1
0
Fork 0
mirror of https://github.com/ytdl-org/youtube-dl.git synced 2024-11-18 08:16:10 +00:00
youtube-dl/youtube_dl/extractor/yahoo.py

173 lines
6.4 KiB
Python
Raw Normal View History

2014-02-11 07:55:24 +00:00
from __future__ import unicode_literals
import itertools
2014-04-04 20:22:11 +00:00
import json
2013-06-23 18:13:52 +00:00
import re
from .common import InfoExtractor, SearchInfoExtractor
2013-06-23 18:13:52 +00:00
from ..utils import (
compat_urllib_parse,
compat_urlparse,
clean_html,
int_or_none,
2013-06-23 18:13:52 +00:00
)
2013-06-23 18:13:52 +00:00
class YahooIE(InfoExtractor):
IE_DESC = 'Yahoo screen and movies'
_VALID_URL = r'https?://(?:screen|movies)\.yahoo\.com/.*?-(?P<id>[0-9]+)(?:-[a-z]+)?\.html'
_TESTS = [
{
2014-02-11 07:55:24 +00:00
'url': 'http://screen.yahoo.com/julian-smith-travis-legg-watch-214727115.html',
'md5': '4962b075c08be8690a922ee026d05e69',
'info_dict': {
2014-04-04 16:31:30 +00:00
'id': '214727115',
'ext': 'mp4',
2014-02-11 07:55:24 +00:00
'title': 'Julian Smith & Travis Legg Watch Julian Smith',
'description': 'Julian and Travis watch Julian Smith',
},
},
{
2014-02-11 07:55:24 +00:00
'url': 'http://screen.yahoo.com/wired/codefellas-s1-ep12-cougar-lies-103000935.html',
'md5': 'd6e6fc6e1313c608f316ddad7b82b306',
'info_dict': {
2014-04-04 16:31:30 +00:00
'id': '103000935',
'ext': 'mp4',
2014-02-11 07:55:24 +00:00
'title': 'Codefellas - The Cougar Lies with Spanish Moss',
'description': 'Agent Topple\'s mustache does its dirty work, and Nicole brokers a deal for peace. But why is the NSA collecting millions of Instagram brunch photos? And if your waffles have nothing to hide, what are they so worried about?',
},
2013-06-27 18:46:46 +00:00
},
{
'url': 'https://movies.yahoo.com/video/world-loves-spider-man-190819223.html',
'md5': '410b7104aa9893b765bc22787a22f3d9',
'info_dict': {
'id': '516ed8e2-2c4f-339f-a211-7a8b49d30845',
'ext': 'mp4',
'title': 'The World Loves Spider-Man',
'description': '''People all over the world are celebrating the release of \"The Amazing Spider-Man 2.\" We're taking a look at the enthusiastic response Spider-Man has received from viewers all over the world.''',
}
}
]
2013-06-23 18:13:52 +00:00
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('id')
webpage = self._download_webpage(url, video_id)
items_json = self._search_regex(
r'mediaItems: ({.*?})$', webpage, 'items', flags=re.MULTILINE,
default=None)
if items_json is None:
long_id = self._search_regex(
r'YUI\.namespace\("Media"\)\.CONTENT_ID\s*=\s*"([^"]+)"',
webpage, 'content ID')
video_id = long_id
else:
items = json.loads(items_json)
info = items['mediaItems']['query']['results']['mediaObj'][0]
# The 'meta' field is not always in the video webpage, we request it
# from another page
long_id = info['id']
2013-12-05 13:31:54 +00:00
return self._get_info(long_id, video_id)
def _get_info(self, long_id, video_id):
query = ('SELECT * FROM yahoo.media.video.streams WHERE id="%s"'
' AND plrs="86Gj0vCaSzV_Iuf6hNylf2" AND region="US"'
' AND protocol="http"' % long_id)
data = compat_urllib_parse.urlencode({
'q': query,
'env': 'prod',
'format': 'json',
})
2014-04-04 16:31:30 +00:00
query_result = self._download_json(
'http://video.query.yahoo.com/v1/public/yql?' + data,
2014-02-11 07:55:24 +00:00
video_id, 'Downloading video info')
info = query_result['query']['results']['mediaObj'][0]
meta = info['meta']
formats = []
for s in info['streams']:
format_info = {
'width': int_or_none(s.get('width')),
'height': int_or_none(s.get('height')),
'tbr': int_or_none(s.get('bitrate')),
}
host = s['host']
path = s['path']
if host.startswith('rtmp'):
format_info.update({
'url': host,
'play_path': path,
'ext': 'flv',
})
else:
format_url = compat_urlparse.urljoin(host, path)
format_info['url'] = format_url
formats.append(format_info)
self._sort_formats(formats)
return {
'id': video_id,
'title': meta['title'],
'formats': formats,
'description': clean_html(meta['description']),
'thumbnail': meta['thumbnail'],
}
2013-06-23 18:13:52 +00:00
class YahooNewsIE(YahooIE):
IE_NAME = 'yahoo:news'
_VALID_URL = r'http://news\.yahoo\.com/video/.*?-(?P<id>\d*?)\.html'
_TESTS = [{
2014-02-11 07:55:24 +00:00
'url': 'http://news.yahoo.com/video/china-moses-crazy-blues-104538833.html',
'md5': '67010fdf3a08d290e060a4dd96baa07b',
'info_dict': {
'id': '104538833',
'ext': 'mp4',
'title': 'China Moses Is Crazy About the Blues',
'description': 'md5:9900ab8cd5808175c7b3fe55b979bed0',
},
}]
def _real_extract(self, url):
mobj = re.match(self._VALID_URL, url)
video_id = mobj.group('id')
webpage = self._download_webpage(url, video_id)
2014-02-11 07:55:24 +00:00
long_id = self._search_regex(r'contentId: \'(.+?)\',', webpage, 'long id')
return self._get_info(long_id, video_id)
class YahooSearchIE(SearchInfoExtractor):
2014-02-11 07:55:24 +00:00
IE_DESC = 'Yahoo screen search'
_MAX_RESULTS = 1000
2014-02-11 07:55:24 +00:00
IE_NAME = 'screen.yahoo:search'
_SEARCH_KEY = 'yvsearch'
def _get_n_results(self, query, n):
"""Get a specified number of results for a query"""
2014-04-04 16:31:30 +00:00
entries = []
for pagenum in itertools.count(0):
2014-02-11 07:55:24 +00:00
result_url = 'http://video.search.yahoo.com/search/?p=%s&fr=screen&o=js&gs=0&b=%d' % (compat_urllib_parse.quote_plus(query), pagenum * 30)
2014-04-04 16:31:30 +00:00
info = self._download_json(result_url, query,
note='Downloading results page '+str(pagenum+1))
2014-02-11 07:55:24 +00:00
m = info['m']
results = info['results']
for (i, r) in enumerate(results):
2014-04-04 16:31:30 +00:00
if (pagenum * 30) + i >= n:
break
mobj = re.search(r'(?P<url>screen\.yahoo\.com/.*?-\d*?\.html)"', r)
e = self.url_result('http://' + mobj.group('url'), 'Yahoo')
2014-04-04 16:31:30 +00:00
entries.append(e)
if (pagenum * 30 + i >= n) or (m['last'] >= (m['total'] - 1)):
break
2014-04-04 16:31:30 +00:00
return {
'_type': 'playlist',
'id': query,
'entries': entries,
}