equinor / webviz-subsurface

Webviz-config plugins for subsurface data.
https://github.com/orgs/equinor/projects/24
GNU General Public License v3.0
57 stars 59 forks source link

Strange coloring of 4D maps with positive and negative values #311

Open ezaezaeza opened 4 years ago

ezaezaeza commented 4 years ago

I suspect that there is something wrong with the coloring of maps in the Surface viewer, when there are both positive and negative values in the map. The maps in the attachment should look similar, but in webvis, the coloring does not seem to be symmetric around 0. webvis_4D_map coviz_4D_map

asnyv commented 4 years ago

@ezaezaeza have you tried defining the max/min range in the yaml-config, similar to how it is done for stoiip in the example? See the attribute_settings below. Should be similar to the user defined range in the other figure you attached. Adding a symmetric: True option could of course be an idea, as the range isn't always well known in advance. And as you say, it could also be considered to be default behavior if both positive and negative values are present.

  - title: Surface viewer (FMU)
    content:
      - SurfaceViewerFMU:
          ensembles:
            - iter-0
            - iter-3
          wellfolder:
            ../observed_data/wells
          attribute_settings:
            stoiip:
              min: 0
              max: 2
              color: inferno
ezaezaeza commented 4 years ago

Thanks for the feedback. I have set the min and max in the config file ( to -.5, +,5) , but it does not solve the issue. But, yes, I think a symmetric: True option would be the solution here (this is how it is done for 4D maps in Rogers Jupyter notebook). I tried, but I guess it is not implemented in webviz yet...? It could be set as a default option when you have both positive and negative values, but there are also examples where you do not want ithe color scale to be symmetric around 0.

asnyv commented 4 years ago

No, symmetric is not a valid option today, but the min/max should have worked as far as I know 🤔 Can't see the bounds on the color scale from your figure since they are outside your snapshot, but I assume they are also not -0.05 to 0.05?

ezaezaeza commented 4 years ago

The color scale shows -0.05 to 0.05, so it is a bit strange.

image

asnyv commented 4 years ago

Interesting 🤔