Open jGaboardi opened 2 days ago
Can it have something to do with normalization of scale? Only the scales in the 2 plots are different: on the left it goes from 0 to 3 (or so) and on the right it's much more fine grained (from 0.9 to 1.1), so i think the values are the same (if you look closely, the greyish stuff on the left - has values close to 1 - so does the blueish stuff on the right)
but this one: https://matplotlib.org/stable/api/_as_gen/matplotlib.colors.CenteredNorm.html doesn't seem to have changed from 3.8.x to 3.9 , so not sure where else the rescaling could be happening
Seems to maybe be some strange new behavior in
matplotlib==3.9.0
that is causing a CI failure. I tested this on bothgeopandas==1.0.0
andgeopandas==1.0.1
, both were successful onmatplotlib==3.8.4
and both failed onmatplotlib==3.9.0
. Very odd that actual values would differ (see legend).