Amend the .nkg file to have LH and RH vertex numbers

[neukym]

As a result of #140, we need to add in different hexel numbers for both LH and RH.

Note that these
a) needn't be the same size
b) needn't have the same vertex numbers.

• @caiw - we should probably check that Kymata does actually use these numbers, and not the index so we are not make work for ourselves (I assume we do, but we should check).
• Revert these hacks: Amend the .nkg file to have LH and RH vertex numbers #141 (comment)
  • → Revert gridsearch vertex workaround #160
• Relies on HexelExpressionSets need to handle cases where L and R hemis have mismatched vertices #147.

[neukym]

in particular, `gridsearch/plain.py'

            hexels=channel_names[1,:10239], # TODO: HACK - FIX WITH ISSUE #141
            data_lh=log_pvalues_lh[:10239,], # TODO: HACK - FIX WITH ISSUE #141
            data_rh=log_pvalues_rh[:10239,], # TODO: HACK - FIX WITH ISSUE #141

Should be something like:

            hexels=[channel_names[:len(log_pvalues_lh)], len(log_pvalues_lh):],
            data_lh=log_pvalues_lh,
            data_rh=log_pvalues_rh,

[caiw]

I'm happy to take this one. I will take a little bit of doing because I wrongly thought this was a logical constraint — sorry about that.

[neukym]

👍 happily this isn't a really a blocker as my above hack forces the code to trim the data by a couple of vertices so they are equal. Let's chat about priorities at some point.