X-Git-Url: http://git.cielonegro.org/gitweb.cgi?a=blobdiff_plain;f=youtube_dl%2Fextractor%2Fyoutube.py;h=b7b91f354dbf07f0aa4e639d75d57fd51fe37687;hb=e72c7e41238681338c92427f6731c74947645308;hp=e719560711dbdd9d3f31ccc7432ee705132a91d4;hpb=ce08a86462984ff82368c1acd7285e9625855191;p=youtube-dl.git
diff --git a/youtube_dl/extractor/youtube.py b/youtube_dl/extractor/youtube.py
index e71956071..b7b91f354 100644
--- a/youtube_dl/extractor/youtube.py
+++ b/youtube_dl/extractor/youtube.py
@@ -264,9 +264,9 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
'266': {'ext': 'mp4', 'height': 2160, 'format_note': 'DASH video', 'acodec': 'none', 'preference': -40, 'vcodec': 'h264'},
# Dash mp4 audio
- '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 48, 'preference': -50},
- '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 128, 'preference': -50},
- '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'vcodec': 'none', 'abr': 256, 'preference': -50},
+ '139': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 48, 'preference': -50, 'container': 'm4a_dash'},
+ '140': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 128, 'preference': -50, 'container': 'm4a_dash'},
+ '141': {'ext': 'm4a', 'format_note': 'DASH audio', 'acodec': 'aac', 'vcodec': 'none', 'abr': 256, 'preference': -50, 'container': 'm4a_dash'},
# Dash webm
'167': {'ext': 'webm', 'height': 360, 'width': 640, 'format_note': 'DASH video', 'acodec': 'none', 'container': 'webm', 'vcodec': 'VP8', 'preference': -40},
@@ -394,6 +394,23 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
'format': '141',
},
},
+ # JS player signature function name containing $
+ {
+ 'url': 'https://www.youtube.com/watch?v=nfWlot6h_JM',
+ 'info_dict': {
+ 'id': 'nfWlot6h_JM',
+ 'ext': 'm4a',
+ 'title': 'Taylor Swift - Shake It Off',
+ 'description': 'md5:2acfda1b285bdd478ccec22f9918199d',
+ 'uploader': 'TaylorSwiftVEVO',
+ 'uploader_id': 'TaylorSwiftVEVO',
+ 'upload_date': '20140818',
+ },
+ 'params': {
+ 'youtube_include_dash_manifest': True,
+ 'format': '141',
+ },
+ },
# Controversy video
{
'url': 'https://www.youtube.com/watch?v=T4XJQO3qol8',
@@ -465,6 +482,20 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
'skip_download': 'requires avconv',
}
},
+ # Non-square pixels
+ {
+ 'url': 'https://www.youtube.com/watch?v=_b-2C3KPAM0',
+ 'info_dict': {
+ 'id': '_b-2C3KPAM0',
+ 'ext': 'mp4',
+ 'stretched_ratio': 16 / 9.,
+ 'upload_date': '20110310',
+ 'uploader_id': 'AllenMeow',
+ 'description': 'made by Wacom from Korea | åå¹&å 油添é by TY\'s Allen | æè¬heylisa00cavey1001åå¸ç±æ
æä¾æ¢åç¿»è¯',
+ 'uploader': 'å«è¾å«',
+ 'title': '[A-made] è®æ
å¦åå¹ç å¤ªå¦ æå°±æ¯é樣ç人',
+ },
+ }
]
def __init__(self, *args, **kwargs):
@@ -574,7 +605,7 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
def _parse_sig_js(self, jscode):
funcname = self._search_regex(
- r'\.sig\|\|([a-zA-Z0-9]+)\(', jscode,
+ r'\.sig\|\|([a-zA-Z0-9$]+)\(', jscode,
'Initial JS player signature function name')
jsi = JSInterpreter(jscode)
@@ -1051,6 +1082,16 @@ class YoutubeIE(YoutubeBaseInfoExtractor, SubtitlesInfoExtractor):
f['preference'] = f.get('preference', 0) - 10000
formats.extend(dash_formats)
+ # Check for malformed aspect ratio
+ stretched_m = re.search(
+ r'[0-9]+):(?P[0-9]+)">',
+ video_webpage)
+ if stretched_m:
+ ratio = float(stretched_m.group('w')) / float(stretched_m.group('h'))
+ for f in formats:
+ if f.get('vcodec') != 'none':
+ f['stretched_ratio'] = ratio
+
self._sort_formats(formats)
return {
@@ -1641,11 +1682,17 @@ class YoutubeTruncatedURLIE(InfoExtractor):
IE_NAME = 'youtube:truncated_url'
IE_DESC = False # Do not list
_VALID_URL = r'''(?x)
- (?:https?://)?[^/]+/watch\?(?:
+ (?:https?://)?
+ (?:\w+\.)?[yY][oO][uU][tT][uU][bB][eE](?:-nocookie)?\.com/
+ (?:watch\?(?:
feature=[a-z_]+|
- annotation_id=annotation_[^&]+
- )?$|
- (?:https?://)?(?:www\.)?youtube\.com/attribution_link\?a=[^&]+$
+ annotation_id=annotation_[^&]+|
+ x-yt-cl=[0-9]+|
+ )?
+ |
+ attribution_link\?a=[^&]+
+ )
+ $
'''
_TESTS = [{
@@ -1654,6 +1701,12 @@ class YoutubeTruncatedURLIE(InfoExtractor):
}, {
'url': 'http://www.youtube.com/watch?',
'only_matching': True,
+ }, {
+ 'url': 'https://www.youtube.com/watch?x-yt-cl=84503534',
+ 'only_matching': True,
+ }, {
+ 'url': 'https://www.youtube.com/watch?feature=foo',
+ 'only_matching': True,
}]
def _real_extract(self, url):
@@ -1669,7 +1722,7 @@ class YoutubeTruncatedURLIE(InfoExtractor):
class YoutubeTruncatedIDIE(InfoExtractor):
IE_NAME = 'youtube:truncated_id'
IE_DESC = False # Do not list
- _VALID_URL = r'https?://(?:www\.)youtube\.com/watch\?v=(?P[0-9A-Za-z_-]{1,10})$'
+ _VALID_URL = r'https?://(?:www\.)?youtube\.com/watch\?v=(?P[0-9A-Za-z_-]{1,10})$'
_TESTS = [{
'url': 'https://www.youtube.com/watch?v=N_708QY7Ob',