mht文件图片解析工具（兼容Chrome/Blink）

之前写过一个mht文件的解析工具，不过当时解析的文件都是ie生成的。没有测试过chrome解析的文件。今天在github上看到一个反馈：https://github.com/obaby/mht-image-extractor/issues/1 qq浏览器保存的文件无法提取，chrome保存的文件会直接崩溃。下载附件的文件解析后发现，这两个文件的文件格式与ie的文件格式并不一致，文件头改成了如下的内容：

From: 
Snapshot-Content-Location: https://mp.weixin.qq.com/s?__biz=MzU1NzQ3MTg5OQ==&mid=2247483652&idx=1&sn=a16979f8b088cb60fb63f210536d5288&chksm=fc3400f0cb4389e698a5a3ce1bf6a6ab3ff6f547bb4db409893850b0c502053d1fea40f70fda&sessionid=0&scene=126&subscene=0&clicktime=1599463540&enterid=1599463540&ascene=3&devicetype=android-28&version=27001237&nettype=ctnet&abtest_cookie=AAACAA%3D%3D&lang=zh_CN&exportkey=AUPVIV8Yt1hvPJ2dYKFWhvM%3D&pass_ticket=eTzcuEu%2BGavsf30E3HDErOhtb18ThPDhge008pIBzY7AFq0IuG1LUgojTpufwqUZ&wx_header=1
Subject: =?utf-8?Q?=E6=B1=89=E6=9C=8D=E4=B8=A8=E5=BD=BC=E5=B2=B8=E8=8A=B1=E5=BC=80?=
Date: Sun, 20 Sep 2020 00:50:44 -0000
MIME-Version: 1.0
Content-Type: multipart/related;
    type="text/html";
    boundary="----MultipartBoundary--Bx5ubV1DnfL8hvvsySfZL6MQeLa58tWkfwrQGpothO----"

而ie保存的文件头则是如下格式的：

Content-Type: multipart/related; start=op.mhtml.1267442701515.fe60c16c115c15f9@169.254.195.209; boundary=----------pMKI1vNl6U7UKeGzbfNTyN Content-Location: http://a.10xjw.com/feizhuliu/89905.html
Subject: =?utf-8?Q?=E8=B6=85=E7=BE=8E=E4=B8=9D=E6=8E=A7=E5=A7=90=E5=A6=B9=E8=8A=B1=E7=A7=92=E6=9D=80=E4=BD=A0=E6=B2=A1=E9=97=AE=E9=A2=98[26P]-=2037kxw.com=20-=20=E4=B8=AD=E5=9B=BD=E6=9C=80=E5=A4=A7=E7=9A=84=E8=89=B2=E6=83=85=E5=88=86=E4=BA=AB=E7=BD=91=E7=AB=99?= MIME-Version: 1.0

其实文件的不同不止这两处，在chrome保存的文件中图片信息可能以二进制形式的存在，而不是之前的base64的编码。新的图片内容数据如下：

ie保存的文件，图片内容如下：

由于之前的版本并没有兼容该编码方式，因而即使找到了合适的分隔符依然无法解析图片，由于文件内容包含二进制内容所以只能切换为二进制模式读取。

于是要解决这个问题，可以采用如下的两个办法：

统一使用二进制模式读文件，然后定位边界线。
只对chrome保存的文件进行处理，我选择这种方法，主要是不用在测试第一种方法的兼容性。

关键代码如下：

# -*- coding: utf-8 -*-
"""
@author: obaby
@license: (C) Copyright 2013-2020, obaby@mars.
@contact: root@obaby.org.cn
@link: http://www.obaby.org.cn
        http://www.h4ck.org.cn
        http://www.findu.co
@file: baby_mht_image_extractor.py
@time: 2020/5/22 20:46
@desc:
"""

import base64
import getopt
import os
import quopri
import sys
import hashlib

from pyfiglet import Figlet

current_path = os.path.dirname(os.path.abspath(__file__))
dirname, filename = os.path.split(os.path.abspath(sys.argv[0]))
current_path = dirname
OUT_PATH = os.path.join(current_path, 'out')


def convert_mht_to_list(boundary, html_content):
    return str(html_content).split(boundary)


def convert_mht_to_list_chrome(boundary, html_content):
    return html_content.split(boundary)


def get_boundary(html_content):
    return '--' + str(html_content).split(';')[-1].split('=')[-1]


def get_boundary_chrome(f):
    for i in range(1, 30):
        l = f.readline()
        if 'boundary' in str(l):
            l = l.replace(b'"', b'').replace(b'\r', b'').replace(b'\n', b'').replace(b'\\', b'')
            bb = bytes.decode(l).split('=')
            return bb[-1]
    return ''


def make_dir(floder_name):
    PATH = os.path.join(OUT_PATH, floder_name)
    if not os.path.exists(PATH):
        os.makedirs(PATH)
        os.chdir(PATH)
    return PATH


def save_image_file(image_content, path, file_name):
    try:
        file_path = os.path.join(path, file_name)
        make_dir(path)
        with open(file_path, 'wb') as f:
            f.write(image_content)
            print('[S] 保存图片成功')
        return file_path
    except Exception as e:
        # print(e)
        print('[S] 保存图片失败: ' + str(e))
        return None


def get_content_type(sub_content):
    content_type = 'Unknown'
    for l in sub_content:
        if 'Content-Type' in l:
            content_type = l.split(';')[0].split(':')[1]
            break
    return content_type


def get_content_encoding(sub_content):
    content_encoding = 'unknown'
    pass_count = 0
    for l in sub_content:
        if 'Content-Transfer-Encoding' in l:
            content_encoding = l.split(':')[1].replace(' ', '')
            break
        pass_count += 1
    return content_encoding, pass_count


def get_content_type_and_content(line, sub_path_name, index):
    line = str(line)
    sub_content = line.split('\n')
    if 'Content-Disposition' in line:
        try:
            file_name = sub_content[0].split(';')[1].split('=')[1]
            if 'filename*0' in sub_content[0]:
                file_name = 'default.jpg'
        except:
            file_name = 'default.jpg'

        content_type = get_content_type(sub_content)
        content_encoding, psc = get_content_encoding(sub_content)
        content = ''.join(sub_content[psc + 1:])

        if 'image' in content_type:
            print('_'*100)
            filename = str(index) + '_' + file_name
            print('[S] 正在保存图片文件:', filename)
            decoded_body = None
            if content_encoding.lower() == 'quoted-printable':
                decoded_body = quopri.decodestring(content)
            if content_encoding.lower() == 'base64':
                decoded_body = base64.b64decode(content)
            if decoded_body:
                save_image_file(decoded_body, sub_path_name, filename)
            else:
                print('[S] 图片解码失败，无法保存')
    return


def print_usage():
    print('*' * 100)
    # f = Figlet(font='slant')
    f = Figlet()
    print(f.renderText('obaby@mars'))
    print('mht image extractor by obaby')
    print('Verson: 0.9.22')
    print('baby_mht_image_extractor -f  -o  -p ')
    print('Need Arguments:')
    print('\t -f ')
    print('\t -o  ')
    print('Option Arguments:')
    print('\t -p ')
    print('Blog: http://www.h4ck.org.cn')
    print('*' * 100)


def save_mht_all_images(input_path):
    sub_path_name = os.path.join(OUT_PATH, os.path.basename(input_path).title())
    with open(input_path, 'r', encoding='utf8') as f:
        first_line = f.readline()
        body_content = f.read()
        boundary = get_boundary(first_line)
        content_list = convert_mht_to_list(boundary, html_content=body_content)
        index = 0
        for l in content_list:
            get_content_type_and_content(l, sub_path_name, index)
            index += 1


def save_mht_all_images_chrome(input_path):
    sub_path_name = os.path.join(OUT_PATH, os.path.basename(input_path).title())
    with open(input_path, 'rb') as f:
        boundary = get_boundary_chrome(f)
        content_type = content_location = content_transfer_encoding = ''
        content = b''
        for line in f:
            # print(line)
            if str(boundary) in str(line):
                # 结束当前读取
                if 'image' in content_type:
                    print('-' * 150)
                    print('[S] Content Type：', content_type)
                    print('[S] Content Transfer Encoding:', content_transfer_encoding)
                    # print(content)
                    print('[S] Content Location：', content_location)

                    image_name = hashlib.md5(bytes(content_location, encoding='utf8')).hexdigest()
                    file_ext = '.' + str(content_type).split('/')[-1]
                    filename = image_name + file_ext
                    print('[I] 图片文件名: ', filename)
                    decoded_body = None
                    if 'quoted-printable' in content_transfer_encoding:
                        decoded_body = quopri.decodestring(content)
                    if 'base64' in content_transfer_encoding:
                        decoded_body = base64.b64decode(content)
                    if 'binary' in content_transfer_encoding:
                        decoded_body = content
                    if decoded_body:
                        save_image_file(decoded_body, sub_path_name, filename)
                    else:
                        print('[S] 图片解码失败，无法保存')
                # 开始下一次读取
                content = b''
                content_type = content_transfer_encoding = ''
            elif 'Content-Type' in str(line):
                l = line.replace(b'\r', b'').replace(b'\n', b'').replace(b'\'', b'')
                content_type = bytes.decode(l).split(':')[-1]
                # print(content_type)
            elif 'Content-Transfer-Encoding' in str(line):
                content_transfer_encoding = str(line).split(':')[-1]
            elif 'Content-Location' in str(line):
                content_location = str(line).split(':')[-1]
            else:
                if b'\r\n' == line:
                    pass
                    # print('blank line')
                else:
                    content += line


def get_browser_type(input_path):
    with open(input_path, 'rb') as f:
        first_line = f.readline()
        if 'boundary' in str(first_line):
            return 'ie'
    return 'chrome'


def main(argv):
    global OUT_PATH
    input_path = ''
    outputpath = ''
    input_file = ''
    try:
        opts, args = getopt.getopt(argv, "hf:o:p:", ["file=", "opath=", "ipath="])
    except getopt.GetoptError:
        print_usage()
        sys.exit(2)
    for opt, arg in opts:
        if opt == '-h':
            print_usage()
            sys.exit()
        elif opt in ("-f", "--file"):
            input_file = arg
        elif opt in ("-p", "--ipath"):
            input_path = arg
        elif opt in ("-o", "--opath"):
            outputpath = arg

    if input_file == '' and input_path == '':
        print_usage()
        sys.exit(2)

    if outputpath != '':
        OUT_PATH = outputpath

    print('*' * 100)
    print('[S] 开始任务......')
    print('[C] 输入文件:' + input_file)
    print('[C] 输入目录:' + input_path)
    print('[C] 输出目录:' + OUT_PATH)

    btype = get_browser_type(input_file)
    print('[B] 浏览器：', btype)

    if btype == 'ie':
        if os.path.isfile(input_file):
            save_mht_all_images(input_file)
            print('[D] 导出全部完成。')
            print('*' * 100)
        else:
            if os.path.isdir(input_path):
                for root, dirs, files in os.walk(input_path):
                    for file in files:
                        print('[S] 开始处理文件:', file)
                        save_mht_all_images(os.path.join(root, file))
                        print('-' * 80)
                print('[D] 导出全部完成。')
                print('*' * 100)
            else:
                print_usage()
    else:
        if os.path.isfile(input_file):
            save_mht_all_images_chrome(input_file)
            print('[D] 导出全部完成。')
            print('*' * 100)
        else:
            if os.path.isdir(input_path):
                for root, dirs, files in os.walk(input_path):
                    for file in files:
                        print('[S] 开始处理文件:', file)
                        save_mht_all_images_chrome(os.path.join(root, file))
                        print('-' * 80)
                print('[D] 导出全部完成。')
                print('*' * 100)
            else:
                print_usage()


if __name__ == '__main__':
    main(sys.argv[1:])

上一篇文章地址：

https://img.h4ck.org.cn/2020/05/mht文件图片解析工具/

完整代码请参考github：

https://github.com/obaby/mht-image-extractor

原创文章，转载请注明： 转载自 obaby@mars

本文标题: 《mht文件图片解析工具（兼容Chrome/Blink）》

本文链接地址: http://h4ck.org.cn/2020/09/mht%e6%96%87%e4%bb%b6%e5%9b%be%e7%89%87%e8%a7%a3%e6%9e%90%e5%b7%a5%e5%85%b7%ef%bc%88%e5%85%bc%e5%ae%b9chrome-blink%ef%bc%89/

分享文章：