fix(playlist): More robust handling of the passed objects.

[blacklight]

Apparently we can't always assume that the JSON object passed to parse_playlist always contains a data envelope.

Closes: #349

[blacklight]

@tehkillerbee let me know if this may be due to recent changes in the API layer or if there's a better way to handle this.

[blacklight]

Ok I've actually noticed that this is the same thing as this #345

@tehkillerbee not sure if you made some progress about this, but it looks like my workaround of fetching data and using the object itself as root as the fallback works. But I've just given it a try in high-tide and it seems that if I have large playlists I can't fetch more than 50 objects (not sure if it's an issue on our side though, or if it should now be the client to handle paginated requests).

[tehkillerbee]

not sure if you made some progress about this, but it looks like my workaround of fetching data and using the object itself as root as the fallback work

The v2 endpoints require no more than 50 objects per fetch. So pagination should be used for large lists. Anyways, I plan to add the threaded worker threads to tidalapi and then they can be used when fetching larger sets of playlists, similar to what we do in mopidy tidal.

This fix is still a good idea so lets merge it 👍