Right now generate_annotation_similarity_graph has a switch statement to select which method is actually used for calculating the similarity overlap, and that switch statement is inside the calc_similarity function that is defined locally.
I'm curious if we move that switch statement outside of calc_similarity and instead pass it in as an argument, will we see a speed improvement? Worth trying.
Right now
generate_annotation_similarity_graphhas a switch statement to select which method is actually used for calculating the similarity overlap, and that switch statement is inside thecalc_similarityfunction that is defined locally.I'm curious if we move that switch statement outside of
calc_similarityand instead pass it in as an argument, will we see a speed improvement? Worth trying.