[python-pytrakt] Refactor: Add IdsMixin to replace extract_ids() utility method

[glensc]

Extracting the "ids" on each response is tedious, and such duplication results users being confused:

• UserList.add_items not working #78 (comment)
• AttributeError: 'Movie' object has no attribute 'trakt' #158 (comment)

Also, the existing extraction is inconsistent (Not applied for TVShow and Movie in this example):

• PyTrakt/trakt/sync.py

  Lines 331 to 342 in 8a6d4f1

  	for d in data:
  	if 'episode' in d:
  	from trakt.tv import TVEpisode
  	show = d.pop('show')
  	extract_ids(d['episode'])
  	results.append(TVEpisode(show.get('title', None), **d['episode']))
  	elif 'movie' in d:
  	from trakt.movies import Movie
  	results.append(Movie(**d.pop('movie')))
  	elif 'show' in d:
  	from trakt.tv import TVShow
  	results.append(TVShow(**d.pop('show')))

This will make the ids extraction automatic, and prevent updating the attributes directly.

Additionally it came out that properties like tmdb_id, imdb_id, trakt_id are unused. they are initialized as None in the constructor but never any other value. these are commented with @deprecated and are to be removed in 4.x.

• Person
• Movie
• Calendar
• SearchResult
• TVShow
• TVEpisode
• TVSeason
• UserList

TODO:

• rebase after [python-pytrakt] Cleanup: Drop classes extending from object #187

[glensc]

It's actually more confusing than that, seems the _id suffixed ones are abandoned:

• PyTrakt/trakt/people.py

  Line 17 in 8a6d4f1

  self.biography = self.birthplace = self.tmdb_id = self.birthday = None

as the ids property returns without the "_id":

• PyTrakt/trakt/people.py

  Lines 70 to 76 in 8a6d4f1

  	@property
  	def ids(self):
  	"""Accessor to the trakt, imdb, and tmdb ids, as well as the trakt.tv
  	slug
  	"""
  	return {'ids': {'trakt': self.trakt, 'slug': self.slug,
  	'imdb': self.imdb, 'tmdb': self.tmdb}}

[glensc]

Also, the existing extraction is inconsistent. not applied for TVShow and Movie in this example:

• PyTrakt/trakt/sync.py

  Lines 331 to 342 in 8a6d4f1

  	for d in data:
  	if 'episode' in d:
  	from trakt.tv import TVEpisode
  	show = d.pop('show')
  	extract_ids(d['episode'])
  	results.append(TVEpisode(show.get('title', None), **d['episode']))
  	elif 'movie' in d:
  	from trakt.movies import Movie
  	results.append(Movie(**d.pop('movie')))
  	elif 'show' in d:
  	from trakt.tv import TVShow
  	results.append(TVShow(**d.pop('show')))

[glensc]

Can't finish UserList as it uses namedtuple:

E   TypeError: Multiple inheritance with NamedTuple is not supported

• https://bugs.python.org/issue36517
• bpo-36517: Raise error on multiple inheritance with NamedTuple python/cpython#19363

[glensc]

A change is needed to namedtuple definitions (UserList):

• remove 'trakt', 'slug'
• add 'ids'

since namedtuple doesn't allow optional values, this would be a breaking change, meaning this needs to land in 4.x

[glensc]

This is ready now, but not sure if this breakage is big enough that this needs to go to 4.x?

• [python-pytrakt] Refactor: Add IdsMixin to replace extract_ids() utility method #186 (comment)

if you're not creating UserList(...dict...) yourself, you're not affected

[glensc]

External projects calling extract_ids(data) directly need to omit it after this PR.

[glensc]

Merging in my fork:

• Refactor: Add IdsMixin to replace extract_ids() utility method glensc/python-pytrakt#14