ActiveFedora aggregations - How to and Best Practices

Table of Contents

Setup

How to make an object an activefedora-aggregation?

      aggregates :members, :predicate => RDFVocabularies::PCDMTerms.hasMember, :class_name => "ActiveFedora::Base"

How to set the rdf type?

      type RDFVocabularies::PCDMTerms.Collection

Questions: Is there a way to append another type using multiple statements? Something like...

      type RDFVocabularies::PCDMTerms.Object
      type RDFVocabularies::WorksTerms.GenericWork

Behaviors

Append to an aggregation.

coll1 << coll2
coll1 << obj1

obj1 << obj2

obj1 << file1  # NOT AN AGGREGATION, but contains for files should be similar

Append and multiple proxies for the same member

coll1 << coll2   # <- [coll2]
coll1 << coll3   # <- [coll2,coll3]
coll1 << coll4   # <- [coll2,coll3,coll4]
coll1 << coll2   # <- [coll3,coll4,coll2]  OR RAISE ERROR because coll2 is already in coll1

coll1 << obj1   # <- [obj1]
coll1 << obj2   # <- [obj1,obj2]
coll1 << obj3   # <- [obj1,obj2,obj3]
coll1 << obj2   # <- [obj1,obj2,obj3,obj2]  - Object can be repeated in an aggregation

# again, not an aggregation, but similar behaviors
obj1 << file1   # <- [file1]
obj1 << file2   # <- [file1,file2]
obj1 << file1   # <- [file1,file2] OR [file1,file2,file1]  ??? Does this replace file 1 ???

Set all

coll1.collections = [coll2,coll3,coll4]        # does there need to be a check to disallow repeats in this array?
coll1.collections = [coll2,coll3,coll4,coll2]  # RAISE ERROR?  because of repeat of coll2

coll1.objects = [obj1,obj2,obj3]
coll1.objects = [obj1,obj2,obj3,obj2]          # repeat of obj2 is OK

It is less efficient to set all instead of appending. Currently, set all will rewrite all the proxies.

Insert at

TBA - Is there a way to do this? What would it look like?

Reordering

TBA - what method would be called and how would the insertion point be identified?

Delete one from an aggregation

TBA - what method would be called and how would the member to delete be identified?