Skip to content
James Baker edited this page Apr 27, 2017 · 2 revisions

This page contains a list of major features and changes that we may want to consider for future versions of Baleen. Please contribute any ideas to this list (inclusion on the list is not a guarantee that it will be included in Baleen).

Branching pipelines

The ability to split pipelines and run different sets of annotators based on previous annotators and the content of the document. For instance, if one annotator determined the document type of each document we might then want to run different sets of annotators based on that type.

Extensible properties on entity types

If annotators could add their own properties to entity types as required, rather than adhering to the strict type system then this would give us a lot more freedom and allow us to simplify and standardise the entity type system (although possibly make it more difficult to exploit the outputs). This could be achieved with, for example, a properties map on all entity types. However, maps aren't currently supported by UIMA 2.x (though might be in UIMA 3.0+).