Alessio Buccino

In this case, would you expect a `spike_amplitudes` folder with an empty numpy array? Or should the function just warn and return nothing?

I'm in favor of using logging everywhere! Thanks for bringing this up @yger

@bramn22 sorry for the super delay on this. We actually made quite some improvements to the motion/drift in more recent releases (>=0.98). Are you still experiencing issues?

ok, what is the output of `print(len(sorting.to_spike_vector()))`?

Ah I think we have a bug in the torch detection! Can you try with the `locally_exclusive`?

After discussing with @samuelgarcia we agree that we can be more "relaxed" about this, but we need to add a warning and use the `recording` sampling frequency as it is...

The only real application for this is `fast_templates|templates` (or viceversa). Maybe it would be better to only support this two with a specific function.

Guys, I'm reading this again.. So everything boils down to `get_extension("templates|fast_tamplates")`. I'm not seeing other use cases at the moment. My suggestion then is to handle this downstream, when needed....

> > I'm not seeing other use cases at the moment. > > For example, let's say in the future someone want to implement an extension `select_spikes` (similar to `random_spikes`,...

> analyzer.get_extension_data("templates|fast_templates") `analyzer.get_extension_data("templates") or analyzer.get_extension_data("fast_templates")`? Not a lot of extra typing