Renaming: from_time_labels -> from_samples_labels

[alejoe91]

and adding from_times_labels

fixes #3723

[yger]

Looks good to me, and I agree it would be easier to understand

[zm711]

I think this is clutch for understandability. I think the other longer term thing we should think about (pre-1.0 in my opinion) is moving from the frame -> sample terminology in all places.

[h-mayorquin]

How comes the argument is still called times here, is this something we should change?

[h-mayorquin]

Two request:

1. Can we add a deprecation so we don't break code.
2. Can the method be from_samples_and_labels and from_times_and_labels

I find the lack of and very confusing in how I read the naming. As in, if the labels were the labels of the times.

As I am writing this I am thinking, aren't the labels the unit_ids? Shouldn't we use that terminology? from_samples_and_units_ids or something like that? I feel less strong about this last point but I feel strongly about the and.

[zm711]

I think those are great points Heberto! I was thinking something similar to point 2, but didn't want the double break. But I think you're right that if this breaking change is being done might as well as make it as clear as possible moving forward. I like the from_samples_and_labels but I take your point that in spikeinterface terminology we use unit_ids as our neuron "labels".

[samuelgarcia]

I like the idea but we need to be backward compatible here (with a warning)
So maybe the idea of @h-mayorquin is the good way and so we keep the old one.

stale