Here's the kind of thing I'd like to be able to do for DistantSupervision instead of using the current main method. Something similar for FeatureGeneration would be good as well.
KnowledgeBase kb = new KnowledgeBase("knowledge-bases/kb-facts-train.tsv.gz", "knowledge-bases/entity-names-train.tsv.gz", "knowledge-bases/target-relations.tsv");
Corpus c = new Corpus("doc-entity-split/training-corpus.derby/");
DistantSupervision ds = new DistantSupervision(NERArgumentIdentification.getInstance(), NERSententialInstanceGeneration.getInstance(), NERRelationMatching(), negExFlag);
writeToFile("ner-train/training-instances.tsv.gz", ds.produceTrainingInstances(kb, c));
Here's the kind of thing I'd like to be able to do for DistantSupervision instead of using the current main method. Something similar for FeatureGeneration would be good as well.