EDIT

implemented but no tests yet, but maybe you can integrate into controller already

please can you review my changes?

Have you tested with termType = Character which is a subtype for Feature. I don't think the "Restrictions" implementation will manage this correctly.

Andreas Müller wrote:

Have you tested with termType = Character which is a subtype for Feature. I don't think the "Restrictions" implementation will manage this correctly.

I did not understand this as requirement and I see two problems whith this:

1. From the TermTypes enum which is being used as filter it is not clear that using Feature will also include Character.
2. How would it be possible to get TermTrees for Feature without Character ?

Therefore I suggest to allow passing a Set of TermTypes as filter:

public List<TermTree> list(Set<TermType> termType, Integer limit, Integer start, List<OrderHint> orderHints, List<String> propertyPaths);


public Pager<TermTree> page(Set<TermType> termType, Integer pageSize, Integer pageIndex, List<OrderHint> orderHints, List<String> propertyPaths);

which is supported by the restrictions api out of the box.

Andreas Kohlbecker wrote:

Andreas Müller wrote:

Have you tested with termType = Character which is a subtype for Feature. I don't think the "Restrictions" implementation will manage this correctly.

I did not understand this as requirement and I see two problems whith this:

1. From the TermTypes enum which is being used as filter it is not clear that using Feature will also include Character.
2. How would it be possible to get TermTrees for Feature without Character ?

Therefore I suggest to allow passing a Set of TermTypes as filter:

public List<TermTree> list(Set<TermType> termType, Integer limit, Integer start, List<OrderHint> orderHints, List<String> propertyPaths);


public Pager<TermTree> page(Set<TermType> termType, Integer pageSize, Integer pageIndex, List<OrderHint> orderHints, List<String> propertyPaths);

which is supported by the restrictions api out of the box.

The javadoc says "Is subtype of {@link #Feature}" which to me is clear. Also the idea of the term types is to replace classes as classes are difficult to handle sometimes. But the semantics should be the same. Therefore if using list(class, ....) you also do not expect that you need to pass all the subclasses as a collection of classes. This is exactely the same behavior as for classes and it is by intention. Therefore I do not agree with the above solution.

Agreed ... and implemented in the way you pointed out.

Andreas Kohlbecker wrote:

Agreed ... and implemented in the way you pointed out.

The current implementation is not generic as my first implementation was. It handles Characters as an explicit case. My first implementation did handle ALL sub-types. Currently only this 1 subtype exists, but in future there might be more.

Also it might be nice to have a little test that it works. Have you tested in some environment?

my implementation now is generic.

Also it might be nice to have a little test that it works. Have you tested in some environment?

I positively tested manually with the cdm_additivity_ontology database. Unit-tests do not yet exist.