ancestral branch thicknesses (below CA) should be minimal

[jameshadfield]

This issue may represent two independent problems, but I suspect they are highly related.

1. Upon filtering the data, all branches ancestral to the common ancestor of the selected strains should have minimum thickness. These branches are circled in the following screenshot.

2. Upon filtering the data, the tree should zoom in to the CA node - in other words, the same behavior as if one clicked on that branch.

Complication to 2: unsure what behavior is best when toggling selected filters as visible / invisible if that action were to modify the CA

---

Figure 1: https://nextstrain.org/zika?f_region=North%20America, displayed using auspice 2.20.1

[trvrb]

I considered (2) previously and even implemented it on a demo branch, but felt that it was too "jumpy" especially for things like date slider interactions and you'd lose your place. I think I'd still recommend a "zoom to selected" button in the tree panel to accomplish this. #1132 is related.

[emmahodcroft]

I agree with 1 but I also have concerns about 2. I think this could get really jumpy and I'm also not sure if it is always desired behaviour. Often (now that filters can turn on and off more easily) I'm flipping these on and off pretty fast for various countries in Europe (for example) and part of what I'm looking at is how they distribute across the tree, part of tree, or cluster that I've zoomed to. So if it was rezooming all the time this would be super annoying. (Ex: Do we have early samples from Finland in 19A? What about France? What about Spain? Are samples from Ireland in 'both branches' of 439 cluster? What about Switzerland?)

I'd second Trevor's idea about a 'zoom to selected' button, to make zooming to filtered easy, though!

[jameshadfield]

Thanks both! Sounds like the best course of action is to not implement (2) here, instead going with #1132

[jameshadfield]

@joverlee521 -- without diving into the code, my instinct would be that the visibility array is at fault here. As a bit of background:

• visibility is an array such that the visibility of node[i] is found at visibility[i]. It refers to whether a node is within the currently active (sub)tree - for instance all the "thin" branches in the figure above should have visibility[nodeIdx] !== NODE_VISIBLE -> true. The visibility of nodes above (ancestral to) the CA should be NODE_NOT_VISIBLE
• the inView property of a node is subtly different to visibility and refers to the nodes that are out of the current tree-panel view. This is relevant when zooming into the tree.
• Branch thicknesses are calculated using the visibility data, and are stored in a similar fashion, where node[i]'s branch thickness is found at branchThickness[i].

(This is all going off my memory, so take it with a grain of salt!)