We are struggling quite a bit with widget signals in Text. Currently, widgets look at signal type (Table, Corpus, Distances, etc.) and take the first suitable input. However, we'd often like to be more specific.
I.e. Ontology outputs Words, which is a Table. Connecting it to Semantic Viewer first, results in Words being used as a Corpus instead of Words.
Preferably, one could define a more specific input type, so when Words are on the input, they would be used as Words, not Corpus. This would currently be most useful for Text, but surely there could be other use cases.
We are struggling quite a bit with widget signals in Text. Currently, widgets look at signal type (Table, Corpus, Distances, etc.) and take the first suitable input. However, we'd often like to be more specific. I.e. Ontology outputs Words, which is a Table. Connecting it to Semantic Viewer first, results in Words being used as a Corpus instead of Words.
Preferably, one could define a more specific input type, so when Words are on the input, they would be used as Words, not Corpus. This would currently be most useful for Text, but surely there could be other use cases.