[Feature] Add support for extracellular electrophysiology probe geometries in 1D tracing

[adamltyson]

brainglobe-segmentation is agnostic to the feature being segmented and analysed. It should continue to be possible to analyse any structure, but there's no reason why we can't improve the features for specific use cases. One of these is the analysis of implanted extracellular probe trajectories. We should:

• Allow the user to choose a specific probe geometry (ideally from ProbeInterface)
• Analyse not just arbitrary positions along the probe, but the coordinate (in real space) and brain region of every recording site (or possibly specific chosen sites, in the case of e.g. Neuropixels)
• Export to standard formats (e.g. NWB - https://github.com/catalystneuro/ndx-anatomical-localization)
• Document the above
• Update the probe segmentation tutorial

FYI @bendichter, @h-mayorquin

[adamltyson]

Might be worth also:

• Documenting the specification of this output
• Including this specification somehow in the file itself so it can "stand alone" in downstream applications

[adamltyson]

Probably requires:

• Neuropixel probe SpikeInterface/probeinterface_library#4 for the most relevant probes
• where is the "tip" of a probe? SpikeInterface/probeinterface#299 to get the physical end of the probe

Currently probeinterface doesn't have all the info we need to integrate with brainglobe-segmentation. It's probably not worth reinventing the wheel, so I think this feature is on hold until the information we need is in probeinterface.