RootInterative figure transformation

[miranov25]

Goal:

In the current RootInteractive we can use custom data transformations and present transformed data. It is actively used, for example for logarithmic, sqrt, power, and rotation transformation of histogram content using a parameterized data source. Such an approach requires a high programming effort not only for the transformation of the data but also for the corresponding intervals (error bars, confidence intervals).

Better support and simplification of usaga can be achieved by centrally supporting the transformation of figures and axes. Transform figures instead of the data source. The goal is to add support for creating custom axis transformations via custom transformation functions (monotonic, differentiable)

Transformations:

x* ->Tx(x)
y* -> Ty(y)
(x*,y*)-> Txy(x,y)  

Requirements/interface:

• Transformations are attributes of the figure
• All elements belonging to the figure are transformed (markers, lines, errors, confidence intervals)
• List of transformation functions should be possible to define on the server (python) or directly o the client (javascript)
• Transformation should be parametrizible. E.g:
  • rotation angle
  • logarithm (base)
  • power (base)

Inspiration/convention to check:

• bokeh
• https://matplotlib.org/stable/tutorials/advanced/transforms_tutorial.html

Compared to other packages, our package should support interactivity and custom transformation (e.g. as text input on the client)

[miranov25]

Axis description should be a string similar to the hints e.g. {transform} nameOfvariable {transform} length(cm)

[miranov25]

Summary

missing part z transform