When RDF data is upload to PROV Show, a selected strategy is applied to it before visualisation which either removes or adds RDF data to/from the graph to allow for certain types of visualisation. For example, the basic strategy adds wasGeneratedBy relationship whenever it encounters generated relationships (i.e. A generated B becomes B wasGeneratedBy A with the two properties being defined as inverse in PROV-O) so that all arrows in a visualised node network graph point backwards in time.
Where objects do not have an rdfs:label property, a label is made from it's URI. This applied to blank nodes too.
The actual code used for each strategy is available in GitHub repository.
The following strategies are currently implemented with more on the way:
This basic strategy is the default. It aims to represent simple Entity, Activity & Agent classes of object with basic relationships only. It performs the following actions on uploaded data:
Entity, Activity & Agent and treat them as just Entity, Activity & Agent objectsActivity classes between any Entities that indicate derivation
<B> wasDerivedFrom <A> is indicated, the following RDF is added:
<B> wasGeneratedBy <AnonymousActivity><AnonymousActivity> used <A>Entity/Activity chains everywherewasGeneratedBy inverse properties for all generated properties
This strategy removes everything from uploaded RDF data other than Entities and builds relationships between Entities where it can reason about them from other relationships. Specifically it will:
Entity and treat them as just as Entitiesused or wasGeneratedBy Activities or wasAttributedTo Agents as EntitiesEntitywasDerivedFrom relationships between Entities that are indicated as being related through Activities
<B> wasGeneratedBy <AnonymousActivity> , <AnonymousActivity> used <A> ⟶ <B> wasDerivedFrom <A>