Lucene FST linking engine does not check if FST models are available in canEnhance(..)

EnhancementEngines need to check in the canEnhance(..) method if they can enhance a ContentItem.

While the Lucene FST Linking engine does check for the language and the required NLP analysis results it fails to check if an fitting FST model is present.

Because of that the Constructor of the LinkingSession throws an IllegalStateException as it expects such calls to be already rejected by the Engine.

The fix is that the canEnhance checks if the configuration provides a FST model for the language OR a FST model for the default matching lanugage. If both is not available it needs to return EnhancementEngine.CANNOT_ENHANCE.