In the clustering widget (and potentially also other places), the selection in the measurements-list is reset and that's hard to see for the user.
To reproduce:
Load data (e.g. blobs), segment it, appyly measurements.
Use the dimensionality reduction widget, e.g. UMAP
Use the clustering widget:
Select UMAP0 and UMAP1 only
Cluster, e.g. K-Means
After clustering, all measurments will be selected. If the user wants to also use HDBSCAN, they first have to select UMAP0 and UMAP1 again - and that's not obvious.
Potential solution: When updating the measurements-list, keep the selection.
In the clustering widget (and potentially also other places), the selection in the measurements-list is reset and that's hard to see for the user.
To reproduce:
After clustering, all measurments will be selected. If the user wants to also use HDBSCAN, they first have to select UMAP0 and UMAP1 again - and that's not obvious.
Potential solution: When updating the measurements-list, keep the selection.