TREEcg/treemunica

Implement actors that score based on non-literal data

Opened this issue · 0 comments

Related to #2, except that there's nothing blocking the implementation of these features at the moment. Specific additional criteria that could be used in scoring include the literal's language, or the used predicate.

We could either implement this like one of the existing actors, or incorporate it in a post-processing reranking step.