Diagrams for `dendrogram.compute` documentation

[ChrisBeaumont]

To explain things like delta with pictures

[astrofrog]

We need to think of how to best illustrate this. In the process of thinking about this, I realized that this diagram may be incorrect:

That is, the dark orange contours should technically touch, as should the lighter orange contours, right? That is, we are not drawing contours at equal steps then determining the hierarchical nesting of the contours - instead, a new structure is created at the point where two sub-structures meet (hence why I'm suggesting that the sub-structures in the above figure should be touching).

I guess we might want something similar to Figure 3 of Alyssa's Nature paper, but we don't really have an equivalent for the light gray lines in that plot - we just go from pixel to pixel (and delta in our case is something different).

[ChrisBeaumont]

An alternative statement: this diagram isn't wrong -- it just shows equal contour spacing, which doesn't necessarily intersect the dendrogram at the merger points. You could show the dendrogram, as well as the equally-spaced contour levels as horizontal lines, to illustrate this.