Selection Tools
Overview
core.selection_tools makes the platform’s active selection visible and
usable. It bridges the gap between selecting rows in another panel and turning
those rows into a dataset for continued analysis.
What You Do
Create a selection using a scatter plot, image gallery, active-learning query or another plugin.
Open Selection Set to preview the selected records.
Move focus through the selected IDs while linked detail panels remain open.
Remove the selection when it is no longer needed, or materialise it as a new dataset.
Optionally make the new subset the active dataset and continue exploring it.
Mappings
The record_id mapping is optional. When unavailable, the plugin can fall
back to a dataframe index where the source supports it.
A stable record ID is strongly preferred because it survives sorting, filtered views and exchange with other plugins more reliably than an index position.
Panel
Selection Set
The panel provides:
the current selection count;
a preview of selected rows;
previous and next focus navigation within the set;
a clear-selection control;
controls for naming and creating a derived dataset;
an option to activate the new dataset immediately.
Changing focus through the panel does not alter the membership of the set.
Action
Create Dataset From Active Selection looks up the selected rows, registers a new dataset and can make it active. The action allows the same operation to be used by a workflow without opening the panel.
Typical Uses
Selection Tools is useful for:
turning a lassoed region into a working dataset;
inspecting an active-learning query one row at a time;
comparing selected images or spectra;
exporting or transforming a scientifically interesting subset.
Caution
Index fallback is less portable than a stable identifier. Map
record_id whenever possible, especially before saving or sharing a
selection-derived result.