fix: Subsurface viewer - wells layer separate log layers

[Anders2303]

In preparation for the data model change for the wells-layer (#2661 ), I did some preemptive refactoring and cleanup in the wells-layer to make it easier to work with. Most notably, I separated the logic for visualizing Well-log curves along the track as a dedicated sub-layer. This also fixes assorted bugs I discovered with the log layer as I was working on it.

Changes

• Halves the size of the wellsLayer file; many of the local utilities have been moved to the layer's utility files
• Added the LogCurveLayer class; this is now used as a sub-layer to handle log curve visualizations
  • I think the wellslayer's properties that related to the log layer are terribly unintuitive, so the log layer has different property names that I think are clearer. We might want to consider using these in the wells layer as-well, and deprecating the old properties....
  • NOTE: I also moved the wells-layer's "selection" layer into the new LogCurveLayer, as I understood this to be intended as a way of highlighting log tracks specifically; this is technically a regression, as the new selection layer is not able to highlight outside of it's related logs MD-domain
• Fixed the selection layer not being able to follow log curve width if logRadius was set to 0
• Fixes layer auto-highlights not always picking paired logs/trajectories
• Made the well-log layer better align with the trajectory path. In short terms, there are now interpolated points that match the trajectories points, so the log curve has the same curvature. It looks nice in 3D now too!
• Fixed the MapAndWellLogViewer storybook example, clicking wells, and domain selection now works

[hkfb]

LGTM, only skimmed most of the moved code.

[hkfb]

I guess if you switch these two conditions around, then you wouldn't need the @ts-expect-error assertion?