NB! Found a bug where weights are (unintentionally) modified after the first feature plot. This seems to have implications for contour lines, including the equilibrium line. Need to dive a bit deeper into what happens. The latest commit per 18. feb implements a fix with a ".copy()" invocation, but we might want to revert this. The expected changes we need (but this needs verification):
- All current contour lines are calculated on the "doubly normalized" data (I believe). I don't think this is critical. We can adjust the contour levels and achieve a very similar plot, since the second normalization is just a linear transformation.
- Equilibrium lines will be moved. These are set based on specific thresholds and are expected to refer to values on the normalized data, and not on the doubly normalized data. Need to take extra care when handling this, so that we are sure this gets right.
- PC lines seems only marginally affected.
NB! Found a bug where weights are (unintentionally) modified after the first feature plot. This seems to have implications for contour lines, including the equilibrium line. Need to dive a bit deeper into what happens. The latest commit per 18. feb implements a fix with a ".copy()" invocation, but we might want to revert this. The expected changes we need (but this needs verification):