Add a clock for LFP data in `NeuropixelsV1eDataFrame`

[cjsha]

In its current implementation, the Clock member of NeuropixelsV1eDataFrame aligns only with the SpikeData member and not the LfpData member.

This ended up happening because the NP1 probe samples LFP data 12x slower than AP data - therefore the LfpData buffer is 12x shorter than the SpikeData buffer. Clock is sampled whenever SpikeData is sampled - there is no corresponding clock that aligns with LfpData.

One potential solution is to add another member to the NeuropixelsV1eDataFrame which is a clock that does align with LfpData.

[aacuevas]

It is true that the LFP in this case is a bit of an outlier in the sense that it is not produced every hardware frame, but every 12. The fact that the NeuropixelsV1eDataFrame.BufferSize property is limited to multiples of 12 to account for this and ensure every software frame includes at least one LFP is a proof of this deviation from the usual frames we deal with.

So while Clock is a representation of hardware frames, we should discuss if yet another field is required. This would just be one every 12 values of Clock.

An alternative solution is use Clock and a Bonsai operator that takes one every 12 elements of the array, and include that in the examples for clarity.

[jonnew]

I think this is wontfix. We just need to provide an example loading script that shows how to do this. I marked this as documentation so that it can be moved over the docs repo. After that, feel free to close.