extractFontKey(files):
    keys = []

    for filename in files:
        if not filename.endswith("fontTable.xml"):
            continue
        xmlStr = open(filename).read()
        xml = soup(xmlStr)
        # print xml
        for i, f in enumerate(xml.findAll('w:font')):
            name = f['w:name']
            x = soup(str(f))

            try:
                fontkey =  x.findAll('w:embedregular')[0]['w:fontkey']
                fontkey = "".join(fontkey[1:-1].split("-"))
            except: # some entry does not have fontkey. skip this
                continue

            print name, fontkey
            keys.append( (name, fontkey))
        return keys

def deobf( fontKeyBin, obfFontStr):
    fontStr = [ord(x) for x in obfFontStr]

    for i in range(16):
        fontStr[i] = ord(obfFontStr[i]) ^ ord(fontKeyBin[15-i])
        fontStr[i+16] = ord(obfFontStr[i+16]) ^ ord(fontKeyBin[15-i])

    fontStr = "".join( [ "%c" % x for x in fontStr])
    return fontStr

def deobfAll(files, keys):
    files.sort() # odttf filenames seem to be sorted as it was appeared in .xml

    i = 0
    for filename in files:
        if not filename.endswith(".odttf"):
            continue
        print filename
        obfStr = open( filename, 'rb').read()
        (fontname, fontkey) = keys[i]
        print fontkey

        deobfStr = deobf(fontkey.decode("hex"), obfStr)

        open( fontname + ".ttf", 'wb').write(deobfStr)
        i += 1

def decompress(docxFilename):
    z = zipfile.ZipFile(docxFilename)

    try:
        os.mkdir( "./tmp")
    except:
        pass

    z.extractall("./tmp")
    files = ["./tmp/" + x for x in z.namelist()]
    return files

if __name__ == '__main__':
    docxFilename = sys.argv[1]

    files = decompress( docxFilename)
    keys = extractFontKey(files)
    deobfAll(files, keys)