考虑到上述文档的结构，这就是您需要的表达式。它将匹配所有图像URL和每个图像描述。

<image\sxlink:href="(https:\/\/[^"\s]+)".*?<title><\/title><desc>(.+?)<\/desc><\/image>

演示
在php中的用法：

$html = '<p>Members of our tim</p><image xlink:href="https://lh4.googleusercontent.com/z3GK1MdYyLTo0Q0xLmawvcptIrK4qkQx7XJWUgTK_i6Psm22GBqZXBh-w0TeQ5xgKxckQOB2wHWySSIpNj3tXx65MPXmaxKjK4ye_Xu-wAUFKLVhvWFgIedtzxo" width="100%" height="100%" preserveAspectRatio="none"><title></title><desc>Bradley Byrne.jpg</desc></image><h1>Some big title</h1><p>Something <span>more</span> here</p><image xlink:href="https://lh5.googleusercontent.com/fWYh7qTWqu4_4oxAiNhmnMCmD6DScZ6bIvkF5nSFunU8NxKlBT1T-1J85MJCqghhbChFzoLi-p4ZFVDCA2DWWBP9Paagp9ZgshqnGK5CQQF6D7IoBGihcFZoOms" width="100%" height="100%" preserveAspectRatio="none"><title></title><desc>Spencer Bachus 113th Congress.jpg</desc></image><h1>TITLE</h1><p>Testing, testing, testing</p><image xlink:href="https://lh5.googleusercontent.com/VAHzM6OkdtxT61j9XSgTDKlpVi99WsFfzNAlvqmnpCi90XFs9aUNMfuCeeeQ3e26fykjveoxldHvv5jO1Bk9IeEmeU7DdGVAM1N9xXoB8tJTYBeTeFBxigXtT5s" width="100%" height="100%" preserveAspectRatio="none"><title></title><desc>Kyrsten Sinema 113th Congress.jpg</desc></image><p>Last updated on 25th of July, 2014</p>';
$pattern = '/<image\sxlink:href="(https:\/\/[^"\s]+)".*?<title><\/title><desc>(.+?)<\/desc><\/image>/';
if(preg_match_all($pattern, $html, $matches)){
  $size_of_matches = count($matches[0]);
  for($i = 0; $i < $size_of_matches; $i++){
    echo $matches[1][$i] . " -> " . $matches[2][$i] . "<br />";
  }
}

输出：

https://lh4.googleusercontent.com/z3GK1MdYyLTo0Q0xLmawvcptIrK4qkQx7XJWUgTK_i6Psm22GBqZXBh-w0TeQ5xgKxckQOB2wHWySSIpNj3tXx65MPXmaxKjK4ye_Xu-wAUFKLVhvWFgIedtzxo -> Bradley Byrne.jpg
https://lh5.googleusercontent.com/fWYh7qTWqu4_4oxAiNhmnMCmD6DScZ6bIvkF5nSFunU8NxKlBT1T-1J85MJCqghhbChFzoLi-p4ZFVDCA2DWWBP9Paagp9ZgshqnGK5CQQF6D7IoBGihcFZoOms -> Spencer Bachus 113th Congress.jpg
https://lh5.googleusercontent.com/VAHzM6OkdtxT61j9XSgTDKlpVi99WsFfzNAlvqmnpCi90XFs9aUNMfuCeeeQ3e26fykjveoxldHvv5jO1Bk9IeEmeU7DdGVAM1N9xXoB8tJTYBeTeFBxigXtT5s -> Kyrsten Sinema 113th Congress.jpg

我没有使用mapreduce或importtsv的经验，所以我使用c#使用了不同的方法。正如hex4949所指出的，图像确实有与之相关联的文本。您只需从文档的源获取该数据（即，右键单击-->查看页面源）。
此代码读入文档的源代码，尝试将政治家与图像文件（基于发布的可用信息）匹配，并将结果写入文本文件。代码中有许多regex的c风格示例。这里有一个输出示例。