Children that are only due to logical definitions should have special handling

[csnyulas]

The "Get children" API call and "children_XYZ" small files should not contain classes that are children of a given class only due to logical definitions (and not as direct, asserted, subclasses).

The old iCAT did not show children due to logical definitions in the class hierarchy, and the ICD browser does not show them either, which helps the users to have a clearer view of the asserted hierarchy.

while the new iCAT-X does show them, like in case of the Amoebic cystitis in the "WHOFIC Foundation Post Beta" ontology:

While in iCAT-X we provide already special handling for these "pure" logical definition parents, as described in issues #241 (with some problems pending #289), we don't have a similar handling for the children.

Some ideas to handle the problem:
We could leave them in the UI, but render them distinctively (e.g. in italic and in bracket).
We should not return them as children in the API calls.

[alexsilaghi]

for the icatx-api-gateway i've fixed the children in protegeproject/webprotege-backend-service#146

[alexsilaghi]

ensuring children uniqueness in protegeproject/webprotege-backend-service@83e109c

[csnyulas]

Moving this to Milestone 9.1, as the part with "Get Children" API call, which was critical for Milestone 8 was implemented by @alexsilaghi, while the UI part is a bit more complicated, and less critical, and will be handled by @matthewhorridge .