Base precision and recall upon a minimum similarity score?

[DallanQ]

I wonder if an alternate precision_recall function could be passed a minimum similarity threshold instead of a number of candidates. It seems like that could result in better scores, since if a tree name only has record names that aren't very similar, including the less-similar names would hurt precision.

[colonialjelly]

That's a good point, I can write a function to filter out candidates that don't meet a certain threshold.