meowgorithm/dotfiles

"""
Validator for Word document XML files against XSD schemas.
"""

import random
import re
import tempfile
import zipfile

import defusedxml.minidom
import lxml.etree

from .base import BaseSchemaValidator


class DOCXSchemaValidator(BaseSchemaValidator):

    WORD_2006_NAMESPACE = "http://schemas.openxmlformats.org/wordprocessingml/2006/main"
    W14_NAMESPACE = "http://schemas.microsoft.com/office/word/2010/wordml"
    W16CID_NAMESPACE = "http://schemas.microsoft.com/office/word/2016/wordml/cid"

    ELEMENT_RELATIONSHIP_TYPES = {}

    def validate(self):
        if not self.validate_xml():
            return False

        all_valid = True
        if not self.validate_namespaces():
            all_valid = False

        if not self.validate_unique_ids():
            all_valid = False

        if not self.validate_file_references():
            all_valid = False

        if not self.validate_content_types():
            all_valid = False