2025.1 tests update, Excel ingestion fix, and SubjectPseudoIdentifier fix

• Main changes are related to the unit tests and they are not important to review
• I spotted a bug in the Excel ingestion which probably was introduced when we defined ReferenceInterpretationResult as a supporting concept. Due to the fact that Excel sheets can have only 31 characters max, there is an overlap with the concept ReferenceInterpretation. I've updated the logic to handle this case and analogous cases.
• There is an issue with the database schema as well. It is related to the sphn_SubjectPseudoIdentifier table. According to the schema, the property sphn_hasSharedIdentifier has cardinality 0..*. A case with cardinality higher than one cannot be inserted in the table as the primary key is defined by the id field only. I therefore removed the primary key on the table (alternative would be to add sphn_hasSharedIdentifier to the primary key). This leads to an issue when multiple values are defined because the JSON data for sphn:SubjectPseudoIdenfier is then a list and not an object anymore. I therefore added some logic after the database extraction to unify the multiple objects in the list into a single object so that we do not need to change the RML mapping rules.