[ted] Improve extraction and update tests

This commit is contained in:
Sergey M․ 2018-07-28 22:09:53 +07:00
parent 9a984265b9
commit cd3a3ff93b
No known key found for this signature in database
GPG key ID: 2C393E0F18A9236D

View file

@ -7,8 +7,10 @@
from ..compat import compat_str
from ..utils import (
float_or_none,
int_or_none,
try_get,
url_or_none,
)
@ -30,7 +32,7 @@ class TEDIE(InfoExtractor):
'''
_TESTS = [{
'url': 'http://www.ted.com/talks/dan_dennett_on_our_consciousness.html',
'md5': '0de43ac406aa3e4ea74b66c9c7789b13',
'md5': 'b0ce2b05ca215042124fbc9e3886493a',
'info_dict': {
'id': '102',
'ext': 'mp4',
@ -42,24 +44,30 @@ class TEDIE(InfoExtractor):
'uploader': 'Dan Dennett',
'width': 853,
'duration': 1308,
}
}, {
'url': 'http://www.ted.com/watch/ted-institute/ted-bcg/vishal-sikka-the-beauty-and-power-of-algorithms',
'md5': 'b899ac15e345fb39534d913f7606082b',
'info_dict': {
'id': 'tSVI8ta_P4w',
'ext': 'mp4',
'title': 'Vishal Sikka: The beauty and power of algorithms',
'thumbnail': r're:^https?://.+\.jpg',
'description': 'md5:6261fdfe3e02f4f579cbbfc00aff73f4',
'upload_date': '20140122',
'uploader_id': 'TEDInstitute',
'uploader': 'TED Institute',
'view_count': int,
'comment_count': int,
'tags': list,
},
'params': {
'skip_download': True,
},
}, {
# missing HTTP bitrates
'url': 'https://www.ted.com/talks/vishal_sikka_the_beauty_and_power_of_algorithms',
'info_dict': {
'id': '6069',
'ext': 'mp4',
'title': 'The beauty and power of algorithms',
'thumbnail': r're:^https?://.+\.jpg',
'description': 'md5:734e352710fb00d840ab87ae31aaf688',
'uploader': 'Vishal Sikka',
},
'params': {
'skip_download': True,
},
'add_ie': ['Youtube'],
}, {
'url': 'http://www.ted.com/talks/gabby_giffords_and_mark_kelly_be_passionate_be_courageous_be_your_best',
'md5': '71b3ab2f4233012dce09d515c9c39ce2',
'md5': 'e6b9617c01a7970ceac8bb2c92c346c0',
'info_dict': {
'id': '1972',
'ext': 'mp4',
@ -68,6 +76,9 @@ class TEDIE(InfoExtractor):
'description': 'md5:5174aed4d0f16021b704120360f72b92',
'duration': 1128,
},
'params': {
'skip_download': True,
},
}, {
'url': 'http://www.ted.com/playlists/who_are_the_hackers',
'info_dict': {
@ -91,22 +102,6 @@ class TEDIE(InfoExtractor):
'params': {
'skip_download': True,
},
}, {
# YouTube video
'url': 'http://www.ted.com/talks/jeffrey_kluger_the_sibling_bond',
'add_ie': ['Youtube'],
'info_dict': {
'id': 'aFBIPO-P7LM',
'ext': 'mp4',
'title': 'The hidden power of siblings: Jeff Kluger at TEDxAsheville',
'description': 'md5:3d7a4f50d95ca5dd67104e2a20f43fe1',
'uploader': 'TEDx Talks',
'uploader_id': 'TEDxTalks',
'upload_date': '20111216',
},
'params': {
'skip_download': True,
},
}, {
# no nativeDownloads
'url': 'https://www.ted.com/talks/tom_thum_the_orchestra_in_my_mouth',
@ -116,6 +111,9 @@ class TEDIE(InfoExtractor):
'title': 'The orchestra in my mouth',
'description': 'md5:5d1d78650e2f8dfcbb8ebee2951ac29a',
'uploader': 'Tom Thum',
'view_count': int,
'comment_count': int,
'tags': list,
},
'params': {
'skip_download': True,
@ -174,24 +172,11 @@ def _talk_info(self, url, video_name):
info = self._extract_info(webpage)
talk_info = try_get(
info, lambda x: x['__INITIAL_DATA__']['talks'][0],
dict) or info['talks'][0]
data = try_get(info, lambda x: x['__INITIAL_DATA__'], dict) or info
talk_info = data['talks'][0]
title = talk_info['title'].strip()
external = talk_info.get('external')
if external:
service = external['service']
self.to_screen('Found video from %s' % service)
ext_url = None
if service.lower() == 'youtube':
ext_url = external.get('code')
return {
'_type': 'url',
'url': ext_url or external['uri'],
}
native_downloads = try_get(
talk_info,
(lambda x: x['downloads']['nativeDownloads'],
@ -211,10 +196,24 @@ def _talk_info(self, url, video_name):
player_talk = talk_info['player_talks'][0]
external = player_talk.get('external')
if isinstance(external, dict):
service = external.get('service')
if isinstance(service, compat_str):
ext_url = None
if service.lower() == 'youtube':
ext_url = external.get('code')
return {
'_type': 'url',
'url': ext_url or external['uri'],
}
resources_ = player_talk.get('resources') or talk_info.get('resources')
http_url = None
for format_id, resources in resources_.items():
if not isinstance(resources, dict):
continue
if format_id == 'h264':
for resource in resources:
h264_url = resource.get('file')
@ -243,8 +242,12 @@ def _talk_info(self, url, video_name):
'tbr': int_or_none(resource.get('bitrate')),
})
elif format_id == 'hls':
stream_url = url_or_none(resources.get('stream'))
if not stream_url:
continue
formats.extend(self._extract_m3u8_formats(
resources.get('stream'), video_name, 'mp4', m3u8_id=format_id, fatal=False))
stream_url, video_name, 'mp4', m3u8_id=format_id,
fatal=False))
m3u8_formats = list(filter(
lambda f: f.get('protocol') == 'm3u8' and f.get('vcodec') != 'none',
@ -254,9 +257,13 @@ def _talk_info(self, url, video_name):
bitrate = self._search_regex(r'(\d+k)', m3u8_format['url'], 'bitrate', default=None)
if not bitrate:
continue
bitrate_url = re.sub(r'\d+k', bitrate, http_url)
if not self._is_valid_url(
bitrate_url, video_name, '%s bitrate' % bitrate):
continue
f = m3u8_format.copy()
f.update({
'url': re.sub(r'\d+k', bitrate, http_url),
'url': bitrate_url,
'format_id': m3u8_format['format_id'].replace('hls', 'http'),
'protocol': 'http',
})
@ -282,7 +289,11 @@ def _talk_info(self, url, video_name):
'description': self._og_search_description(webpage),
'subtitles': self._get_subtitles(video_id, talk_info),
'formats': formats,
'duration': talk_info.get('duration'),
'duration': float_or_none(talk_info.get('duration')),
'view_count': int_or_none(data.get('viewed_count')),
'comment_count': int_or_none(
try_get(data, lambda x: x['comments']['count'])),
'tags': try_get(talk_info, lambda x: x['tags'], list),
}
def _get_subtitles(self, video_id, talk_info):