Add support for spanish episodes on South Park

[dilworks]

South Park Studios just added their entire catalog of episodes dubbed to Latin American Spanish (http://southpark.cc.com/blog/2015/04/16/all-seasons-of-south-park-now-available-in-spanish), which is insanely great for us at the south of the border. However, youtube-dl can't deal with those as the URLs are slightly different.

After looking at the extractor source (southpark.py), the fix might be trivial:

_VALID_URL = r'https?://(www\.)?(?P<url>southpark\.cc\.com/(clips|full-episodes)/(?P<id>.+?)(\?|#|$))'

The path for the Spanish episodes is "episodios-en-espanol", so it may be just matter of adding that string to the valid URLs regex.

[dilworks]

The suggested regex fix (just add "episodios-en-espanol" to the valid URLs regex) indeed does the work:

Sample URL: http://southpark.cc.com/episodios-en-espanol/s10e08-haz-el-amor-no-el-warcraft
Fix results:

tomman@tomman-lp-c2:~/workbench/tsdx-mbench/youtube-dl$ ./youtube-dl --list-formats http://southpark.cc.com/episodios-en-espanol/s10e08-haz-el-amor-no-el-warcraft
[southpark.cc.com] s10e08-haz-el-amor-no-el-warcraft: Downloading webpage
[southpark.cc.com] 5ff1eda8-ecfd-11e0-aca6-0026b9414f30: Downloading info
[southpark.cc.com] e50791a2-ed00-11e0-aca6-0026b9414f30: Extracting information
[southpark.cc.com] e50791a2-ed00-11e0-aca6-0026b9414f30: Downloading video urls
[southpark.cc.com] e507fd68-ed00-11e0-aca6-0026b9414f30: Extracting information
[southpark.cc.com] e507fd68-ed00-11e0-aca6-0026b9414f30: Downloading video urls
[southpark.cc.com] e50957d0-ed00-11e0-aca6-0026b9414f30: Extracting information
[southpark.cc.com] e50957d0-ed00-11e0-aca6-0026b9414f30: Downloading video urls
[download] Downloading playlist: None
[southpark.cc.com] playlist None: Collected 3 video ids (downloading 3 of them)
[download] Downloading video 1 of 3
[info] Available formats for e50791a2-ed00-11e0-aca6-0026b9414f30:
format code  extension  resolution note
300          mp4        384x216    
450          mp4        512x288    
700          mp4        768x432    
1000         mp4        960x540    
1200         mp4        1280x720   (best)
[download] Downloading video 2 of 3
[info] Available formats for e507fd68-ed00-11e0-aca6-0026b9414f30:
format code  extension  resolution note
300          mp4        384x216    
450          mp4        512x288    
700          mp4        768x432    
1000         mp4        960x540    
1200         mp4        1280x720   (best)
[download] Downloading video 3 of 3
[info] Available formats for e50957d0-ed00-11e0-aca6-0026b9414f30:
format code  extension  resolution note
300          mp4        384x216    
450          mp4        512x288    
700          mp4        768x432    
1000         mp4        960x540    
1200         mp4        1280x720   (best)

[dstftw]

Episodios en espanol will be available in the next version. Thanks for the report.

[dilworks]

Odd, while testing the episodes stream with English audio, not with Spanish audio as expected.

I'll need to perform more testing to figure out how to pick the proper audio track.

[dilworks]

Found something:
When the playlist XML is requested from the server, there is a parameter in the URL where a language code is expected. In my case, the URL turns to be this:

/feeds/player/mediagen?uri=mgid:arc:video:southparkstudios.com:e50791a2-ed00-11e0-aca6-0026b9414f30&aspectRatio=16:9&lang=en&acceptMethods=fms

As we can see, "en" means "English, no matter that you've entered through a Spanish page". The expected language code for Spanish episodes is "sp" (yes, not "es"). Changing the language code yields the correct playlist. However, I don't know where to poke in the code for fixing this...

[dstftw]

Too late. That's already implemented in a542e37 and e4a5e77.

[dilworks]

Any clues on how to use it then? I'm not familiar with the endless array of commandline options yet...

[dilworks]

Ah, OK, let me upgrade again, didn't noticed the full commit yet :D

[dstftw]

This has been fixed in youtube-dl 2015.04.26. See our FAQ if you have problems updating.