xml-to-har.py

#!/usr/bin/env python
"""
Converts Internet Explorer 'capture network traffic' XML to a HAR file.

Turns out that XML is just a HAR file anyways, but in XML form. So this
just converts it to JSON, and Bob's your uncle.

Requires Python 2.7+ and LXML.
"""
from __future__ import unicode_literals

import argparse
import json
from lxml import objectify
import sys


if sys.version_info > (3,):
    str_type = str
else:
    str_type = unicode


list_things = {
    'pages',
    'entries',
    'cookies',
    'queryString',
    'headers',
}


def xml_to_dict(element):
    if element.tag in list_things:
        return [xml_to_dict(e) for e in element.getchildren()]
    else:
        if element.getchildren():
            return {e.tag: xml_to_dict(e) for e in element.getchildren()}
        else:
            return str_type(element.pyval)


def main():
    parser = argparse.ArgumentParser(description="Convert IE's crazy XML-HAR into a real HAR file")
    parser.add_argument('infile', type=argparse.FileType('r'), default=sys.stdin)
    parser.add_argument('outfile', type=argparse.FileType('w'), default=sys.stdout)
    args = parser.parse_args()

    tree = objectify.parse(args.infile)
    root = tree.getroot()
    d = {root.tag: xml_to_dict(root)}
    json.dump(d, args.outfile, indent=2, sort_keys=True)


if __name__ == '__main__':
    main()

For anyone struggling with file encodings on the Windows (MSYSGIT) command line like me, I created a fork that forces UTF-8 when reading and writing the files. This small fix cost me almost three hours today since I was puzzled why lxml or an extra open() would ignore my encoding (unless I specify 'ascii'). Turns out argparse has a bit more magic in it than I thought.

Coming from php; had to do this to the json in order for it to be validated in http://www.softwareishard.com/har/viewer/ Source xml came from IE11.

$contents = file_get_contents('NetworkData.har');
$json = json_decode($contents, TRUE);
foreach ($json['log']['entries'] as $key => &$values) {
  $values['time'] = (int) $values['time'];
  $values['timings']['receive'] = (int) $values['timings']['receive'];
  $values['timings']['send'] = (int) $values['timings']['send'];
  $values['timings']['wait'] = (int) $values['timings']['wait'];
  if (!isset($values['response']['redirectURL'])) {
    $values['response']['redirectURL'] = '';
  }
  if (!isset($values['cache'])) {
    $values['cache'] = array();
  }
}
$output = json_encode($json, JSON_HEX_TAG | JSON_HEX_APOS | JSON_HEX_AMP | JSON_HEX_QUOT | JSON_PRETTY_PRINT);
file_put_contents('NetworkData2.har', $output);

Does it work in Python 3, or only in 2.7?

Thanks a lot for this handy tool, much appreciated!

No Bob is not my uncle :P

I want to record chrome session in HAR and playback it in IE. Is there an option?

So glad I landed here, it's very useful :D Thank you so much!