[youtube] Users: download from the api in json to simplify extraction (fixes #1358)

There could be duplicate videos or other videos if the description have links.
This commit is contained in:
Jaime Marquínez Ferrándiz 2013-09-06 10:41:46 +02:00
parent 0638ad9999
commit fd9cf73836

View file

@ -1018,8 +1018,7 @@ class YoutubeUserIE(InfoExtractor):
_VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?)|ytuser:)([A-Za-z0-9_-]+)' _VALID_URL = r'(?:(?:(?:https?://)?(?:\w+\.)?youtube\.com/(?:user/)?)|ytuser:)([A-Za-z0-9_-]+)'
_TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s' _TEMPLATE_URL = 'http://gdata.youtube.com/feeds/api/users/%s'
_GDATA_PAGE_SIZE = 50 _GDATA_PAGE_SIZE = 50
_GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d' _GDATA_URL = 'http://gdata.youtube.com/feeds/api/users/%s/uploads?max-results=%d&start-index=%d&alt=json'
_VIDEO_INDICATOR = r'/watch\?v=(.+?)[\<&]'
IE_NAME = u'youtube:user' IE_NAME = u'youtube:user'
def suitable(cls, url): def suitable(cls, url):
@ -1048,13 +1047,15 @@ class YoutubeUserIE(InfoExtractor):
page = self._download_webpage(gdata_url, username, page = self._download_webpage(gdata_url, username,
u'Downloading video ids from %d to %d' % (start_index, start_index + self._GDATA_PAGE_SIZE)) u'Downloading video ids from %d to %d' % (start_index, start_index + self._GDATA_PAGE_SIZE))
try:
response = json.loads(page)
except ValueError as err:
raise ExtractorError(u'Invalid JSON in API response: ' + compat_str(err))
# Extract video identifiers # Extract video identifiers
ids_in_page = [] ids_in_page = []
for entry in response['feed']['entry']:
for mobj in re.finditer(self._VIDEO_INDICATOR, page): ids_in_page.append(entry['id']['$t'].split('/')[-1])
if mobj.group(1) not in ids_in_page:
ids_in_page.append(mobj.group(1))
video_ids.extend(ids_in_page) video_ids.extend(ids_in_page)
# A little optimization - if current page is not # A little optimization - if current page is not