aboutsummaryrefslogtreecommitdiff
path: root/youtube_dl/extractor/nintendo.py
diff options
context:
space:
mode:
authorSergey M․ <dstftw@gmail.com>2016-07-17 00:01:25 +0700
committerSergey M․ <dstftw@gmail.com>2016-07-17 00:01:25 +0700
commit49bc16b95ed7d418b353fda46ba845ac0eca648b (patch)
tree3bb630040157c0d7b12928e3930b15825e7adb3c /youtube_dl/extractor/nintendo.py
parenta2f9ca1e67e6b926957abe3b35a4d78355bef7d4 (diff)
[nintendo] Improve playlist extraction (Closes #9986)
Diffstat (limited to 'youtube_dl/extractor/nintendo.py')
-rw-r--r--youtube_dl/extractor/nintendo.py33
1 files changed, 16 insertions, 17 deletions
diff --git a/youtube_dl/extractor/nintendo.py b/youtube_dl/extractor/nintendo.py
index 57333ada0..4b4e66b05 100644
--- a/youtube_dl/extractor/nintendo.py
+++ b/youtube_dl/extractor/nintendo.py
@@ -1,13 +1,15 @@
+# coding: utf-8
from __future__ import unicode_literals
+import re
+
from .common import InfoExtractor
from .ooyala import OoyalaIE
-
-import re
+from ..utils import unescapeHTML
class NintendoIE(InfoExtractor):
- _VALID_URL = r'https?://(?:www\.)?nintendo\.com/games/detail/(?P<id>[\w-]+)'
+ _VALID_URL = r'https?://(?:www\.)?nintendo\.com/games/detail/(?P<id>[^/?#&]+)'
_TESTS = [{
'url': 'http://www.nintendo.com/games/detail/yEiAzhU2eQI1KZ7wOHhngFoAHc1FpHwj',
'info_dict': {
@@ -24,24 +26,21 @@ class NintendoIE(InfoExtractor):
'url': 'http://www.nintendo.com/games/detail/tokyo-mirage-sessions-fe-wii-u',
'info_dict': {
'id': 'tokyo-mirage-sessions-fe-wii-u',
+ 'title': 'Tokyo Mirage Sessions ♯FE',
},
- 'params': {
- 'skip_download': True,
- },
- 'add_ie': ['Ooyala'],
- 'playlist_count': 4,
+ 'playlist_count': 3,
}]
def _real_extract(self, url):
- video_id = self._match_id(url)
- webpage = self._download_webpage(url, video_id)
+ page_id = self._match_id(url)
- ooyala_codes = re.findall(
- r'data-video-code=(["\'])(?P<code>.+?)\1',
- webpage)
+ webpage = self._download_webpage(url, page_id)
- entries = []
- for ooyala_code in ooyala_codes:
- entries.append(OoyalaIE._build_url_result(ooyala_code[1]))
+ entries = [
+ OoyalaIE._build_url_result(m.group('code'))
+ for m in re.finditer(
+ r'class=(["\'])embed-video\1[^>]+data-video-code=(["\'])(?P<code>(?:(?!\2).)+)\2',
+ webpage)]
- return self.playlist_result(entries, video_id, self._og_search_title(webpage))
+ return self.playlist_result(
+ entries, page_id, unescapeHTML(self._og_search_title(webpage, fatal=False)))