Make edge thickness consistent across layouts and visualizations #276
Labels
Comments
fedarko
added a commit
to fedarko/empress
that referenced
this issue
Jul 28, 2020
eventually thickness should be determined by assessing the "between- tip distance" (i.e. how much is y=1 expressed in terms of the scaled rect layout?), and using that to impact barplot positioning. for circular layout i imagine something similar with "how much is theta = 1?" or something vaguely like that. wait does that relate to biocore#276? probably.
|
As more motivation to improve this -- the current method for scaling line widths uses the log of the number of leaves in the tree as a denominator in the equation. A problem with this is that if there's only 1 leaf in the tree (as in the silly corner-case where the tree is a straight line from the root to the single tip), then
Having the layouts computed in JS will make recording the scaling factors used a lot easier, which should be the way to fix this. Hopefully. |
This was referenced Aug 15, 2020
fedarko
added a commit
to fedarko/empress
that referenced
this issue
Sep 21, 2020
Relevant to biocore#276. The reliance on max displacement (which doesn't seem to have an analogue in the unrooted layout? or maybe we can sort of treat it the same way as circ layout, idk) means that this might be too specific a solution to apply to line width thicknesses, but it should make using barplots a lot less painful
|
Also applies to:
|
This was referenced Nov 3, 2020
Merged
ElDeveloper
pushed a commit
that referenced
this issue
Nov 10, 2020
… barplot lengths look thicker; fix a small Colorer bug (#442) * ENH: Scale barplot lengths by max displacement Relevant to #276. The reliance on max displacement (which doesn't seem to have an analogue in the unrooted layout? or maybe we can sort of treat it the same way as circ layout, idk) means that this might be too specific a solution to apply to line width thicknesses, but it should make using barplots a lot less painful * BUG: precomp sm barplot lyr len&report correctly * Document layout factor stuff * add note re: 1.1 * BUG: Fix CLR_COL / "freebie" problem * Remove unneeded comments re: freebie thing * Test rgb2gl; prettify * improve rgb2gl docs * TST/MNT: Improve/test layout supported & freebies * Do away with CLR_COL_GL, and document stuff should address @ElDeveloper comment * BUG: Also apply barplot unit to border lens WHOOPS * TST: fix barplot border test
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
When coloring the tree and changing the width of the lines the thickness will be dependent on the tree, and the layout. It would be nice if this wasn't the case.
See #144 and #135 for more discussion.
The text was updated successfully, but these errors were encountered: