Brian McFee

> Thanks for noting this. In the short-term, should we at least enforce a < 3.8 requirement for matplotlib in setup.py? Yeah, I think that would make sense - both...

That's almost certainly true. Quick github search has 51 code hits: https://github.com/search?q=mir_eval.display&type=code and most of them are either our own (for appropriately defined "ours") or one-off code from a few...

Jotting down some thoughts here... The basic problem is that we want something that works essentially like `ax.plot` - in that subsequent calls automatically increment the style cycler, so that...

Gave this a bit more of a think, and decided we should avoid stashing custom stylers in the axes object. Instead, I think we can do a slightly less direct...

Congrats! Seems like there's a lot of that going around these days :grin: (my leave is just winding down now, hence the spare cycles before everyone else notices...) I can...

> or we could do something weird like call `validate` when an exception is raised in the metric function (to get a more helpful error message) I don't think that...

This is unfortunately a difficult question to answer, and the causes for deviation are threaded all throughout the computation stack, from BLAS up to the BSS eval algorithms. I'm having...

Aha, reference thread: https://github.com/craffel/mir_eval/issues/239

> but the devil's in the details, and we can't crack matlab open to see how the \ operator really works. Stack overflow to the rescue: apparently we CAN (sort...

https://www.semanticscholar.org/paper/A-Scape-Plot-Representation-for-Visualizing-of-M%C3%BCller-Jiang/62ea16ed458ffbc140dd9761a274a7e0e72c201e gives a good overview of the method. They're useful for multi level segment viz, which can include things like local->global key estimation.