Manifold colors - Githubissues

champsproject / ldds

Python package for computing and visualizing Lagrangian Descriptors in Dynamical Systems

BSD 3-Clause "New" or "Revised" License

17 stars 2 forks source link

Manifold colors #20

Closed vkrajnak closed 3 years ago

vkrajnak commented 4 years ago

Red for unstable, blue for stable. Will require to give draw_lagrangian_descriptor forward and backward LDs separately instead of sum.

broncio123 commented 3 years ago

Maybe work on redefinition of colourmap parameters to achieve this.

@VikJGG plots do something like that and when overlaid they look great (pics below). However, is this colour scale tweaking manual?

Just a thought: One possible approach would be to take the LD values distribution and define vmin and vmax passed to plt.contourf as a function of the mean and std of LD values. I did something like this for the function draw_lagrangrian_descriptors, defining a new variable called colormap_mode that implements these changes of vmin/vmax if called with value 2.

vkrajnak commented 3 years ago

Maybe it would be enough to redefine the plotting function to accept LD_forward and LD_backward optionally and depending on which ones are provided, it will draw forward, backward or both on top of each other. This should also be relatively straight-forward.

broncio123 commented 3 years ago

Maybe it would be enough to redefine the plotting function to accept LD_forward and LD_backward optionally and depending on which ones are provided, it will draw forward, backward or both on top of each other. This should also be relatively straight-forward.

Nice, can you do this @vkrajnak , please? I think this will enhance the graphical output for the EXAMPLES notebook, both for the book and the package.

vkrajnak commented 3 years ago

I can put it on the waiting list

broncio123 commented 3 years ago

NOTE: Laplacian filter. Thanks, @VikJGG

vkrajnak commented 3 years ago

For filters, you opened issue #11. Let's keep them out of this issue

broncio123 commented 3 years ago

Maybe work on redefinition of colourmap parameters to achieve this.

@VikJGG plots do something like that and when overlaid they look great (pics below). However, is this colour scale tweaking manual?

Just a thought: One possible approach would be to take the LD values distribution and define vmin and vmax passed to plt.contourf as a function of the mean and std of LD values. I did something like this for the function draw_lagrangrian_descriptors, defining a new variable called colormap_mode that implements these changes of vmin/vmax if called with value 2.

IMPORTANT NOTE

For filters, you opened issue #11. Let's keep them out of this issue

Gradient- and Laplacian-based filters should be defined under a common umbrella. I will stick to the other open issue though atm.

vkrajnak commented 3 years ago

Addressed in pull request #30