The metadata app contains the generic metadata system used in LASS, which allows objects in the LASS model set to contain key-value stores of textual, image-based and other formats of metadata by inheriting a mixin and providing a subclass of the standard metadata models.
The metadata app is dedicated to the LASS metadata system, which allows various different strands of data (generally text, but also image-based and other formats) to be attached to items.
LASS uses the metadata system, for example, to provide shows with names and descriptions that have full recorded history and hooks for an approval system. It is also used to associate images (thumbnails and player insets) with podcasts.
Each model can have zero or more strands of metadata attached to it. Each strand is its own model (see below for more information on how to create a metadata provider model), and represents a specific collection of metadata on objects in the subject model.
Strands are indexed by name; an entire strand (as a dictionary-like object) can be retrieved from an implementor of MetadataSubjectMixin with object.metadata()['strand-name']. Generally, there will be a text strand containing all textual metadata, and an images strand containing thumbnail images and other related pictorial metadata.
All metadata strands are implemented as key-value stores, the key store being implemented as one unified model for simplicity reasons and the value stores being separate for each strand for each model.
Two classes (metadata.mixins.MetadataSubjectMixin and metadata.models.GenericMetadata) provide the core framework for defining a metadata subject and a metadata strand. There are descendents of GenericMetadata available for specific commonly used strand types.
In the LASS project, examples of how to use the metadata system can be found in schedule.models.show, uryplayer.models.podcast and people.models.role.