Spike: make hovering calculate only useful data without any loss of data

Nobody has claimed this yet.

As part of fixing #396 one of the attempts was to limit the calculation of the hover points. Even though this approach was reverted, it is worthwhile to explore this separately in more depth. Specifically, the comments regarding hovering and the usefulness of the calculations should be revisited.

The following comments and `const` were temporarily found in **src/utils/useHoverPointIndices.ts** before being reverted.

/************************************************************************
```* When calculating hovered points in a graph:
*
* 1. The graph is divided into sections or "bins".
* Bigger graphs (width, height), have more bins than smaller graphs.
*
* 2. All of the data points are assigned to the bins.
* When comparing the same set of data points:
* - large graphs have many bins with fewer data points per bin
* ex: 50 data points = 10 bins x 5 data points each
*
* - small graphs have fewer bins with many data points per bin
* ex: 50 data points = 2 bins x 25 data points each
*
* 3. When hovering, a typical screen size (laptop or desktop) fits around 200 rows
* in a readable font size. Any additional rows calculated will not be viewable
* on the screen. So, any additional calculation is useless because no useful
* information for the user will be viewable due to screen size.
*
* 4. Therefore, any excessively large data set does a lot of useless work and
* adversely affects browser performance.
*
* 5. Limit the total number of points in each dimension and also each bin.
* Keep in mind the considerations in 2.
*/
const TOTAL_POINTS_PER_HOVER_DIMENSION = 100_000
```

Start by reading src/utils/useHoverPointIndices.ts and the context from issue #396. Review the reverted hovering comments and TOTAL_POINTS_PER_HOVER_DIMENSION assumption, then determine how to reduce unnecessary calculations without losing useful hover data. Done means the approach is defined and validated against the stated data-preservation goal.

Written by the indexing model from the issue text.