Add support for NPX 2.0 single-shank probes

[bparks13]

This PR supersedes #493.

Adding support for single-shank Neuropixels 2.0 probes, with the additional functionality of allowing the user to change the probe type in the GUI.

Note that due to the restructuring of the abstract/concrete classes, in the XML file the name of the property is ProbeConfiguration for the quad-shank probe (this is simply the name of the property, and is what all older versions of the files will be named), but for the single-shank probe it is renamed to SingleShankProbeConfiguration. We could keep the same name (ProbeConfiguration) and add the type name in the XML tag (i.e., xsi:type="onix1:SingleShank"), but this will break backward compatibility, since the deserializer cannot know which concrete type to use for the probe configuration property when loading old workflows. If we want to break compatibility for this, we can inform users that they can add the line xsi:type="onix1:QuadShank" to the ProbeConfiguration properties to let it load, but it seemed easier to maintain backward compatibility.

The text above is now outdated, with the most recent commits there is now a non-browsable, non-externalizable property that is used to handle serialization, leveraging the attributes to automatically add the type; this is called ConfigurationA/B. Additionally, the original ProbeConfigurationA/B properties are not serialized, but they are deserialized into a quad-shank probe configuration, allowing for backward compatibility.

@anjaldoshi We do not have any single-shank probes to test with, could you build this branch and run some recordings with a single-shank probe to make sure it records data correctly?

Fixes #446

[bparks13]

Before merging, make sure that the question of channel numbers is answered; there are differences in the documentation, between the given equations and the spreadsheet with the channel numbers for each electrode. We want to make sure that we have the correct channel map before adding this functionality.

[bparks13]

Channel mapping is now corrected, as per confirmed from Josh and Jan. The documentation contains the wrong equations, but the electrode channel mapping spreadsheet has the correct values; the equations now implemented return the correct channel number for each electrode, same as Neuropixels-PXI plugin.