def apply(self, block: Block, **kwargs ) -> Optional[xa.DataArray]:
if (self.learned_mapping is None) or (self.learned_mapping.embedding is None):
Task.taskNotAvailable( "Workflow violation", "Must learn a classication before it can be applied", **kwargs )
return None
point_data: xa.DataArray = block.getPointData( **kwargs )
embedding: np.ndarray = self.learned_mapping.transform( point_data )
return self.wrap_embedding( point_data.coords['samples'], embedding )
def build_model(self, *args, **kwargs):
if self.block is None:
Task.taskNotAvailable( "Workflow violation", "Must load a block first", **kwargs )
else:
labels: xa.DataArray = self.getExtendedLabelPoints()
umapManager.embed( self.block, labels, **kwargs )
self.plot_markers_volume()
def learn_classification( self, **kwargs ):
if self.block is None:
Task.taskNotAvailable( "Workflow violation", "Must load a block and spread some labels first", **kwargs )
else:
full_labels: xa.DataArray = self.getExtendedLabelPoints()
print( f"Learning Classification, labels shape = {full_labels.shape}, nLabels = {np.count_nonzero( full_labels > 0 )}")
event = dict(event="classify", type="learn", data=self.block, labels=full_labels )
self.submitEvent( event, EventMode.Gui )