search

ubermag / discretisedfield

Python package for the analysis and visualisation of finite-difference fields.
http://ubermag.github.io
BSD 3-Clause "New" or "Revised" License
18 stars 13 forks source link

Make divergent colormap the default one? #32

Closed marijanbeg closed 5 years ago

marijanbeg commented 7 years ago

viridis is the default colormap. For magnetisation (out-of-plane component), divergent colormap might be better.

marijanbeg commented 7 years ago

The best options would be coolwarm, bwr, and seismic colormaps. However, these colormaps become very misleading if the range of data is not full [-1, 1] which is most often the case in micromagnetics.

fangohr commented 7 years ago

Can we have an example of using different (user specified?) colormaps in the notebook documentation? As long as people know how to use the colormaps, they can then do what they like and deviate from the default. I would like this directly back to matplotlib's colormap handling, so that we don't introduce an extra layer of code.

We could include the colormaps you mention above as examples, of course.

Regarding the data range: I agree with your concern. It might be sensible to fix the data range at -1 and 1 for the normalised magnetisation for some plots though? Again, if we show an example of how to do that (and how not to do that), we give the users all power and flexibility.

marijanbeg commented 5 years ago

For simplicity, I would prefer not to have the magnetisation range fixed. After some testing, this opens a lot of questions if there are two different Ms in the system. Colormap can always be chosen by the user and this is going to be documented.