[extractor/common] _parse_mpd_formats fails on parsing AdaptationSet with contentType="text" and mimeType="application/mp4"

[tobijjah]

• I've verified and I assure that I'm running youtube-dl 2017.10.29
• At least skimmed through the README, most notably the FAQ and BUGS sections
• Searched the bugtracker for similar issues including closed ones
• Bug report (encountered problems with youtube-dl)

---

Description

Traceback (most recent call last):
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/youtube_dl/YoutubeDL.py", line 784, in extract_info
    ie_result = ie.extract(url)
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/youtube_dl/extractor/common.py", line 434, in extract
    ie_result = self._real_extract(url)
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/youtube_dl/extractor/hbo.py", line 309, in _real_extract
    return self._extract_from_path('series/%s' % path)
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/youtube_dl/extractor/hbo.py", line 154, in _extract_from_path
    return self._extract_from_xml(api_data, path)
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/youtube_dl/extractor/hbo.py", line 147, in _extract_from_xml
    data = self._extract_mpd_formats(url, 12)
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/youtube_dl/extractor/common.py", line 1753, in _extract_mpd_formats
    formats_dict=formats_dict, mpd_url=mpd_url)
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/youtube_dl/extractor/common.py", line 2007, in _parse_mpd_formats
    self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/youtube_dl/extractor/common.py", line 702, in report_warning
    '[%s] %s%s' % (self.IE_NAME, idstr, msg))
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/test/helper.py", line 244, in _report_warning
    real_warning(w)
  File "/home/ilex/Documents/code/python/projects/yt-dl/youtube-dl/test/test_download.py", line 52, in report_warning
    raise ExtractorError(message)
youtube_dl.utils.ExtractorError: [hbo:series] Unknown MIME type application/mp4 in DASH manifest; please report this issue on https://yt-dl.org/bug . Make sure you are using the latest version; see  https://yt-dl.org/update  on how to update. Be sure to call youtube-dl with the --verbose flag and include its complete output.

If you trying to parse with _parse_mpd_formats (respectively _extract_mpd_formats) the provided example MPD manifest file, the method fails and triggers an ExtractorError. The error trigger occurs if the parser reaches the AdaptationSet with the contentType="text" and the mimeType="application/mp4". In the attached code snippet (where the error occurs) it looks like the parser should skip an AdaptationSet with this contentType but it does not do it. Indeed, the 'content_type' attribute is set to application and the conditional statement is not fulfilled (see comments in the attached code snippet). I attached three possible solutions because I don't know if this behavior is wished or this is a bug of the parser. Additionally, I don't know if the provided solutions have unknown side effects etc. (I tested them just with a few examples and in my case it worked as it should)

• Example MPD manifest: https://dash.pro35.lv3.cdn.hbo.com/av/videos/specials/night-of-too-many-stars/trailer-37690706_PRO35/manifest.mpd
• Relevant code snippet from common.InfoExtractor._parse_mpd_formats():

for representation in adaptation_set.findall(_add_ns('Representation')):
    if is_drm_protected(representation):
        continue
    representation_attrib = adaptation_set.attrib.copy()
    representation_attrib.update(representation.attrib)  # contentType is text
    # According to [1, 5.3.7.2, Table 9, page 41], @mimeType is mandatory
    mime_type = representation_attrib['mimeType']  # mimeType is application/mp4
    content_type = mime_type.split('/')[0]  # sets content_type to application
    if content_type == 'text':  # therefore this condition is not fulfilled and this AdaptationSet is not skipped
        # TODO implement WebVTT downloading
        pass
    elif content_type in ('video', 'audio'):
        ...
    else:  # trigger error
        self.report_warning('Unknown MIME type %s in DASH manifest' % mime_type)

• Possible solutions for common.InfoExtractor._parse_mpd_formats():

# First: tested and my preferred 
representation_attrib.update(representation.attrib)
mime_type = representation_attrib['mimeType'] 
content_type = mime_type.split('/')[0] 
if content_type == 'text' or representation_attrib.get('contentType') == 'text':  # change
     pass

# Second: untested
representation_attrib.update(representation.attrib)
mime_type = representation_attrib['mimeType'] 
content_type = representation_attrib.get('contentType')  # change
if content_type == 'text': 
     pass

# Third: tested
representation_attrib.update(representation.attrib)
mime_type = representation_attrib['mimeType'] 
content_type = mime_type.split('/')[0] 
if content_type in ('text', 'application'):  # change
     pass