Extracts embedded fonts from a Microsoft Word Document (.docx)

docx2ttf.py

#!/usr/bin/env python3

# Standard Library
import zipfile, xml.etree.ElementTree, shutil, os, sys, pathlib

# Classes and Associations
# [Docx] (1)---(1) [DocxFonts] (1)---(1..n) [DocxFont] (1)---(0..n) [DocxOdttfFont]

class Docx():

    # https://ecma-international.org/publications-and-standards/standards/ecma-376/
    # Part 1 “Fundamentals And Markup Language Reference”, 5th edition, December 2016
    # 11. WordprocessingML

    def __init__(self, file_name):
        self.file_name = file_name
        self.zipfile = zipfile.PyZipFile(file_name)
        self.yaml_log = f"---\nfile_name: {self.file_name}\n"

    def extract_fonts(self):
        DocxFonts(self).extract()
        return self.yaml_log


class DocxFonts(): # [Docx] (1)---(1) [DocxFonts]

    # [Content_Types].xml
    # <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
    # <Types xmlns="http://schemas.openxmlformats.org/package/2006/content-types">
    #   <Default Extension="odttf" ContentType="application/vnd.openxmlformats-officedocument.obfuscatedFont"/>
    #     ...
    #   <Override PartName="/word/fontTable.xml" ContentType="application/vnd.openxmlformats-officedocument.wordprocessingml.fontTable+xml"/>


    # <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
    # <w:fonts xmlns:mc="http://schemas.openxmlformats.org/markup-compatibility/2006"
    #     xmlns:r="http://schemas.openxmlformats.org/officeDocument/2006/relationships"
    #     xmlns:w="http://schemas.openxmlformats.org/wordprocessingml/2006/main"
    # ...
    #    <w:font w:name="Aptos">

    zipped_xml_table_name = 'word/fontTable.xml'

    r_namspace_uri = 'http://schemas.openxmlformats.org/officeDocument/2006/relationships'
    w_namspace_uri = 'http://schemas.openxmlformats.org/wordprocessingml/2006/main'
    namespaces = {'r': r_namspace_uri, 'w': w_namspace_uri}

    def __init__(self, docx):
        self.docx = docx
        self.xml_table = docx.zipfile.read(self.zipped_xml_table_name)
        self.xml_table = xml.etree.ElementTree.fromstring(self.xml_table)
        self.xml_fonts = self.xml_table.findall('.//w:font', DocxFonts.namespaces)

    def extract(self):
        self.extracted_directory = pathlib.Path(f"{self.docx.file_name}.fonts")
        shutil.rmtree(self.extracted_directory, ignore_errors = True)
        os.makedirs(self.extracted_directory)
        self.docx.yaml_log += f"fonts:\n"
        for xml_font in self.xml_fonts:
            DocxFont(self, xml_font).extract()

class DocxFont(): # [DocxFonts] (1)---(1..n) [DocxFont]

    # https://ecma-international.org/publications-and-standards/standards/ecma-376/
    # Part 1 “Fundamentals And Markup Language Reference”, 5th edition, December 2016
    # 17. WordprocessingML Reference Material
    # 17.8 Fonts
    # 17.8.1 Font Embedding

    # <w:font w:name="Times New Roman">
    #   <w:panose1 w:val="02020603050405020304"/>
    #   <w:charset w:val="00"/>
    #   <w:family w:val="roman"/>
    #   <w:pitch w:val="variable"/>
    #   <w:sig w:usb0="E0002AFF" w:usb1="C0007843" w:usb2="00000009" w:usb3="00000000" w:csb0="000001FF" w:csb1="00000000"/>
    # </w:font>
    # <w:font w:name="Aptos">
    #   <w:charset w:val="00"/>
    #   <w:family w:val="swiss"/>
    #   <w:pitch w:val="variable"/>
    #   <w:sig w:usb0="20000287" w:usb1="00000003" w:usb2="00000000" w:usb3="00000000" w:csb0="0000019F" w:csb1="00000000"/>
    #   <w:embedRegular r:id="rId1" w:fontKey="{695C1753-A322-488E-9BAB-90BBBA6B2839}"/>
    #   <w:embedItalic r:id="rId2" w:fontKey="{5CB96690-C171-4250-B3CC-DF32ED694182}"/>
    # </w:font>

    def __init__(self, docx_fonts, xml_font):
        self.docx_fonts = docx_fonts
        self.xml_font = xml_font
        self.name = self.xml_font.attrib[f"{{{ DocxFonts.w_namspace_uri}}}name"]
        self.xml_odttf_fonts = self.xml_font.findall('.//*[@w:fontKey]', DocxFonts.namespaces)
        self.docx_fonts.docx.yaml_log += f"  - font: {self.name}\n"

    def extract(self):
        for xml_odttf_font in self.xml_odttf_fonts:
            DocxOdttfFont(self, xml_odttf_font).extract()

class DocxOdttfFont(): # [DocxFont] (1)---(0..n) [DocxOdttfFont]

    # [Content_Types].xml
    # <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
    # <Types xmlns="http://schemas.openxmlformats.org/package/2006/content-types">
    #     <Default Extension="odttf" ContentType="application/vnd.openxmlformats-officedocument.obfuscatedFont"/>

    extension = 'odttf'

    # word/_rels/fontTable.xml.rels
    # <?xml version="1.0" encoding="UTF-8" standalone="yes"?>
    # <Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">
    #   <Relationship Id="rId8" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/font" Target="fonts/font8.odttf"/>

    zipped_directory = 'word/fonts'
    prefix = 'font'

    # <w:embedItalic r:id="rId2" w:fontKey="{5CB96690-C171-4250-B3CC-DF32ED694182}"/>
    # AptosItalic.ttf

    def __init__(self, docx_font, xml_odttf_font):
        self.docx_font = docx_font
        self.font_name = self.docx_font.name.replace(' ','')
        self.xml_font = xml_odttf_font
        self.style = self.xml_font.tag.replace(f"{{{DocxFonts.w_namspace_uri}}}embed", '')
        self.ttf_file_name = f"{self.font_name}{self.style}.ttf"
        self.id = self.xml_font.attrib[f"{{{DocxFonts.r_namspace_uri}}}id"][3:]
        self.key_hexa = self.xml_font.attrib[f"{{{DocxFonts.w_namspace_uri}}}fontKey"][1:-1].replace('-','')

    def extract(self):
        zipped_odttf_path = f"{self.zipped_directory}/{self.prefix}"
        zipped_odttf_path += f"{self.id}.{self.extension}"
        odttf_data = self.docx_font.docx_fonts.docx.zipfile.read(zipped_odttf_path)
        self.ttf_path =  self.docx_font.docx_fonts.extracted_directory / self.ttf_file_name
        self.ttf_path.write_bytes(odttf_data)
        self.deobfuscate()
        self.docx_font.docx_fonts.docx.yaml_log += f"    ttf:  {self.ttf_file_name}\n"
        
    def deobfuscate(self):
        # https://discuss.python.org/t/xor-operator-between-bytes/17910/13
        obfucated_bytes_count = 32
        key_bytes = bytearray.fromhex(self.key_hexa)
        # 32 bytes big endian int : 16 bytes big endian + 16 bytes big endian
        double_key_int = int.from_bytes(key_bytes * 2, byteorder='big', signed=False)
        ttf_file = self.ttf_path.open('r+b') # rw binary
        obfucated_int = int.from_bytes(ttf_file.read(obfucated_bytes_count), byteorder='little', signed=False)
        deobfucated_int = obfucated_int ^ double_key_int
        ttf_file.seek(0)
        ttf_file.write(deobfucated_int.to_bytes(obfucated_bytes_count, byteorder='little', signed=False))
        ttf_file.close()

if __name__ == '__main__':
    if len(sys.argv) < 2 :
        print(f"""
usage: {sys.argv[0]} first.docx ... last.docx

    first.docx.fonts/FontOneStyle.ttf
    first.docx.fonts/FontTwoStyle.ttf
    ...
    ...
    last.docx.fonts/FontOneStyle.ttf
        """)
        exit()

    # doxc2ttf.py *.docx
    for file in sys.argv[1:]:
        try:
            print(Docx(file).extract_fonts())
        except Exception as exception:
            print(exception)