update project structure (comment on the olms2 branch init)
There are some decisions to make and, therefore, some discussions. Below are some questions I am asking myself:
- LinguisticRealisation
- Should we make it a specific implementation of a Data Container?
- It would mean assigning a UID, which might make sense to later export an RDF-based linguistic ontology
- Should corpus occurrences be provided when creating the LinguisticRealisation? Or should we define a method to fetch them?
- It might be worth creating a method since we could use it for regular updates.
- Since we define a class and not an abstract class, is the filename
..._schema
still relevant?
- Should we make it a specific implementation of a Data Container?
- Corpus occurrence
- Do we need two create two different classes, one for Concept and one for (Meta)Relations?
- Data Containers
- Do we provide a UID when creating the data container instance, or do we create one automatically?
- In the first case, we would allow assigning URIs, for instance
- In the second case, we would consider UIDs internal to the ontology learning process and consider IDs (or URIs) creation only when exporting the Knowledge Representation in a specific format.
- I would do the latter.
- Do we provide a UID when creating the data container instance, or do we create one automatically?