Given a noun phrase (subject) and a verb (predicate), the sq.graph method should return the resulting concepts and their weights that appear as direct objects in the same sentences as the given subject and predicate.
This is likely due to how solr is being queried, and the available metadata for filtering. It is currently a two-step process:
1) get all the sentence_ids that have preflabel matching the noun phrase and are the subjectof the verb
2) get all the preflabels in those sentence_ids that are objectof the same verb
This should work in theory but is not returning all the expected concepts. This needs to be formally tested against manually annotated sentences.
Given a noun phrase (subject) and a verb (predicate), the sq.graph method should return the resulting concepts and their weights that appear as direct objects in the same sentences as the given subject and predicate.
This is likely due to how solr is being queried, and the available metadata for filtering. It is currently a two-step process: 1) get all the sentence_ids that have
preflabel
matching the noun phrase and are thesubjectof
the verb 2) get all thepreflabels
in those sentence_ids that areobjectof
the same verbThis should work in theory but is not returning all the expected concepts. This needs to be formally tested against manually annotated sentences.