full_title = track_id
token = mobj.group('secret_token')
if token:
- info_json_url += "&secret_token=" + token
+ info_json_url += '&secret_token=' + token
elif mobj.group('player'):
query = compat_urlparse.parse_qs(compat_urlparse.urlparse(url).query)
real_url = query['url'][0]
resource = mobj.group('rsrc') or 'all'
base_url = self._BASE_URL_MAP[resource] % user['id']
- next_href = None
+ COMMON_QUERY = {
+ 'limit': 50,
+ 'client_id': self._CLIENT_ID,
+ 'linked_partitioning': '1',
+ }
+
+ query = COMMON_QUERY.copy()
+ query['offset'] = 0
+
+ next_href = base_url + '?' + compat_urllib_parse.urlencode(query)
entries = []
for i in itertools.count():
- if not next_href:
- data = compat_urllib_parse.urlencode({
- 'offset': i * 50,
- 'limit': 50,
- 'client_id': self._CLIENT_ID,
- 'linked_partitioning': '1',
- 'representation': 'speedy',
- })
- next_href = base_url + '?' + data
-
response = self._download_json(
next_href, uploader, 'Downloading track page %s' % (i + 1))
collection = response['collection']
-
if not collection:
- self.to_screen('%s: End page received' % uploader)
break
def resolve_permalink_url(candidates):
if permalink_url:
entries.append(self.url_result(permalink_url))
- if 'next_href' in response:
- next_href = response['next_href']
- if not next_href:
- break
- else:
- next_href = None
+ next_href = response.get('next_href')
+ if not next_href:
+ break
+
+ parsed_next_href = compat_urlparse.urlparse(response['next_href'])
+ qs = compat_urlparse.parse_qs(parsed_next_href.query)
+ qs.update(COMMON_QUERY)
+ next_href = compat_urlparse.urlunparse(
+ parsed_next_href._replace(query=compat_urllib_parse.urlencode(qs, True)))
return {
'_type': 'playlist',