Improvements in Glossaries module

[dlpzx]

Here is a list of improvements that can be currently be taken in the Glossaries module:

1. Implement server-side permissions for glossary operations
   This should be the first priority. For glossaries there are no permissions decorating the service. Which means that for example if a team owns glossary1, another team could delete glossary1. This is avoided in the UI but not programmatically.

2. Ensure business logic is in GlossariesService and not in db.repositories
   The service layer in the glossaries module offloads all the business logic to the GlossariesRepository. Instead we should execute the business logic in the service and use the repository exclusively for db operations.

3. Clearly split Glossaries and Catalog. At the moment in the backend there is a catalog module which is in fact handling all APIs for Glossaries which are not necessary the same. In the frontend we separate among Glossaries and Catalog.

4. Remove duplicated gql fields for Glossaries/Categories/Terms
   Categories, Terms and Glossaries share most of their gql types and input_types. We should reduce the lines of code and re-use a single input_type and single type for them. This is also applicable to the frontend query definition.

5. Simplify API calls and make them more straight-forward
   Secondly, the way the API calls are designed to get information from a glossary can be simplified. We currently define 3 APIs in the frontend:

• getGlossary
• getGlossaryTree - which in reality is getGlossary with a bunch of extra fields
• listAssociations - which in reality is getGlossaryTree - which in reality is getGlossary

getGlossary is doing too many things. We should break it down by responsibility:

• getGlossary - input: glossary.nodeUri, output: glossary db metadata + resolve_stats
• NEW `getGlossaryTree - input: glossary.nodeUri, output: tree object of categories, terms with their parentUri
• NEW listAssociations - input: glossary.nodeUri, output: paginated list of associations

6. Remove duplicated code
   Glossaries API calls return a great number of unused fields. First of all we should remove unused fields from both backend and frontend to improve performance. I am mostly referring to the Glossary object.