diff options
Diffstat (limited to 'youtube_dl/extractor/common.py')
| -rw-r--r-- | youtube_dl/extractor/common.py | 46 |
1 files changed, 38 insertions, 8 deletions
diff --git a/youtube_dl/extractor/common.py b/youtube_dl/extractor/common.py index 82f5de2d8..5a2d0d995 100644 --- a/youtube_dl/extractor/common.py +++ b/youtube_dl/extractor/common.py @@ -27,7 +27,9 @@ from ..utils import ( bug_reports_message, clean_html, compiled_regex_type, + determine_ext, ExtractorError, + fix_xml_ampersands, float_or_none, int_or_none, RegexNotFoundError, @@ -706,10 +708,23 @@ class InfoExtractor(object): 'twitter card player') @staticmethod - def _form_hidden_inputs(html): - return dict(re.findall( - r'<input\s+type="hidden"\s+name="([^"]+)"\s+(?:id="[^"]+"\s+)?value="([^"]*)"', - html)) + def _hidden_inputs(html): + return dict([ + (input.group('name'), input.group('value')) for input in re.finditer( + r'''(?x) + <input\s+ + type=(?P<q_hidden>["\'])hidden(?P=q_hidden)\s+ + name=(?P<q_name>["\'])(?P<name>.+?)(?P=q_name)\s+ + (?:id=(?P<q_id>["\']).+?(?P=q_id)\s+)? + value=(?P<q_value>["\'])(?P<value>.*?)(?P=q_value) + ''', html) + ]) + + def _form_hidden_inputs(self, form_id, html): + form = self._search_regex( + r'(?s)<form[^>]+?id=(["\'])%s\1[^>]*>(?P<form>.+?)</form>' % form_id, + html, '%s form' % form_id, group='form') + return self._hidden_inputs(form) def _sort_formats(self, formats, field_preference=None): if not formats: @@ -821,10 +836,14 @@ class InfoExtractor(object): self.to_screen(msg) time.sleep(timeout) - def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None): + def _extract_f4m_formats(self, manifest_url, video_id, preference=None, f4m_id=None, + transform_source=lambda s: fix_xml_ampersands(s).strip()): manifest = self._download_xml( manifest_url, video_id, 'Downloading f4m manifest', - 'Unable to download f4m manifest') + 'Unable to download f4m manifest', + # Some manifests may be malformed, e.g. prosiebensat1 generated manifests + # (see https://github.com/rg3/youtube-dl/issues/6215#issuecomment-121704244) + transform_source=transform_source) formats = [] manifest_version = '1.0' @@ -834,8 +853,19 @@ class InfoExtractor(object): media_nodes = manifest.findall('{http://ns.adobe.com/f4m/2.0}media') for i, media_el in enumerate(media_nodes): if manifest_version == '2.0': - manifest_url = ('/'.join(manifest_url.split('/')[:-1]) + '/' + - (media_el.attrib.get('href') or media_el.attrib.get('url'))) + media_url = media_el.attrib.get('href') or media_el.attrib.get('url') + if not media_url: + continue + manifest_url = ( + media_url if media_url.startswith('http://') or media_url.startswith('https://') + else ('/'.join(manifest_url.split('/')[:-1]) + '/' + media_url)) + # If media_url is itself a f4m manifest do the recursive extraction + # since bitrates in parent manifest (this one) and media_url manifest + # may differ leading to inability to resolve the format by requested + # bitrate in f4m downloader + if determine_ext(manifest_url) == 'f4m': + formats.extend(self._extract_f4m_formats(manifest_url, video_id, preference, f4m_id)) + continue tbr = int_or_none(media_el.attrib.get('bitrate')) formats.append({ 'format_id': '-'.join(filter(None, [f4m_id, compat_str(i if tbr is None else tbr)])), |