Test spikeinterface-like behavior for get_video with one element

[h-mayorquin]

This is something important for the chunk iterator that is being developed by @weiglszonja. Note that this should be merged after #180 as it has it as a base.

[CodyCBakerPhD]

Great, anything else to add?

[h-mayorquin]

Great, anything else to add?

@CodyCBakerPhD @weiglszonja
I don't know if you care about the case when start_frame == end frame, in that case spikeinterface returns an array of shape (0, num_channels). Is that a edge case of the iterator you are building? What do you think?

[CodyCBakerPhD]

I don't know if you care about the case when start_frame == end frame, in that case spikeinterface returns an array of shape (0, num_channels). Is that a edge case of the iterator you are building? What do you think?

The slicing semantics of the GenericDataChunkIterator should not have an issue with endpoints like that but good to think of...

So in SpikeInterface, if you want to get all channel data for a single frame, you need to slice from start_frame to start_frame+1? And to access the final frame that way you would need to do from final_frame to final_frame+1?

[h-mayorquin]

So in SpikeInterface, if you want to get all channel data for a single frame, you need to slice from start_frame to start_frame+1? And to access the final frame that way you would need to do from final_frame to final_frame+1?

start_frame=0, end_frame=1 -> you get the first frame
start_frame=last_frame - 1, end_frame=last_frame -> you get the last frame