[brightcove:new] Combine <video> tag searches

Per @yan12125.

Instead of searching for <video> followed by <script>, now
search for <video> optionally followed by <script>. If we
find <script>, use it. Otherwise, extract_attributes() from the video
tag and use those.

Move stripping the 'ref': prefix of data-video-id="ref:12345" out of
regexp into code, since we now use extract_attributes().
This commit is contained in:
John Hawkinson 2017-03-07 01:00:42 -05:00
parent 68ba9882a3
commit 49571c1c2f

View File

@ -482,8 +482,12 @@ class BrightcoveNewIE(InfoExtractor):
# Reference:
# 1. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideoiniframe
# 2. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#setvideousingjavascript
# 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/embed-in-page.html
# 3. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/embed-in-page.html
# 4. https://support.brightcove.com/en/video-cloud/docs/dynamically-assigning-videos-player
# 5. http://docs.brightcove.com/en/video-cloud/brightcove-player/guides/publish-video.html#tag
# [5] looks like:
# <video data-video-id="5320421710001" data-account="245991542" data-player="SJWAiyYWg" data-embed="default" class="video-js" controls itemscope itemtype="http://schema.org/VideoObject">
entries = []
@ -492,31 +496,30 @@ class BrightcoveNewIE(InfoExtractor):
r'<iframe[^>]+src=(["\'])((?:https?:)?//players\.brightcove\.net/\d+/[^/]+/index\.html.+?)\1', webpage):
entries.append(url if url.startswith('http') else 'http:' + url)
# Look for embed_in_page embeds [2]
for video_id, account_id, player_id, embed in re.findall(
# According to examples from [3] it's unclear whether video id
# may be optional and what to do when it is
# According to [4] data-video-id may be prefixed with ref:
r'''(?sx)
<video[^>]+
data-video-id=["\'](\d+|ref:[^"\']+)["\'][^>]*>.*?
</video>.*?
<script[^>]+
src=["\'](?:https?:)?//players\.brightcove\.net/
(\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
''', webpage):
entries.append(
'http://players.brightcove.net/%s/%s_%s/index.html?videoId=%s'
% (account_id, player_id, embed, video_id))
# <video data-brightcove-video-id="5320421710001" data-account="245991542" data-player="SJWAiyYWg" data-embed="default" class="video-js" controls itemscope itemtype="http://schema.org/VideoObject">
for video in re.findall(r'(?i)(<video[^>]+>)', webpage):
# Look for embed_in_page embeds [2] and <video> tags [5]
for video, script_tag, account_id, player_id, embed in re.findall(
r'''(?isx)
(<video[^>]+>)
(?:.*?
(<script[^>]+
src=["\'](?:https?:)?//players\.brightcove\.net/
(\d+)/([^/]+)_([^/]+)/index(?:\.min)?\.js
)
)?
''', webpage):
attrs = extract_attributes(video)
# According to examples from [3] it's unclear whether video id
# may be optional and what to do when it is
video_id = attrs.get('data-video-id')
account_id = attrs.get('data-account')
player_id = attrs.get('data-player')
embed = attrs.get('data-embed')
# See PR#12099/bostonglobe.py for 'data-brightcove-video-id' variant
if not account_id: account_id = attrs.get('data-account')
if not player_id: player_id = attrs.get('data-player')
if not embed: embed = attrs.get('data-embed')
# According to [4] data-video-id may be prefixed with 'ref:'
video_id = video_id.rpartition('ref:')[2]
if video_id and account_id and player_id and embed:
entries.append(