dynamically set map deme radii

[jameshadfield]

would prevent cases such as this:

[trvrb]

Possibly addressed by changes in #330? We should revisit and see if working as intended.

[jameshadfield]

Doing this for all the different use cases seems difficult to do perfectly, but relatively easy to make vast improvements over the current situation (right now it's hardcoded so only works for a small range of tip values).

We could also expose a user selector to tweak this as desired, e.g:

[tsibley]

What if we define a scaled set of maximum deme radii based on the current map zoom level, assign the largest demes to the zoom level max radii, and scale the rest of the smaller demes based on that?

[jameshadfield]

This would be a relatively easy first step and would probably be a big improvement.

[trvrb]

I think that PR #712 + a dynamically set max zoom (to be ~2 levels past initial zoom set by fitBounds) will fully resolve this. Initial sizing all works well, only looks bad when zoomed too far out. I think reasonable to add in a maxZoom and maxBounds to appropriately bound these more narrow datasets.

[jcw349]

Doing this for all the different use cases seems difficult to do perfectly, but relatively easy to make vast improvements over the current situation (right now it's hardcoded so only works for a small range of tip values).

We could also expose a user selector to tweak this as desired, e.g:

Are there plans to implement this still? I think it will help visualize local cases. As of now, when we hover our mouse over small demes, the mouse misses the circle and nothing happens.

[trvrb]

I'll leave this open, but it hasn't seemed a major issue to me. If others disagree, please chime in.

[emmahodcroft]

I wouldn't say it's disabling, but it can definitely still be annoying sometimes. For example, on focal S:N501 builds, UK circle just about swallows rest of map. For a figure I made recently, I had to run locally and adjust deme sizes in auspice just to get something semi-reasonable-looking. That's a bit of a special case, but it might be nice if we could allow users to somehow scale how radical the difference between demes are, for cases where difference may be really extreme. However, not urgent.

[trvrb]

Thanks Emma. Just bumped to "moderate priority".