Right now, the UI selects only based on the "method" column which contains the K used for the segmentation. Instead, the "method" should just be "GeoStaS" and then the UI select would have only one entry in a "method" section and another UI list with the K. Something like:
Left UI list: Method
Right UI list: Domain count
No segmentation
2
Chainsaw
3
GeoStaS K-means (selected)
4
GeoStaS Hierarchical
5
Merizo
6
7
...
When the user clicks on "Chainsaw", the only number in the right list will be the domain count it produces. When they click on one of the "GeoStaS" options, they get to see all the ndom values for that method type.
"No segmentation" will only have "1" as the domain count. It could be put first as a default, or left last, since if the user didn't want segmentation, they wouldn't have added a file for it.
To do this, the snakemake script that generates it should be updated to remove the K= from the "method" section and the blender UI should have two separate lists with the second one depending on a .selected_segmentation_method string.
Right now, the UI selects only based on the "method" column which contains the K used for the segmentation. Instead, the "method" should just be "GeoStaS" and then the UI select would have only one entry in a "method" section and another UI list with the K. Something like:
When the user clicks on "Chainsaw", the only number in the right list will be the domain count it produces. When they click on one of the "GeoStaS" options, they get to see all the
ndomvalues for that method type."No segmentation" will only have "1" as the domain count. It could be put first as a default, or left last, since if the user didn't want segmentation, they wouldn't have added a file for it.
To do this, the snakemake script that generates it should be updated to remove the
K=from the "method" section and the blender UI should have two separate lists with the second one depending on a.selected_segmentation_methodstring.