diff options
Diffstat (limited to 'youtube_dl/extractor/instagram.py')
| -rw-r--r-- | youtube_dl/extractor/instagram.py | 33 |
1 files changed, 13 insertions, 20 deletions
diff --git a/youtube_dl/extractor/instagram.py b/youtube_dl/extractor/instagram.py index 6ae704efd..ddc42882a 100644 --- a/youtube_dl/extractor/instagram.py +++ b/youtube_dl/extractor/instagram.py @@ -5,12 +5,13 @@ from .common import InfoExtractor class InstagramIE(InfoExtractor): _VALID_URL = r'(?:http://)?instagram.com/p/(.*?)/' _TEST = { - u'url': u'http://instagram.com/p/aye83DjauH/#', + u'url': u'http://instagram.com/p/aye83DjauH/?foo=bar#abc', u'file': u'aye83DjauH.mp4', u'md5': u'0d2da106a9d2631273e192b372806516', u'info_dict': { u"uploader_id": u"naomipq", - u"title": u"Video by naomipq" + u"title": u"Video by naomipq", + u'description': u'md5:1f17f0ab29bd6fe2bfad705f58de3cb8', } } @@ -18,25 +19,17 @@ class InstagramIE(InfoExtractor): mobj = re.match(self._VALID_URL, url) video_id = mobj.group(1) webpage = self._download_webpage(url, video_id) - video_url = self._html_search_regex( - r'<meta property="og:video" content="(.+?)"', - webpage, u'video URL') - thumbnail_url = self._html_search_regex( - r'<meta property="og:image" content="(.+?)" />', - webpage, u'thumbnail URL', fatal=False) - html_title = self._html_search_regex( - r'<title>(.+?)</title>', - webpage, u'title', flags=re.DOTALL) - title = re.sub(u'(?: *\(Videos?\))? \u2022 Instagram$', '', html_title).strip() - uploader_id = self._html_search_regex(r'content="(.*?)\'s video on Instagram', - webpage, u'uploader name', fatal=False) - ext = 'mp4' + uploader_id = self._search_regex(r'"owner":{"username":"(.+?)"', + webpage, u'uploader id', fatal=False) + desc = self._search_regex(r'"caption":"(.*?)"', webpage, u'description', + fatal=False) return [{ 'id': video_id, - 'url': video_url, - 'ext': ext, - 'title': title, - 'thumbnail': thumbnail_url, - 'uploader_id' : uploader_id + 'url': self._og_search_video_url(webpage), + 'ext': 'mp4', + 'title': u'Video by %s' % uploader_id, + 'thumbnail': self._og_search_thumbnail(webpage), + 'uploader_id' : uploader_id, + 'description': desc, }] |