Consider changing the default value of filter_before. about trackpy HOT 3 CLOSED

It should default to 'None' and the engine does what is fastest for it?

from trackpy.

@nkeim and I are both finding filter_before slower in every case, even with numba off, which is a surprise to me. I am sure you can concoct a case where filter_before is worthwhile, but I'm not sure how rare/strange that data would have to be.

I like your idea of making None the default going forward. For now, we might have None become False internally for every case. If and when we learn more about how to optimize, how to recognize cases where True runs faster, we can make the internal logic of locate smarter without changing the API.

from trackpy.

Yes, that is the idea. There is a push to change most of the defaults in mpl to None for this exact reason.

It also enables you to tell if the user said 'do the default thing' or 'do the same thing as the default'

from trackpy.