def testAppletDeserialization(self):
g = Graph()
op = OpFake(graph=g)
dataExportApplet = TrackingBaseDataExportApplet(
op, "Tracking Result Export")
dataExportSerializer = dataExportApplet.dataSerializers[0]
with h5py.File(self.TEST_PROJECT_FILE) as testProject:
# serialize TrackingBaseDataExportApplet's topLevelOperator
opTrackingBaseDataExport = dataExportApplet.topLevelOperator
opTrackingBaseDataExport.SelectedPlugin.setValue('Fiji-MaMuT')
opTrackingBaseDataExport.SelectedExportSource.setValue('Plugin')
opTrackingBaseDataExport.AdditionalPluginArguments.setValue(
{'bdvFilePath': '/tmp/bdv.xml'})
dataExportSerializer.serializeToHdf5(testProject,
self.TEST_PROJECT_FILE)
# create new instance of TrackingBaseDataExportApplet and deserialize its topLevelOperator
dataExportApplet = TrackingBaseDataExportApplet(
op, "Tracking Result Export")
dataExportSerializer = dataExportApplet.dataSerializers[0]
dataExportSerializer.deserializeFromHdf5(testProject,
self.TEST_PROJECT_FILE)
# check deserialized values in applet's topLevelOperator
assert dataExportSerializer.topLevelOperator.SelectedPlugin.value == 'Fiji-MaMuT'
assert dataExportSerializer.topLevelOperator.SelectedExportSource.value == 'Plugin'
assert dataExportSerializer.topLevelOperator.AdditionalPluginArguments.value[
'bdvFilePath'] == '/tmp/bdv.xml'
def __init__(self, shell, headless, workflow_cmdline_args,
project_creation_args, *args, **kwargs):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(ManualTrackingWorkflow, self).__init__(shell,
headless,
workflow_cmdline_args,
project_creation_args,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'\
'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
opDataSelection.DatasetRoles.setValue(['Raw Data', 'Prediction Maps'])
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
self.objectExtractionApplet = ObjectExtractionApplet(
name="Object Feature Computation",
workflow=self,
interactive=False)
self.trackingApplet = ManualTrackingApplet(workflow=self)
self.default_export_filename = '{dataset_dir}/{nickname}-exported_data.csv'
self.dataExportApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_export_filename,
)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
['Manual Tracking', 'Object Identities'])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
#self.dataExportApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def __init__(self, shell, headless, workflow_cmdline_args,
project_creation_args, *args, **kwargs):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(ManualTrackingWorkflow, self).__init__(shell,
headless,
workflow_cmdline_args,
project_creation_args,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'\
'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
batchDataGui=False,
force5d=True,
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
opDataSelection.DatasetRoles.setValue(['Raw Data', 'Prediction Maps'])
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
self.objectExtractionApplet = ObjectExtractionApplet(
name="Object Feature Computation",
workflow=self,
interactive=False)
self.trackingApplet = ManualTrackingApplet(workflow=self)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet = TrackingBaseDataExportApplet(
self, "Tracking Result Export")
self.dataExportApplet.set_exporting_operator(opTracking)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
['Manual Tracking', 'Object Identities'])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
self._applets = []
self._applets.append(self.dataSelectionApplet)
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def testAppletSerialization(self):
g = Graph()
op = OpFake(graph=g)
dataExportApplet = TrackingBaseDataExportApplet(
op, "Tracking Result Export")
dataExportSerializer = dataExportApplet.dataSerializers[0]
with h5py.File(self.TEST_PROJECT_FILE) as testProject:
opTrackingBaseDataExport = dataExportApplet.topLevelOperator
opTrackingBaseDataExport.SelectedPlugin.setValue('Fiji-MaMuT')
opTrackingBaseDataExport.SelectedExportSource.setValue('Plugin')
opTrackingBaseDataExport.AdditionalPluginArguments.setValue(
{'bdvFilePath': '/tmp/bdv.xml'})
dataExportSerializer.serializeToHdf5(testProject,
self.TEST_PROJECT_FILE)
# check serialized values
with h5py.File(self.TEST_PROJECT_FILE) as testProject:
assert testProject[
"Tracking Result Export/SelectedPlugin"].value.decode(
) == 'Fiji-MaMuT'
assert testProject[
"Tracking Result Export/SelectedExportSource"].value.decode(
) == 'Plugin'
assert testProject[
"Tracking Result Export/AdditionalPluginArguments/bdvFilePath"].value.decode(
) == '/tmp/bdv.xml'
def data_export_applet():
op = Operator(graph=Graph())
dataExportApplet = TrackingBaseDataExportApplet(op,
"Tracking Result Export")
opTrackingBaseDataExport = dataExportApplet.topLevelOperator
opTrackingBaseDataExport.SelectedPlugin.setValue("Fiji-MaMuT")
opTrackingBaseDataExport.SelectedExportSource.setValue("Plugin")
opTrackingBaseDataExport.AdditionalPluginArguments.setValue(
{"bdvFilePath": "/tmp/bdv.xml"})
return dataExportApplet
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(ManualTrackingWorkflow, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'\
'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,
"Threshold and Size Filter",
"ThresholdTwoLevels" )
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",
workflow=self, interactive=False)
self.trackingApplet = ManualTrackingApplet( workflow=self )
self.default_export_filename = '{dataset_dir}/{nickname}-exported_data.csv'
self.dataExportApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_export_filename,
)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue( ['Manual Tracking', 'Object Identities'] )
opDataExport.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
#self.dataExportApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def test_applet_deserialization(project_path, data_export_applet):
with h5py.File(project_path, "w") as project_file:
data_export_applet.dataSerializers[0].serializeToHdf5(
project_file, project_path)
op = Operator(graph=Graph())
data_export_applet = TrackingBaseDataExportApplet(
op, "Tracking Result Export")
with h5py.File(project_path, "r") as project_file:
data_export_applet.dataSerializers[0].deserializeFromHdf5(
project_file, project_path)
opTrackingBaseDataExport = data_export_applet.dataSerializers[
0].topLevelOperator
assert opTrackingBaseDataExport.SelectedPlugin.value == "Fiji-MaMuT"
assert opTrackingBaseDataExport.SelectedExportSource.value == "Plugin"
assert opTrackingBaseDataExport.AdditionalPluginArguments.value[
"bdvFilePath"] == "/tmp/bdv.xml"
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(ManualTrackingWorkflow, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'\
'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
force5d=True,
instructionText=data_instructions,
max_lanes=1
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,
"Threshold and Size Filter",
"ThresholdTwoLevels" )
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",
workflow=self, interactive=False)
self.trackingApplet = ManualTrackingApplet( workflow=self )
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export")
self.dataExportApplet.set_exporting_operator(opTracking)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue( ['Manual Tracking', 'Object Identities'] )
opDataExport.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self._applets = []
self._applets.append(self.dataSelectionApplet)
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Variables to store division and cell classifiers to prevent retraining every-time batch processing runs
self.stored_division_classifier = None
self.stored_cell_classifier = None
## Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=None
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Segmentation Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,
"Threshold and Size Filter",
"ThresholdTwoLevels" )
self.objectExtractionApplet = TrackingFeatureExtractionApplet(workflow=self, interactive=False,
name="Object Feature Computation")
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
self.divisionDetectionApplet = self._createDivisionDetectionApplet(configConservation.selectedFeaturesDiv) # Might be None
if self.divisionDetectionApplet:
feature_dict_division = {}
feature_dict_division[config.features_division_name] = { name: {} for name in config.division_features }
opObjectExtraction.FeatureNamesDivision.setValue(feature_dict_division)
selected_features_div = {}
for plugin_name in list(config.selected_features_division.keys()):
selected_features_div[plugin_name] = { name: {} for name in config.selected_features_division[plugin_name] }
# FIXME: do not hard code this
for name in [ 'SquaredDistances_' + str(i) for i in range(config.n_best_successors) ]:
selected_features_div[config.features_division_name][name] = {}
opDivisionDetection = self.divisionDetectionApplet.topLevelOperator
opDivisionDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivisionDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivisionDetection.AllowDeleteLabels.setValue(False)
opDivisionDetection.AllowAddLabel.setValue(False)
opDivisionDetection.EnableLabelTransfer.setValue(False)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount)
selected_features_objectcount = {}
for plugin_name in list(config.selected_features_objectcount.keys()):
selected_features_objectcount[plugin_name] = { name: {} for name in config.selected_features_objectcount[plugin_name] }
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount)
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
self.trackingApplet = ConservationTrackingApplet( workflow=self )
self.default_export_filename = '{dataset_dir}/{nickname}-exported_data.csv'
self.dataExportApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_export_filename,
pluginExportFunc=self._pluginExportFunc,
)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue( ['Object-Identities', 'Tracking-Result', 'Merger-Result'] )
opDataExport.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.prepare_lane_for_export = self.prepare_lane_for_export
# configure export settings
# settings = {'file path': self.default_export_filename, 'compression': {}, 'file type': 'csv'}
# selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
# opTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
if self.divisionDetectionApplet:
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(self, "Batch Processing", self.dataSelectionApplet, self.dataExportApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
self._applets.append(self.batchProcessingApplet)
# Parse export and batch command-line arguments for headless mode
if workflow_cmdline_args:
self._data_export_args, unused_args = self.dataExportApplet.parse_known_cmdline_args( workflow_cmdline_args )
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args( workflow_cmdline_args )
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
if unused_args:
logger.warning("Unused command-line args: {}".format( unused_args ))
class ManualTrackingWorkflow( Workflow ):
workflowName = "Manual Tracking Workflow"
workflowDisplayName = "Manual Tracking Workflow [Inputs: Raw Data, Pixel Prediction Map]"
workflowDescription = "Manual tracking of objects, based on Prediction Maps or (binary) Segmentation Images"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(ManualTrackingWorkflow, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'\
'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
force5d=True,
instructionText=data_instructions,
max_lanes=1
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,
"Threshold and Size Filter",
"ThresholdTwoLevels" )
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",
workflow=self, interactive=False)
self.trackingApplet = ManualTrackingApplet( workflow=self )
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export")
self.dataExportApplet.set_exporting_operator(opTracking)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue( ['Manual Tracking', 'Object Identities'] )
opDataExport.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self._applets = []
self._applets.append(self.dataSelectionApplet)
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opTwoLevelThreshold.InputImage.connect( opData.ImageGroup[1] )
opTwoLevelThreshold.RawInput.connect( opData.ImageGroup[0] ) # Used for display only
# Use OpReorderAxis for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes( parent=self )
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect( opTwoLevelThreshold.CachedOutput )
opObjExtraction.RawImage.connect( op5Raw.Output )
opObjExtraction.BinaryImage.connect( op5Binary.Output )
opTracking.RawImage.connect( op5Raw.Output )
opTracking.BinaryImage.connect( op5Binary.Output )
opTracking.LabelImage.connect( opObjExtraction.LabelImage )
opTracking.ObjectFeatures.connect( opObjExtraction.RegionFeatures )
opTracking.ComputedFeatureNames.connect(opObjExtraction.ComputedFeatureNames)
opDataExport.Inputs.resize(2)
opDataExport.Inputs[0].connect( opTracking.TrackImage )
opDataExport.Inputs[1].connect( opTracking.LabelImage )
opDataExport.RawData.connect( op5Raw.Output )
opDataExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and \
len(thresholdingOutput) > 0
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.ComputedFeatureNames
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = features_ready and \
len(opTracking.Labels) > 0 and \
opTracking.Labels.ready() and \
opTracking.TrackImage.ready()
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.dataExportApplet.busy
busy |= self.trackingApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet, features_ready and not busy)
self._shell.setAppletEnabled(self.dataExportApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configStructured.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configStructured.selectedFeaturesObjectCount)
self.cropSelectionApplet = CropSelectionApplet(self,"Crop Selection","CropSelection")
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
# self.default_training_export_filename = '{dataset_dir}/{nickname}-training_exported_data.csv'
# self.dataExportAnnotationsApplet = TrackingBaseDataExportApplet(self, "Training Export",default_export_filename=self.default_training_export_filename)
# opDataExportAnnotations = self.dataExportAnnotationsApplet.topLevelOperator
# opDataExportAnnotations.SelectionNames.setValue( ['User Training for Tracking', 'Object Identities'] )
# opDataExportAnnotations.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# self.dataExportAnnotationsApplet.set_exporting_operator(opAnnotations)
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export",default_export_filename=self.default_tracking_export_filename)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking Result', 'Merger Result', 'Object Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.cropSelectionApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
# self._applets.append(self.dataExportAnnotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
class ConservationTrackingWorkflowBase( Workflow ):
workflowName = "Automatic Tracking Workflow (Conservation Tracking) BASE"
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Variables to store division and cell classifiers to prevent retraining every-time batch processing runs
self.stored_division_classifier = None
self.stored_cell_classifier = None
## Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=None
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Segmentation Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,
"Threshold and Size Filter",
"ThresholdTwoLevels" )
self.objectExtractionApplet = TrackingFeatureExtractionApplet(workflow=self, interactive=False,
name="Object Feature Computation")
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
self.divisionDetectionApplet = self._createDivisionDetectionApplet(configConservation.selectedFeaturesDiv) # Might be None
if self.divisionDetectionApplet:
feature_dict_division = {}
feature_dict_division[config.features_division_name] = { name: {} for name in config.division_features }
opObjectExtraction.FeatureNamesDivision.setValue(feature_dict_division)
selected_features_div = {}
for plugin_name in list(config.selected_features_division.keys()):
selected_features_div[plugin_name] = { name: {} for name in config.selected_features_division[plugin_name] }
# FIXME: do not hard code this
for name in [ 'SquaredDistances_' + str(i) for i in range(config.n_best_successors) ]:
selected_features_div[config.features_division_name][name] = {}
opDivisionDetection = self.divisionDetectionApplet.topLevelOperator
opDivisionDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivisionDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivisionDetection.AllowDeleteLabels.setValue(False)
opDivisionDetection.AllowAddLabel.setValue(False)
opDivisionDetection.EnableLabelTransfer.setValue(False)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount)
selected_features_objectcount = {}
for plugin_name in list(config.selected_features_objectcount.keys()):
selected_features_objectcount[plugin_name] = { name: {} for name in config.selected_features_objectcount[plugin_name] }
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount)
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
self.trackingApplet = ConservationTrackingApplet( workflow=self )
self.default_export_filename = '{dataset_dir}/{nickname}-exported_data.csv'
self.dataExportApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_export_filename,
pluginExportFunc=self._pluginExportFunc,
)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue( ['Object-Identities', 'Tracking-Result', 'Merger-Result'] )
opDataExport.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.prepare_lane_for_export = self.prepare_lane_for_export
# configure export settings
# settings = {'file path': self.default_export_filename, 'compression': {}, 'file type': 'csv'}
# selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
# opTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
if self.divisionDetectionApplet:
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(self, "Batch Processing", self.dataSelectionApplet, self.dataExportApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
self._applets.append(self.batchProcessingApplet)
# Parse export and batch command-line arguments for headless mode
if workflow_cmdline_args:
self._data_export_args, unused_args = self.dataExportApplet.parse_known_cmdline_args( workflow_cmdline_args )
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args( workflow_cmdline_args )
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
if unused_args:
logger.warning("Unused command-line args: {}".format( unused_args ))
@property
def applets(self):
return self._applets
def _createDivisionDetectionApplet(self,selectedFeatures=dict()):
return ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=selectedFeatures)
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def prepareForNewLane(self, laneIndex):
# Store division and cell classifiers
if self.divisionDetectionApplet:
opDivisionClassification = self.divisionDetectionApplet.topLevelOperator
if opDivisionClassification.classifier_cache.Output.ready() and \
not opDivisionClassification.classifier_cache._dirty:
self.stored_division_classifier = opDivisionClassification.classifier_cache.Output.value
else:
self.stored_division_classifier = None
opCellClassification = self.cellClassificationApplet.topLevelOperator
if opCellClassification.classifier_cache.Output.ready() and \
not opCellClassification.classifier_cache._dirty:
self.stored_cell_classifier = opCellClassification.classifier_cache.Output.value
else:
self.stored_cell_classifier = None
def handleNewLanesAdded(self):
"""
If new lanes were added, then we invalidated our classifiers unecessarily.
Here, we can restore the classifier so it doesn't need to be retrained.
"""
# If we have stored division and cell classifiers, restore them into the workflow now.
if self.stored_division_classifier:
opDivisionClassification = self.divisionDetectionApplet.topLevelOperator
opDivisionClassification.classifier_cache.forceValue(self.stored_division_classifier)
# Release reference
self.stored_division_classifier = None
# If we have stored division and cell classifiers, restore them into the workflow now.
if self.stored_cell_classifier:
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.classifier_cache.forceValue(self.stored_cell_classifier)
# Release reference
self.stored_cell_classifier = None
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction.setDefaultFeatures(configConservation.allFeaturesObjectCount)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(laneIndex)
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(opData.ImageGroup[0]) # Used for display only
# opTwoLevelThreshold.Channel.setValue(1)
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
# # Connect operators ##
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect( op5Binary.Output )
opDivDetection.RawImages.connect( op5Raw.Output )
opDivDetection.SegmentationImages.connect(opObjExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(opObjExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(opObjExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect( op5Binary.Output )
opCellClassification.RawImages.connect( op5Raw.Output )
opCellClassification.SegmentationImages.connect(opObjExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(opObjExtraction.RegionFeaturesVigra)
opCellClassification.ComputedFeatureNames.connect(opObjExtraction.FeatureNamesVigra)
if self.divisionDetectionApplet:
opTracking.ObjectFeaturesWithDivFeatures.connect( opObjExtraction.RegionFeaturesAll)
opTracking.ComputedFeatureNamesWithDivFeatures.connect( opObjExtraction.ComputedFeatureNamesAll )
opTracking.DivisionProbabilities.connect( opDivDetection.Probabilities )
opTracking.RawImage.connect( op5Raw.Output )
opTracking.LabelImage.connect( opObjExtraction.LabelImage )
opTracking.ObjectFeatures.connect( opObjExtraction.RegionFeaturesVigra )
opTracking.ComputedFeatureNames.connect( opObjExtraction.FeatureNamesVigra)
opTracking.DetectionProbabilities.connect( opCellClassification.Probabilities )
opTracking.NumLabels.connect( opCellClassification.NumLabels )
opDataExport.Inputs.resize(3)
opDataExport.Inputs[0].connect( opTracking.RelabeledImage )
opDataExport.Inputs[1].connect( opTracking.Output )
opDataExport.Inputs[2].connect( opTracking.MergerOutput )
opDataExport.RawData.connect( op5Raw.Output )
opDataExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def prepare_lane_for_export(self, lane_index):
# Bypass cache on headless mode and batch processing mode
self.objectExtractionApplet.topLevelOperator[lane_index].BypassModeEnabled.setValue(True)
if not self.fromBinary:
self.thresholdTwoLevelsApplet.topLevelOperator[lane_index].opCache.BypassModeEnabled.setValue(True)
self.thresholdTwoLevelsApplet.topLevelOperator[lane_index].opSmootherCache.BypassModeEnabled.setValue(True)
# Get axes info
maxt = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[0]
maxx = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[1]
maxy = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[2]
maxz = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[3]
time_enum = list(range(maxt))
x_range = (0, maxx)
y_range = (0, maxy)
z_range = (0, maxz)
ndim = 2
if ( z_range[1] - z_range[0] ) > 1:
ndim = 3
parameters = self.trackingApplet.topLevelOperator.Parameters.value
# Save state of axis ranges
if 'time_range' in parameters:
self.prev_time_range = parameters['time_range']
else:
self.prev_time_range = time_enum
if 'x_range' in parameters:
self.prev_x_range = parameters['x_range']
else:
self.prev_x_range = x_range
if 'y_range' in parameters:
self.prev_y_range = parameters['y_range']
else:
self.prev_y_range = y_range
if 'z_range' in parameters:
self.prev_z_range = parameters['z_range']
else:
self.prev_z_range = z_range
if 'numFramesPerSplit' in parameters:
numFramesPerSplit = parameters['numFramesPerSplit']
else:
numFramesPerSplit = 0
self.trackingApplet.topLevelOperator[lane_index].track(
time_range = time_enum,
x_range = x_range,
y_range = y_range,
z_range = z_range,
size_range = parameters['size_range'],
x_scale = parameters['scales'][0],
y_scale = parameters['scales'][1],
z_scale = parameters['scales'][2],
maxDist=parameters['maxDist'],
maxObj = parameters['maxObj'],
divThreshold=parameters['divThreshold'],
avgSize=parameters['avgSize'],
withTracklets=parameters['withTracklets'],
sizeDependent=parameters['sizeDependent'],
divWeight=parameters['divWeight'],
transWeight=parameters['transWeight'],
withDivisions=parameters['withDivisions'],
withOpticalCorrection=parameters['withOpticalCorrection'],
withClassifierPrior=parameters['withClassifierPrior'],
ndim=ndim,
withMergerResolution=parameters['withMergerResolution'],
borderAwareWidth = parameters['borderAwareWidth'],
withArmaCoordinates = parameters['withArmaCoordinates'],
cplex_timeout = parameters['cplex_timeout'],
appearance_cost = parameters['appearanceCost'],
disappearance_cost = parameters['disappearanceCost'],
max_nearest_neighbors = parameters['max_nearest_neighbors'],
numFramesPerSplit = numFramesPerSplit,
force_build_hypotheses_graph = False,
withBatchProcessing = True
)
def _pluginExportFunc(self, lane_index, filename, exportPlugin, checkOverwriteFiles, plugArgsSlot) -> int:
return (
self.trackingApplet
.topLevelOperator
.getLane(lane_index)
.exportPlugin(
filename,
exportPlugin,
checkOverwriteFiles,
plugArgsSlot
)
)
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def onProjectLoaded(self, projectManager):
"""
Overridden from Workflow base class. Called by the Project Manager.
If the user provided command-line arguments, use them to configure
the workflow inputs and output settings.
"""
# Configure the data export operator.
if self._data_export_args:
self.dataExportApplet.configure_operator_with_parsed_args( self._data_export_args )
# Configure headless mode.
if self._headless and self._batch_input_args and self._data_export_args:
logger.info("Beginning Batch Processing")
self.batchProcessingApplet.run_export_from_parsed_args(self._batch_input_args)
logger.info("Completed Batch Processing")
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
opDataSelection = self.dataSelectionApplet.topLevelOperator
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and \
len(thresholdingOutput) > 0
else:
thresholding_ready = True and input_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.ComputedFeatureNamesAll
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
objectCountClassifier_ready = features_ready
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = objectCountClassifier_ready
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportApplet.busy
busy |= self.batchProcessingApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
if self.divisionDetectionApplet:
self._shell.setAppletEnabled(self.divisionDetectionApplet, features_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet, features_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet, objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
self._shell.setAppletEnabled(self.batchProcessingApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
class ManualTrackingWorkflow(Workflow):
workflowName = "Manual Tracking Workflow"
workflowDisplayName = "Manual Tracking Workflow [Inputs: Raw Data, Pixel Prediction Map]"
workflowDescription = "Manual tracking of objects, based on Prediction Maps or (binary) Segmentation Images"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__(self, shell, headless, workflow_cmdline_args,
project_creation_args, *args, **kwargs):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(ManualTrackingWorkflow, self).__init__(shell,
headless,
workflow_cmdline_args,
project_creation_args,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'\
'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
forceAxisOrder='txyzc',
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
opDataSelection.DatasetRoles.setValue(['Raw Data', 'Prediction Maps'])
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
self.objectExtractionApplet = ObjectExtractionApplet(
name="Object Feature Computation",
workflow=self,
interactive=False)
self.trackingApplet = ManualTrackingApplet(workflow=self)
self.default_export_filename = '{dataset_dir}/{nickname}-exported_data.csv'
self.dataExportApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_export_filename)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
['Manual Tracking', 'Object Identities'])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(
laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(
laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(
laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(
opData.ImageGroup[0]) # Used for display only
# Use OpReorderAxis for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(opTwoLevelThreshold.CachedOutput)
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
opTracking.RawImage.connect(op5Raw.Output)
opTracking.BinaryImage.connect(op5Binary.Output)
opTracking.LabelImage.connect(opObjExtraction.LabelImage)
opTracking.ObjectFeatures.connect(opObjExtraction.RegionFeatures)
opTracking.ComputedFeatureNames.connect(opObjExtraction.Features)
opDataExport.Inputs.resize(2)
opDataExport.Inputs[0].connect(opTracking.TrackImage)
opDataExport.Inputs[1].connect(opTracking.LabelImage)
opDataExport.RawData.connect(op5Raw.Output)
opDataExport.RawDatasetInfo.connect(opData.DatasetGroup[0])
def post_process_lane_export(self, lane_index):
# FIXME: This probably only works for the non-blockwise export slot.
# We should assert that the user isn't using the blockwise slot.
settings, selected_features = self.trackingApplet.topLevelOperator.getLane(
lane_index).get_table_export_settings()
if settings:
raw_dataset_info = self.dataSelectionApplet.topLevelOperator.DatasetGroup[
lane_index][0].value
if raw_dataset_info.location == DatasetInfo.Location.FileSystem:
filename_suffix = raw_dataset_info.nickname
else:
filename_suffix = str(lane_index)
req = self.trackingApplet.topLevelOperator.getLane(
lane_index
).export_object_data(
lane_index,
# FIXME: Even in non-headless mode, we can't show the gui because we're running in a non-main thread.
# That's not a huge deal, because there's still a progress bar for the overall export.
show_gui=False,
filename_suffix=filename_suffix)
req.wait()
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and \
len(thresholdingOutput) > 0
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
features_ready = thresholding_ready
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = features_ready and \
len(opTracking.Labels) > 0 and \
opTracking.Labels.ready() and \
opTracking.TrackImage.ready()
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.dataExportApplet.busy
busy |= self.trackingApplet.busy
self._shell.enableProjectChanges(not busy)
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready
and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet,
thresholding_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet, features_ready
and not busy)
self._shell.setAppletEnabled(self.dataExportApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
class StructuredTrackingWorkflowBase( Workflow ):
workflowName = "Structured Learning Tracking Workflow BASE"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configConservation.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount)
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
if SOLVER=="CPLEX" or SOLVER=="GUROBI":
self._solver="ILP"
elif SOLVER=="DPCT":
self._solver="Flow-based"
else:
self._solver=None
opStructuredTracking._solver = self._solver
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_tracking_export_filename,
pluginExportFunc=self._pluginExportFunc
)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking-Result', 'Merger-Result', 'Object-Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.prepare_lane_for_export = self.prepare_lane_for_export
# configure export settings
settings = {'file path': self.default_tracking_export_filename, 'compression': {}, 'file type': 'h5'}
selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
opStructuredTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(self, "Batch Processing", self.dataSelectionApplet, self.dataExportTrackingApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
if workflow_cmdline_args:
if '--testFullAnnotations' in workflow_cmdline_args:
self.testFullAnnotations = True
else:
self.testFullAnnotations = False
self._data_export_args, unused_args = self.dataExportTrackingApplet.parse_known_cmdline_args( workflow_cmdline_args )
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args( workflow_cmdline_args )
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
self.testFullAnnotations = False
if unused_args:
logger.warning("Unused command-line args: {}".format( unused_args ))
def _pluginExportFunc(self, lane_index, filename, exportPlugin, checkOverwriteFiles, plugArgsSlot) -> int:
return (
self.trackingApplet
.topLevelOperator
.getLane(lane_index)
.exportPlugin(
filename,
exportPlugin,
checkOverwriteFiles,
additionalPluginArgumentsSlot
)
)
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator.getLane(laneIndex)
opAnnotations = self.annotationsApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
opStructuredTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataTrackingExport = self.dataExportTrackingApplet.topLevelOperator.getLane(laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect( opData.ImageGroup[1] )
opTwoLevelThreshold.RawInput.connect( opData.ImageGroup[0] ) # Used for display only
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
opObjExtraction.RawImage.connect( op5Raw.Output )
opObjExtraction.BinaryImage.connect( op5Binary.Output )
opTrackingFeatureExtraction.RawImage.connect( op5Raw.Output )
opTrackingFeatureExtraction.BinaryImage.connect( op5Binary.Output )
opTrackingFeatureExtraction.setDefaultFeatures(configConservation.allFeaturesObjectCount)
opTrackingFeatureExtraction.FeatureNamesVigra.setValue(configConservation.allFeaturesObjectCount)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = { name: {} for name in config.division_features }
opTrackingFeatureExtraction.FeatureNamesDivision.setValue(feature_dict_division)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect( op5Binary.Output )
opDivDetection.RawImages.connect( op5Raw.Output )
opDivDetection.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect( op5Binary.Output )
opCellClassification.RawImages.connect( op5Raw.Output )
opCellClassification.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opCellClassification.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesNoDivisions)
opAnnotations.RawImage.connect( op5Raw.Output )
opAnnotations.BinaryImage.connect( op5Binary.Output )
opAnnotations.LabelImage.connect( opObjExtraction.LabelImage )
opAnnotations.ObjectFeatures.connect( opObjExtraction.RegionFeatures )
opAnnotations.ComputedFeatureNames.connect(opObjExtraction.Features)
opAnnotations.DivisionProbabilities.connect( opDivDetection.Probabilities )
opAnnotations.DetectionProbabilities.connect( opCellClassification.Probabilities )
opAnnotations.MaxNumObj.connect (opCellClassification.MaxNumObj)
opStructuredTracking.RawImage.connect( op5Raw.Output )
opStructuredTracking.LabelImage.connect( opTrackingFeatureExtraction.LabelImage )
opStructuredTracking.ObjectFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesVigra )
opStructuredTracking.ComputedFeatureNames.connect( opTrackingFeatureExtraction.FeatureNamesVigra )
if self.divisionDetectionApplet:
opStructuredTracking.ObjectFeaturesWithDivFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesAll)
opStructuredTracking.ComputedFeatureNamesWithDivFeatures.connect( opTrackingFeatureExtraction.ComputedFeatureNamesAll )
opStructuredTracking.DivisionProbabilities.connect( opDivDetection.Probabilities )
opStructuredTracking.DetectionProbabilities.connect( opCellClassification.Probabilities )
opStructuredTracking.NumLabels.connect( opCellClassification.NumLabels )
opStructuredTracking.Annotations.connect (opAnnotations.Annotations)
opStructuredTracking.Labels.connect (opAnnotations.Labels)
opStructuredTracking.Divisions.connect (opAnnotations.Divisions)
opStructuredTracking.Appearances.connect (opAnnotations.Appearances)
opStructuredTracking.Disappearances.connect (opAnnotations.Disappearances)
opStructuredTracking.MaxNumObj.connect (opCellClassification.MaxNumObj)
opDataTrackingExport.Inputs.resize(3)
opDataTrackingExport.Inputs[0].connect( opStructuredTracking.RelabeledImage )
opDataTrackingExport.Inputs[1].connect( opStructuredTracking.MergerOutput )
opDataTrackingExport.Inputs[2].connect( opStructuredTracking.LabelImage )
opDataTrackingExport.RawData.connect( op5Raw.Output )
opDataTrackingExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def prepare_lane_for_export(self, lane_index):
import logging
logger = logging.getLogger(__name__)
maxt = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[0]
maxx = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[1]
maxy = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[2]
maxz = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[3]
time_enum = list(range(maxt))
x_range = (0, maxx)
y_range = (0, maxy)
z_range = (0, maxz)
ndim = 2
if ( z_range[1] - z_range[0] ) > 1:
ndim = 3
parameters = self.trackingApplet.topLevelOperator.Parameters.value
# Save state of axis ranges
if 'time_range' in parameters:
self.prev_time_range = parameters['time_range']
else:
self.prev_time_range = time_enum
if 'x_range' in parameters:
self.prev_x_range = parameters['x_range']
else:
self.prev_x_range = x_range
if 'y_range' in parameters:
self.prev_y_range = parameters['y_range']
else:
self.prev_y_range = y_range
if 'z_range' in parameters:
self.prev_z_range = parameters['z_range']
else:
self.prev_z_range = z_range
# batch processing starts a new lane, so training data needs to be copied from the lane that loaded the project
loaded_project_lane_index=0
self.annotationsApplet.topLevelOperator[lane_index].Annotations.setValue(
self.trackingApplet.topLevelOperator[loaded_project_lane_index].Annotations.value)
def runLearningAndTracking(withMergerResolution=True):
if self.testFullAnnotations:
logger.info("Test: Structured Learning")
weights = self.trackingApplet.topLevelOperator[lane_index]._runStructuredLearning(
z_range,
parameters['maxObj'],
parameters['max_nearest_neighbors'],
parameters['maxDist'],
parameters['divThreshold'],
[parameters['scales'][0],parameters['scales'][1],parameters['scales'][2]],
parameters['size_range'],
parameters['withDivisions'],
parameters['borderAwareWidth'],
parameters['withClassifierPrior'],
withBatchProcessing=True)
logger.info("weights: {}".format(weights))
logger.info("Test: Tracking")
result = self.trackingApplet.topLevelOperator[lane_index].track(
time_range = time_enum,
x_range = x_range,
y_range = y_range,
z_range = z_range,
size_range = parameters['size_range'],
x_scale = parameters['scales'][0],
y_scale = parameters['scales'][1],
z_scale = parameters['scales'][2],
maxDist=parameters['maxDist'],
maxObj = parameters['maxObj'],
divThreshold=parameters['divThreshold'],
avgSize=parameters['avgSize'],
withTracklets=parameters['withTracklets'],
sizeDependent=parameters['sizeDependent'],
detWeight=parameters['detWeight'],
divWeight=parameters['divWeight'],
transWeight=parameters['transWeight'],
withDivisions=parameters['withDivisions'],
withOpticalCorrection=parameters['withOpticalCorrection'],
withClassifierPrior=parameters['withClassifierPrior'],
ndim=ndim,
withMergerResolution=withMergerResolution,
borderAwareWidth = parameters['borderAwareWidth'],
withArmaCoordinates = parameters['withArmaCoordinates'],
cplex_timeout = parameters['cplex_timeout'],
appearance_cost = parameters['appearanceCost'],
disappearance_cost = parameters['disappearanceCost'],
force_build_hypotheses_graph = False,
withBatchProcessing = True
)
return result
if self.testFullAnnotations:
self.result = runLearningAndTracking(withMergerResolution=False)
hypothesesGraph = self.trackingApplet.topLevelOperator[lane_index].LearningHypothesesGraph.value
hypothesesGraph.insertSolution(self.result)
hypothesesGraph.computeLineage()
solution = hypothesesGraph.getSolutionDictionary()
annotations = self.trackingApplet.topLevelOperator[lane_index].Annotations.value
self.trackingApplet.topLevelOperator[lane_index].insertAnnotationsToHypothesesGraph(hypothesesGraph,annotations,misdetectionLabel=-1)
hypothesesGraph.computeLineage()
solutionFromAnnotations = hypothesesGraph.getSolutionDictionary()
for key in list(solution.keys()):
if key == 'detectionResults':
detectionFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if solution[key][i]['id'] == solutionFromAnnotations[key][j]['id'] and \
solution[key][i]['value'] == solutionFromAnnotations[key][j]['value']:
flag = True
break
detectionFlag &= flag
elif key == 'divisionResults':
divisionFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if solution[key][i]['id'] == solutionFromAnnotations[key][j]['id'] and \
solution[key][i]['value'] == solutionFromAnnotations[key][j]['value']:
flag = True
break
divisionFlag &= flag
elif key == 'linkingResults':
linkingFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if solution[key][i]['dest'] == solutionFromAnnotations[key][j]['dest'] and \
solution[key][i]['src'] == solutionFromAnnotations[key][j]['src']:
if solution[key][i]['gap'] == solutionFromAnnotations[key][j]['gap'] and \
solution[key][i]['value'] == solutionFromAnnotations[key][j]['value']:
flag = True
break
linkingFlag &= flag
assert detectionFlag, "Detection results are NOT correct. They differ from your annotated detections."
logger.info("Detection results are correct.")
assert divisionFlag, "Division results are NOT correct. They differ from your annotated divisions."
logger.info("Division results are correct.")
assert linkingFlag, "Transition results are NOT correct. They differ from your annotated transitions."
logger.info("Transition results are correct.")
self.result = runLearningAndTracking(withMergerResolution=parameters['withMergerResolution'])
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def onProjectLoaded(self, projectManager):
"""
Overridden from Workflow base class. Called by the Project Manager.
If the user provided command-line arguments, use them to configure
the workflow inputs and output settings.
"""
# Configure the data export operator.
if self._data_export_args:
self.dataExportTrackingApplet.configure_operator_with_parsed_args( self._data_export_args )
# Configure headless mode.
if self._headless and self._batch_input_args and self._data_export_args:
logger.info("Beginning Batch Processing")
self.batchProcessingApplet.run_export_from_parsed_args(self._batch_input_args)
logger.info("Completed Batch Processing")
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and len(thresholdingOutput) > 0
else:
thresholding_ready = input_ready
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator
trackingFeatureExtractionOutput = opTrackingFeatureExtraction.ComputedFeatureNamesAll
tracking_features_ready = thresholding_ready and len(trackingFeatureExtractionOutput) > 0
objectCountClassifier_ready = tracking_features_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.RegionFeatures
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
opAnnotations = self.annotationsApplet.topLevelOperator
annotations_ready = features_ready and \
len(opAnnotations.Labels) > 0 and \
opAnnotations.Labels.ready() and \
opAnnotations.TrackImage.ready()
opStructuredTracking = self.trackingApplet.topLevelOperator
structured_tracking_ready = objectCountClassifier_ready
withIlpSolver = (self._solver=="ILP")
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.annotationsApplet.busy
# busy |= self.dataExportAnnotationsApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportTrackingApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.trackingFeatureExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, not busy)
self._shell.setAppletEnabled(self.annotationsApplet, features_ready and not busy) # and withIlpSolver)
# self._shell.setAppletEnabled(self.dataExportAnnotationsApplet, annotations_ready and not busy and \
# self.dataExportAnnotationsApplet.topLevelOperator.Inputs[0][0].ready() )
self._shell.setAppletEnabled(self.trackingApplet, objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportTrackingApplet, structured_tracking_ready and not busy and \
self.dataExportTrackingApplet.topLevelOperator.Inputs[0][0].ready() )
def __init__(self, shell, headless, workflow_cmdline_args,
project_creation_args, *args, **kwargs):
graph = kwargs["graph"] if "graph" in kwargs else Graph()
if "graph" in kwargs:
del kwargs["graph"]
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase,
self).__init__(shell,
headless,
workflow_cmdline_args,
project_creation_args,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Variables to store division and cell classifiers to prevent retraining every-time batch processing runs
self.stored_division_classifier = None
self.stored_cell_classifier = None
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
forceAxisOrder=["txyzc"],
instructionText=data_instructions,
max_lanes=None,
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue(
["Raw Data", "Segmentation Image"])
else:
opDataSelection.DatasetRoles.setValue(
["Raw Data", "Prediction Maps"])
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
self.objectExtractionApplet = TrackingFeatureExtractionApplet(
workflow=self,
interactive=False,
name="Object Feature Computation")
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
self.divisionDetectionApplet = self._createDivisionDetectionApplet(
configConservation.selectedFeaturesDiv) # Might be None
if self.divisionDetectionApplet:
feature_dict_division = {}
feature_dict_division[config.features_division_name] = {
name: {}
for name in config.division_features
}
opObjectExtraction.FeatureNamesDivision.setValue(
feature_dict_division)
selected_features_div = {}
for plugin_name in list(config.selected_features_division.keys()):
selected_features_div[plugin_name] = {
name: {}
for name in config.selected_features_division[plugin_name]
}
# FIXME: do not hard code this
for name in [
"SquaredDistances_" + str(i)
for i in range(config.n_best_successors)
]:
selected_features_div[config.features_division_name][name] = {}
opDivisionDetection = self.divisionDetectionApplet.topLevelOperator
opDivisionDetection.SelectedFeatures.setValue(
configConservation.selectedFeaturesDiv)
opDivisionDetection.LabelNames.setValue(
["Not Dividing", "Dividing"])
opDivisionDetection.AllowDeleteLabels.setValue(False)
opDivisionDetection.AllowAddLabel.setValue(False)
opDivisionDetection.EnableLabelTransfer.setValue(False)
self.cellClassificationApplet = ObjectClassificationApplet(
workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount,
)
selected_features_objectcount = {}
for plugin_name in list(config.selected_features_objectcount.keys()):
selected_features_objectcount[plugin_name] = {
name: {}
for name in config.selected_features_objectcount[plugin_name]
}
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(
configConservation.selectedFeaturesObjectCount)
opCellClassification.SuggestedLabelNames.setValue(
["False Detection"] + [str(1) + " Object"] +
[str(i) + " Objects" for i in range(2, 10)])
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
self.trackingApplet = ConservationTrackingApplet(workflow=self)
self.default_export_filename = "{dataset_dir}/{nickname}-exported_data.csv"
self.dataExportApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_export_filename,
pluginExportFunc=self._pluginExportFunc,
)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
["Object-Identities", "Tracking-Result", "Merger-Result"])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.prepare_lane_for_export = self.prepare_lane_for_export
# configure export settings
# settings = {'file path': self.default_export_filename, 'compression': {}, 'file type': 'csv'}
# selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
# opTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
if self.divisionDetectionApplet:
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(
self, "Batch Processing", self.dataSelectionApplet,
self.dataExportApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
self._applets.append(self.batchProcessingApplet)
# Parse export and batch command-line arguments for headless mode
if workflow_cmdline_args:
self._data_export_args, unused_args = self.dataExportApplet.parse_known_cmdline_args(
workflow_cmdline_args)
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args(
workflow_cmdline_args)
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
if unused_args:
logger.warning("Unused command-line args: {}".format(unused_args))
class StructuredTrackingWorkflowBase( Workflow ):
workflowName = "Structured Learning Tracking Workflow BASE"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configConservation.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount)
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
if SOLVER=="CPLEX" or SOLVER=="GUROBI":
self._solver="ILP"
elif SOLVER=="DPCT":
self._solver="Flow-based"
else:
self._solver=None
opStructuredTracking._solver = self._solver
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export",default_export_filename=self.default_tracking_export_filename)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking-Result', 'Merger-Result', 'Object-Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.prepare_lane_for_export = self.prepare_lane_for_export
self.dataExportTrackingApplet.post_process_lane_export = self.post_process_lane_export
# configure export settings
settings = {'file path': self.default_tracking_export_filename, 'compression': {}, 'file type': 'h5'}
selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
opStructuredTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(self, "Batch Processing", self.dataSelectionApplet, self.dataExportTrackingApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
if workflow_cmdline_args:
if '--testFullAnnotations' in workflow_cmdline_args:
self.testFullAnnotations = True
else:
self.testFullAnnotations = False
self._data_export_args, unused_args = self.dataExportTrackingApplet.parse_known_cmdline_args( workflow_cmdline_args )
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args( workflow_cmdline_args )
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
self.testFullAnnotations = False
if unused_args:
logger.warning("Unused command-line args: {}".format( unused_args ))
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator.getLane(laneIndex)
opAnnotations = self.annotationsApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
opStructuredTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataTrackingExport = self.dataExportTrackingApplet.topLevelOperator.getLane(laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect( opData.ImageGroup[1] )
opTwoLevelThreshold.RawInput.connect( opData.ImageGroup[0] ) # Used for display only
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
opObjExtraction.RawImage.connect( op5Raw.Output )
opObjExtraction.BinaryImage.connect( op5Binary.Output )
opTrackingFeatureExtraction.RawImage.connect( op5Raw.Output )
opTrackingFeatureExtraction.BinaryImage.connect( op5Binary.Output )
opTrackingFeatureExtraction.setDefaultFeatures(configConservation.allFeaturesObjectCount)
opTrackingFeatureExtraction.FeatureNamesVigra.setValue(configConservation.allFeaturesObjectCount)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = { name: {} for name in config.division_features }
opTrackingFeatureExtraction.FeatureNamesDivision.setValue(feature_dict_division)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect( op5Binary.Output )
opDivDetection.RawImages.connect( op5Raw.Output )
opDivDetection.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect( op5Binary.Output )
opCellClassification.RawImages.connect( op5Raw.Output )
opCellClassification.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opCellClassification.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesNoDivisions)
opAnnotations.RawImage.connect( op5Raw.Output )
opAnnotations.BinaryImage.connect( op5Binary.Output )
opAnnotations.LabelImage.connect( opObjExtraction.LabelImage )
opAnnotations.ObjectFeatures.connect( opObjExtraction.RegionFeatures )
opAnnotations.ComputedFeatureNames.connect(opObjExtraction.Features)
opAnnotations.DivisionProbabilities.connect( opDivDetection.Probabilities )
opAnnotations.DetectionProbabilities.connect( opCellClassification.Probabilities )
opAnnotations.MaxNumObj.connect (opCellClassification.MaxNumObj)
opStructuredTracking.RawImage.connect( op5Raw.Output )
opStructuredTracking.LabelImage.connect( opTrackingFeatureExtraction.LabelImage )
opStructuredTracking.ObjectFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesVigra )
opStructuredTracking.ComputedFeatureNames.connect( opTrackingFeatureExtraction.FeatureNamesVigra )
if self.divisionDetectionApplet:
opStructuredTracking.ObjectFeaturesWithDivFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesAll)
opStructuredTracking.ComputedFeatureNamesWithDivFeatures.connect( opTrackingFeatureExtraction.ComputedFeatureNamesAll )
opStructuredTracking.DivisionProbabilities.connect( opDivDetection.Probabilities )
opStructuredTracking.DetectionProbabilities.connect( opCellClassification.Probabilities )
opStructuredTracking.NumLabels.connect( opCellClassification.NumLabels )
opStructuredTracking.Annotations.connect (opAnnotations.Annotations)
opStructuredTracking.Labels.connect (opAnnotations.Labels)
opStructuredTracking.Divisions.connect (opAnnotations.Divisions)
opStructuredTracking.Appearances.connect (opAnnotations.Appearances)
opStructuredTracking.Disappearances.connect (opAnnotations.Disappearances)
opStructuredTracking.MaxNumObj.connect (opCellClassification.MaxNumObj)
opDataTrackingExport.Inputs.resize(3)
opDataTrackingExport.Inputs[0].connect( opStructuredTracking.RelabeledImage )
opDataTrackingExport.Inputs[1].connect( opStructuredTracking.MergerOutput )
opDataTrackingExport.Inputs[2].connect( opStructuredTracking.LabelImage )
opDataTrackingExport.RawData.connect( op5Raw.Output )
opDataTrackingExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def prepare_lane_for_export(self, lane_index):
import logging
logger = logging.getLogger(__name__)
maxt = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[0]
maxx = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[1]
maxy = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[2]
maxz = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[3]
time_enum = list(range(maxt))
x_range = (0, maxx)
y_range = (0, maxy)
z_range = (0, maxz)
ndim = 2
if ( z_range[1] - z_range[0] ) > 1:
ndim = 3
parameters = self.trackingApplet.topLevelOperator.Parameters.value
# Save state of axis ranges
if 'time_range' in parameters:
self.prev_time_range = parameters['time_range']
else:
self.prev_time_range = time_enum
if 'x_range' in parameters:
self.prev_x_range = parameters['x_range']
else:
self.prev_x_range = x_range
if 'y_range' in parameters:
self.prev_y_range = parameters['y_range']
else:
self.prev_y_range = y_range
if 'z_range' in parameters:
self.prev_z_range = parameters['z_range']
else:
self.prev_z_range = z_range
# batch processing starts a new lane, so training data needs to be copied from the lane that loaded the project
loaded_project_lane_index=0
self.annotationsApplet.topLevelOperator[lane_index].Annotations.setValue(
self.trackingApplet.topLevelOperator[loaded_project_lane_index].Annotations.value)
def runLearningAndTracking(withMergerResolution=True):
if self.testFullAnnotations:
logger.info("Test: Structured Learning")
weights = self.trackingApplet.topLevelOperator[lane_index]._runStructuredLearning(
z_range,
parameters['maxObj'],
parameters['max_nearest_neighbors'],
parameters['maxDist'],
parameters['divThreshold'],
[parameters['scales'][0],parameters['scales'][1],parameters['scales'][2]],
parameters['size_range'],
parameters['withDivisions'],
parameters['borderAwareWidth'],
parameters['withClassifierPrior'],
withBatchProcessing=True)
logger.info("weights: {}".format(weights))
logger.info("Test: Tracking")
result = self.trackingApplet.topLevelOperator[lane_index].track(
time_range = time_enum,
x_range = x_range,
y_range = y_range,
z_range = z_range,
size_range = parameters['size_range'],
x_scale = parameters['scales'][0],
y_scale = parameters['scales'][1],
z_scale = parameters['scales'][2],
maxDist=parameters['maxDist'],
maxObj = parameters['maxObj'],
divThreshold=parameters['divThreshold'],
avgSize=parameters['avgSize'],
withTracklets=parameters['withTracklets'],
sizeDependent=parameters['sizeDependent'],
detWeight=parameters['detWeight'],
divWeight=parameters['divWeight'],
transWeight=parameters['transWeight'],
withDivisions=parameters['withDivisions'],
withOpticalCorrection=parameters['withOpticalCorrection'],
withClassifierPrior=parameters['withClassifierPrior'],
ndim=ndim,
withMergerResolution=withMergerResolution,
borderAwareWidth = parameters['borderAwareWidth'],
withArmaCoordinates = parameters['withArmaCoordinates'],
cplex_timeout = parameters['cplex_timeout'],
appearance_cost = parameters['appearanceCost'],
disappearance_cost = parameters['disappearanceCost'],
force_build_hypotheses_graph = False,
withBatchProcessing = True
)
return result
if self.testFullAnnotations:
self.result = runLearningAndTracking(withMergerResolution=False)
hypothesesGraph = self.trackingApplet.topLevelOperator[lane_index].LearningHypothesesGraph.value
hypothesesGraph.insertSolution(self.result)
hypothesesGraph.computeLineage()
solution = hypothesesGraph.getSolutionDictionary()
annotations = self.trackingApplet.topLevelOperator[lane_index].Annotations.value
self.trackingApplet.topLevelOperator[lane_index].insertAnnotationsToHypothesesGraph(hypothesesGraph,annotations,misdetectionLabel=-1)
hypothesesGraph.computeLineage()
solutionFromAnnotations = hypothesesGraph.getSolutionDictionary()
for key in list(solution.keys()):
if key == 'detectionResults':
detectionFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if solution[key][i]['id'] == solutionFromAnnotations[key][j]['id'] and \
solution[key][i]['value'] == solutionFromAnnotations[key][j]['value']:
flag = True
break
detectionFlag &= flag
elif key == 'divisionResults':
divisionFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if solution[key][i]['id'] == solutionFromAnnotations[key][j]['id'] and \
solution[key][i]['value'] == solutionFromAnnotations[key][j]['value']:
flag = True
break
divisionFlag &= flag
elif key == 'linkingResults':
linkingFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if solution[key][i]['dest'] == solutionFromAnnotations[key][j]['dest'] and \
solution[key][i]['src'] == solutionFromAnnotations[key][j]['src']:
if solution[key][i]['gap'] == solutionFromAnnotations[key][j]['gap'] and \
solution[key][i]['value'] == solutionFromAnnotations[key][j]['value']:
flag = True
break
linkingFlag &= flag
assert detectionFlag, "Detection results are NOT correct. They differ from your annotated detections."
logger.info("Detection results are correct.")
assert divisionFlag, "Division results are NOT correct. They differ from your annotated divisions."
logger.info("Division results are correct.")
assert linkingFlag, "Transition results are NOT correct. They differ from your annotated transitions."
logger.info("Transition results are correct.")
self.result = runLearningAndTracking(withMergerResolution=parameters['withMergerResolution'])
def post_process_lane_export(self, lane_index, checkOverwriteFiles=False):
# Plugin export if selected
logger.info("Export source is: " + self.dataExportTrackingApplet.topLevelOperator.SelectedExportSource.value)
print("in post_process_lane_export")
if self.dataExportTrackingApplet.topLevelOperator.SelectedExportSource.value == OpTrackingBaseDataExport.PluginOnlyName:
logger.info("Export source plugin selected!")
selectedPlugin = self.dataExportTrackingApplet.topLevelOperator.SelectedPlugin.value
additionalPluginArgumentsSlot = self.dataExportTrackingApplet.topLevelOperator.AdditionalPluginArguments
exportPluginInfo = pluginManager.getPluginByName(selectedPlugin, category="TrackingExportFormats")
if exportPluginInfo is None:
logger.error("Could not find selected plugin %s" % exportPluginInfo)
else:
exportPlugin = exportPluginInfo.plugin_object
logger.info("Exporting tracking result using %s" % selectedPlugin)
name_format = self.dataExportTrackingApplet.topLevelOperator.getLane(lane_index).OutputFilenameFormat.value
partially_formatted_name = self.getPartiallyFormattedName(lane_index, name_format)
if exportPlugin.exportsToFile:
filename = partially_formatted_name
if os.path.basename(filename) == '':
filename = os.path.join(filename, 'pluginExport.txt')
else:
filename = partially_formatted_name
if filename is None or len(str(filename)) == 0:
logger.error("Cannot export from plugin with empty output filename")
return True
self.dataExportTrackingApplet.progressSignal(-1)
exportStatus = self.trackingApplet.topLevelOperator.getLane(lane_index).exportPlugin(
filename, exportPlugin, checkOverwriteFiles, additionalPluginArgumentsSlot)
self.dataExportTrackingApplet.progressSignal(100)
if not exportStatus:
return False
logger.info("Export done")
return True
return True
def getPartiallyFormattedName(self, lane_index, path_format_string):
''' Takes the format string for the output file, fills in the most important placeholders, and returns it '''
raw_dataset_info = self.dataSelectionApplet.topLevelOperator.DatasetGroup[lane_index][0].value
project_path = self.shell.projectManager.currentProjectPath
project_dir = os.path.dirname(project_path)
dataset_dir = PathComponents(raw_dataset_info.filePath).externalDirectory
abs_dataset_dir = make_absolute(dataset_dir, cwd=project_dir)
known_keys = {}
known_keys['dataset_dir'] = abs_dataset_dir
nickname = raw_dataset_info.nickname.replace('*', '')
if os.path.pathsep in nickname:
nickname = PathComponents(nickname.split(os.path.pathsep)[0]).fileNameBase
known_keys['nickname'] = nickname
known_keys['result_type'] = self.dataExportTrackingApplet.topLevelOperator.SelectedPlugin._value
# use partial formatting to fill in non-coordinate name fields
partially_formatted_name = format_known_keys(path_format_string, known_keys)
return partially_formatted_name
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def onProjectLoaded(self, projectManager):
"""
Overridden from Workflow base class. Called by the Project Manager.
If the user provided command-line arguments, use them to configure
the workflow inputs and output settings.
"""
# Configure the data export operator.
if self._data_export_args:
self.dataExportTrackingApplet.configure_operator_with_parsed_args( self._data_export_args )
# Configure headless mode.
if self._headless and self._batch_input_args and self._data_export_args:
logger.info("Beginning Batch Processing")
self.batchProcessingApplet.run_export_from_parsed_args(self._batch_input_args)
logger.info("Completed Batch Processing")
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and len(thresholdingOutput) > 0
else:
thresholding_ready = input_ready
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator
trackingFeatureExtractionOutput = opTrackingFeatureExtraction.ComputedFeatureNamesAll
tracking_features_ready = thresholding_ready and len(trackingFeatureExtractionOutput) > 0
objectCountClassifier_ready = tracking_features_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.RegionFeatures
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
opAnnotations = self.annotationsApplet.topLevelOperator
annotations_ready = features_ready and \
len(opAnnotations.Labels) > 0 and \
opAnnotations.Labels.ready() and \
opAnnotations.TrackImage.ready()
opStructuredTracking = self.trackingApplet.topLevelOperator
structured_tracking_ready = objectCountClassifier_ready
withIlpSolver = (self._solver=="ILP")
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.annotationsApplet.busy
# busy |= self.dataExportAnnotationsApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportTrackingApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.trackingFeatureExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, not busy)
self._shell.setAppletEnabled(self.annotationsApplet, features_ready and not busy) # and withIlpSolver)
# self._shell.setAppletEnabled(self.dataExportAnnotationsApplet, annotations_ready and not busy and \
# self.dataExportAnnotationsApplet.topLevelOperator.Inputs[0][0].ready() )
self._shell.setAppletEnabled(self.trackingApplet, objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportTrackingApplet, structured_tracking_ready and not busy and \
self.dataExportTrackingApplet.topLevelOperator.Inputs[0][0].ready() )
class StructuredTrackingWorkflowBase(Workflow):
workflowName = "Structured Learning Tracking Workflow BASE"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__(self, shell, headless, workflow_cmdline_args,
project_creation_args, *args, **kwargs):
graph = kwargs["graph"] if "graph" in kwargs else Graph()
if "graph" in kwargs:
del kwargs["graph"]
super(StructuredTrackingWorkflowBase,
self).__init__(shell,
headless,
workflow_cmdline_args,
project_creation_args,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder=["txyzc"],
instructionText=data_instructions,
max_lanes=1,
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue(["Raw Data", "Binary Image"])
else:
opDataSelection.DatasetRoles.setValue(
["Raw Data", "Prediction Maps"])
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
self.divisionDetectionApplet = ObjectClassificationApplet(
workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configConservation.selectedFeaturesDiv,
)
self.cellClassificationApplet = ObjectClassificationApplet(
workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount,
)
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(
name="Object Feature Computation",
workflow=self,
interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(
name="Object Feature Computation",
workflow=self,
interactive=False)
self.annotationsApplet = AnnotationsApplet(name="Training",
workflow=self)
opAnnotations = self.annotationsApplet.topLevelOperator
self.trackingApplet = StructuredTrackingApplet(
name="Tracking - Structured Learning", workflow=self)
opStructuredTracking = self.trackingApplet.topLevelOperator
if SOLVER == "CPLEX" or SOLVER == "GUROBI":
self._solver = "ILP"
elif SOLVER == "DPCT":
self._solver = "Flow-based"
else:
self._solver = None
opStructuredTracking._solver = self._solver
self.default_tracking_export_filename = "{dataset_dir}/{nickname}-tracking_exported_data.csv"
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_tracking_export_filename,
pluginExportFunc=self._pluginExportFunc,
)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue(
["Tracking-Result", "Merger-Result", "Object-Identities"])
opDataExportTracking.WorkingDirectory.connect(
opDataSelection.WorkingDirectory)
self.dataExportTrackingApplet.set_exporting_operator(
opStructuredTracking)
self.dataExportTrackingApplet.prepare_lane_for_export = self.prepare_lane_for_export
# configure export settings
settings = {
"file path": self.default_tracking_export_filename,
"compression": {},
"file type": "h5"
}
selected_features = [
"Count", "RegionCenter", "RegionRadii", "RegionAxes"
]
opStructuredTracking.ExportSettings.setValue(
(settings, selected_features))
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(
self, "Batch Processing", self.dataSelectionApplet,
self.dataExportTrackingApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(
configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(["Not Dividing", "Dividing"])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(
configConservation.selectedFeaturesObjectCount)
opCellClassification.SuggestedLabelNames.setValue(
["False Detection"] + [str(1) + " Object"] +
[str(i) + " Objects" for i in range(2, 10)])
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
if workflow_cmdline_args:
if "--testFullAnnotations" in workflow_cmdline_args:
self.testFullAnnotations = True
else:
self.testFullAnnotations = False
self._data_export_args, unused_args = self.dataExportTrackingApplet.parse_known_cmdline_args(
workflow_cmdline_args)
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args(
workflow_cmdline_args)
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
self.testFullAnnotations = False
if unused_args:
logger.warning("Unused command-line args: {}".format(unused_args))
def _pluginExportFunc(self, lane_index, filename, exportPlugin,
checkOverwriteFiles, plugArgsSlot) -> int:
return self.trackingApplet.topLevelOperator.getLane(
lane_index).exportPlugin(filename, exportPlugin,
checkOverwriteFiles, plugArgsSlot)
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(
laneIndex)
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator.getLane(
laneIndex)
opAnnotations = self.annotationsApplet.topLevelOperator.getLane(
laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(
laneIndex)
opStructuredTracking = self.trackingApplet.topLevelOperator.getLane(
laneIndex)
opDataTrackingExport = self.dataExportTrackingApplet.topLevelOperator.getLane(
laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(
laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(
laneIndex)
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(
opData.ImageGroup[0]) # Used for display only
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
opTrackingFeatureExtraction.RawImage.connect(op5Raw.Output)
opTrackingFeatureExtraction.BinaryImage.connect(op5Binary.Output)
opTrackingFeatureExtraction.setDefaultFeatures(
configConservation.allFeaturesObjectCount)
opTrackingFeatureExtraction.FeatureNamesVigra.setValue(
configConservation.allFeaturesObjectCount)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = {
name: {}
for name in config.division_features
}
opTrackingFeatureExtraction.FeatureNamesDivision.setValue(
feature_dict_division)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect(op5Binary.Output)
opDivDetection.RawImages.connect(op5Raw.Output)
opDivDetection.SegmentationImages.connect(
opTrackingFeatureExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(
opTrackingFeatureExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(
opTrackingFeatureExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect(op5Binary.Output)
opCellClassification.RawImages.connect(op5Raw.Output)
opCellClassification.SegmentationImages.connect(
opTrackingFeatureExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(
opTrackingFeatureExtraction.RegionFeaturesAll)
opCellClassification.ComputedFeatureNames.connect(
opTrackingFeatureExtraction.ComputedFeatureNamesNoDivisions)
opAnnotations.RawImage.connect(op5Raw.Output)
opAnnotations.BinaryImage.connect(op5Binary.Output)
opAnnotations.LabelImage.connect(opObjExtraction.LabelImage)
opAnnotations.ObjectFeatures.connect(opObjExtraction.RegionFeatures)
opAnnotations.ComputedFeatureNames.connect(opObjExtraction.Features)
opAnnotations.DivisionProbabilities.connect(
opDivDetection.Probabilities)
opAnnotations.DetectionProbabilities.connect(
opCellClassification.Probabilities)
opAnnotations.MaxNumObj.connect(opCellClassification.MaxNumObj)
opStructuredTracking.RawImage.connect(op5Raw.Output)
opStructuredTracking.LabelImage.connect(
opTrackingFeatureExtraction.LabelImage)
opStructuredTracking.ObjectFeatures.connect(
opTrackingFeatureExtraction.RegionFeaturesVigra)
opStructuredTracking.ComputedFeatureNames.connect(
opTrackingFeatureExtraction.FeatureNamesVigra)
if self.divisionDetectionApplet:
opStructuredTracking.ObjectFeaturesWithDivFeatures.connect(
opTrackingFeatureExtraction.RegionFeaturesAll)
opStructuredTracking.ComputedFeatureNamesWithDivFeatures.connect(
opTrackingFeatureExtraction.ComputedFeatureNamesAll)
opStructuredTracking.DivisionProbabilities.connect(
opDivDetection.Probabilities)
opStructuredTracking.DetectionProbabilities.connect(
opCellClassification.Probabilities)
opStructuredTracking.NumLabels.connect(opCellClassification.NumLabels)
opStructuredTracking.Annotations.connect(opAnnotations.Annotations)
opStructuredTracking.Labels.connect(opAnnotations.Labels)
opStructuredTracking.Divisions.connect(opAnnotations.Divisions)
opStructuredTracking.Appearances.connect(opAnnotations.Appearances)
opStructuredTracking.Disappearances.connect(
opAnnotations.Disappearances)
opStructuredTracking.MaxNumObj.connect(opCellClassification.MaxNumObj)
opDataTrackingExport.Inputs.resize(3)
opDataTrackingExport.Inputs[0].connect(opStructuredTracking.Output)
opDataTrackingExport.Inputs[1].connect(
opStructuredTracking.MergerOutput)
opDataTrackingExport.Inputs[2].connect(
opStructuredTracking.RelabeledImage)
opDataTrackingExport.RawData.connect(op5Raw.Output)
opDataTrackingExport.RawDatasetInfo.connect(opData.DatasetGroup[0])
def prepare_lane_for_export(self, lane_index):
import logging
logger = logging.getLogger(__name__)
maxt = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[0]
maxx = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[1]
maxy = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[2]
maxz = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[3]
time_enum = list(range(maxt))
x_range = (0, maxx)
y_range = (0, maxy)
z_range = (0, maxz)
ndim = 2
if (z_range[1] - z_range[0]) > 1:
ndim = 3
parameters = self.trackingApplet.topLevelOperator.Parameters.value
# Save state of axis ranges
if "time_range" in parameters:
self.prev_time_range = parameters["time_range"]
else:
self.prev_time_range = time_enum
if "x_range" in parameters:
self.prev_x_range = parameters["x_range"]
else:
self.prev_x_range = x_range
if "y_range" in parameters:
self.prev_y_range = parameters["y_range"]
else:
self.prev_y_range = y_range
if "z_range" in parameters:
self.prev_z_range = parameters["z_range"]
else:
self.prev_z_range = z_range
# batch processing starts a new lane, so training data needs to be copied from the lane that loaded the project
loaded_project_lane_index = 0
self.annotationsApplet.topLevelOperator[
lane_index].Annotations.setValue(
self.trackingApplet.
topLevelOperator[loaded_project_lane_index].Annotations.value)
def runLearningAndTracking(withMergerResolution=True):
if self.testFullAnnotations:
logger.info("Test: Structured Learning")
weights = self.trackingApplet.topLevelOperator[
lane_index]._runStructuredLearning(
z_range,
parameters["maxObj"],
parameters["max_nearest_neighbors"],
parameters["maxDist"],
parameters["divThreshold"],
[
parameters["scales"][0], parameters["scales"][1],
parameters["scales"][2]
],
parameters["size_range"],
parameters["withDivisions"],
parameters["borderAwareWidth"],
parameters["withClassifierPrior"],
withBatchProcessing=True,
)
logger.info("weights: {}".format(weights))
logger.info("Test: Tracking")
result = self.trackingApplet.topLevelOperator[lane_index].track(
time_range=time_enum,
x_range=x_range,
y_range=y_range,
z_range=z_range,
size_range=parameters["size_range"],
x_scale=parameters["scales"][0],
y_scale=parameters["scales"][1],
z_scale=parameters["scales"][2],
maxDist=parameters["maxDist"],
maxObj=parameters["maxObj"],
divThreshold=parameters["divThreshold"],
avgSize=parameters["avgSize"],
withTracklets=parameters["withTracklets"],
sizeDependent=parameters["sizeDependent"],
detWeight=parameters["detWeight"],
divWeight=parameters["divWeight"],
transWeight=parameters["transWeight"],
withDivisions=parameters["withDivisions"],
withOpticalCorrection=parameters["withOpticalCorrection"],
withClassifierPrior=parameters["withClassifierPrior"],
ndim=ndim,
withMergerResolution=withMergerResolution,
borderAwareWidth=parameters["borderAwareWidth"],
withArmaCoordinates=parameters["withArmaCoordinates"],
cplex_timeout=parameters["cplex_timeout"],
appearance_cost=parameters["appearanceCost"],
disappearance_cost=parameters["disappearanceCost"],
force_build_hypotheses_graph=False,
withBatchProcessing=True,
)
return result
if self.testFullAnnotations:
self.result = runLearningAndTracking(withMergerResolution=False)
hypothesesGraph = self.trackingApplet.topLevelOperator[
lane_index].LearningHypothesesGraph.value
hypothesesGraph.insertSolution(self.result)
hypothesesGraph.computeLineage()
solution = hypothesesGraph.getSolutionDictionary()
annotations = self.trackingApplet.topLevelOperator[
lane_index].Annotations.value
self.trackingApplet.topLevelOperator[
lane_index].insertAnnotationsToHypothesesGraph(
hypothesesGraph, annotations, misdetectionLabel=-1)
hypothesesGraph.computeLineage()
solutionFromAnnotations = hypothesesGraph.getSolutionDictionary()
for key in list(solution.keys()):
if key == "detectionResults":
detectionFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if (solution[key][i]["id"]
== solutionFromAnnotations[key][j]["id"]
and solution[key][i]["value"] ==
solutionFromAnnotations[key][j]["value"]):
flag = True
break
detectionFlag &= flag
elif key == "divisionResults":
divisionFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if (solution[key][i]["id"]
== solutionFromAnnotations[key][j]["id"]
and solution[key][i]["value"] ==
solutionFromAnnotations[key][j]["value"]):
flag = True
break
divisionFlag &= flag
elif key == "linkingResults":
linkingFlag = True
for i in range(len(solution[key])):
flag = False
for j in range(len(solutionFromAnnotations[key])):
if (solution[key][i]["dest"]
== solutionFromAnnotations[key][j]["dest"]
and solution[key][i]["src"]
== solutionFromAnnotations[key][j]["src"]):
if (solution[key][i]["gap"] ==
solutionFromAnnotations[key][j]["gap"]
and solution[key][i]["value"]
== solutionFromAnnotations[key][j]
["value"]):
flag = True
break
linkingFlag &= flag
assert detectionFlag, "Detection results are NOT correct. They differ from your annotated detections."
logger.info("Detection results are correct.")
assert divisionFlag, "Division results are NOT correct. They differ from your annotated divisions."
logger.info("Division results are correct.")
assert linkingFlag, "Transition results are NOT correct. They differ from your annotated transitions."
logger.info("Transition results are correct.")
self.result = runLearningAndTracking(
withMergerResolution=parameters["withMergerResolution"])
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and all(
[sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def onProjectLoaded(self, projectManager):
"""
Overridden from Workflow base class. Called by the Project Manager.
If the user provided command-line arguments, use them to configure
the workflow inputs and output settings.
"""
# Configure the data export operator.
if self._data_export_args:
self.dataExportTrackingApplet.configure_operator_with_parsed_args(
self._data_export_args)
# Configure headless mode.
if self._headless and self._batch_input_args and self._data_export_args:
logger.info("Beginning Batch Processing")
self.batchProcessingApplet.run_export_from_parsed_args(
self._batch_input_args)
logger.info("Completed Batch Processing")
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and len(thresholdingOutput) > 0
else:
thresholding_ready = input_ready
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator
trackingFeatureExtractionOutput = opTrackingFeatureExtraction.ComputedFeatureNamesAll
tracking_features_ready = thresholding_ready and len(
trackingFeatureExtractionOutput) > 0
objectCountClassifier_ready = tracking_features_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.RegionFeatures
features_ready = thresholding_ready and len(objectExtractionOutput) > 0
opAnnotations = self.annotationsApplet.topLevelOperator
annotations_ready = (features_ready and len(opAnnotations.Labels) > 0
and opAnnotations.Labels.ready()
and opAnnotations.TrackImage.ready())
opStructuredTracking = self.trackingApplet.topLevelOperator
structured_tracking_ready = objectCountClassifier_ready
withIlpSolver = self._solver == "ILP"
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.annotationsApplet.busy
# busy |= self.dataExportAnnotationsApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportTrackingApplet.busy
self._shell.enableProjectChanges(not busy)
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet,
input_ready and not busy)
self._shell.setAppletEnabled(self.trackingFeatureExtractionApplet,
thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet,
tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet,
tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, not busy)
self._shell.setAppletEnabled(self.annotationsApplet, features_ready
and not busy) # and withIlpSolver)
# self._shell.setAppletEnabled(self.dataExportAnnotationsApplet, annotations_ready and not busy and \
# self.dataExportAnnotationsApplet.topLevelOperator.Inputs[0][0].ready() )
self._shell.setAppletEnabled(self.trackingApplet,
objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(
self.dataExportTrackingApplet,
structured_tracking_ready and not busy
and self.dataExportTrackingApplet.topLevelOperator.Inputs[0]
[0].ready(),
)
def __init__( self, shell, headless, workflow_cmdline_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase, self).__init__(shell, headless, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
forceAxisOrder='txyzc',
instructionText=data_instructions,
max_lanes=1
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,
"Threshold and Size Filter",
"ThresholdTwoLevels" )
if self.withOptTrans:
self.opticalTranslationApplet = OpticalTranslationApplet(workflow=self)
self.objectExtractionApplet = TrackingFeatureExtractionApplet(workflow=self, interactive=False,
name="Object Feature Computation")
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection")
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification (optional)",
projectFileGroupName="CountClassification")
self.trackingApplet = ConservationTrackingApplet( workflow=self )
self.default_export_filename = '{dataset_dir}/{nickname}-exported_data.csv'
self.dataExportApplet = TrackingBaseDataExportApplet(self,
"Tracking Result Export",
default_export_filename=self.default_export_filename)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue( ['Object Identities', 'Tracking Result', 'Merger Result'] )
opDataExport.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
if self.withOptTrans:
self._applets.append(self.opticalTranslationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
class StructuredTrackingWorkflowBase( Workflow ):
workflowName = "Structured Learning Tracking Workflow BASE"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder='txyzc',
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configStructured.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configStructured.selectedFeaturesObjectCount)
self.cropSelectionApplet = CropSelectionApplet(self,"Crop Selection","CropSelection")
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
# self.default_training_export_filename = '{dataset_dir}/{nickname}-training_exported_data.csv'
# self.dataExportAnnotationsApplet = TrackingBaseDataExportApplet(self, "Training Export",default_export_filename=self.default_training_export_filename)
# opDataExportAnnotations = self.dataExportAnnotationsApplet.topLevelOperator
# opDataExportAnnotations.SelectionNames.setValue( ['User Training for Tracking', 'Object Identities'] )
# opDataExportAnnotations.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# self.dataExportAnnotationsApplet.set_exporting_operator(opAnnotations)
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export",default_export_filename=self.default_tracking_export_filename)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking Result', 'Merger Result', 'Object Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.cropSelectionApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
# self._applets.append(self.dataExportAnnotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator.getLane(laneIndex)
opAnnotations = self.annotationsApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
# opDataAnnotationsExport = self.dataExportAnnotationsApplet.topLevelOperator.getLane(laneIndex)
opCropSelection = self.cropSelectionApplet.topLevelOperator.getLane(laneIndex)
opStructuredTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataTrackingExport = self.dataExportTrackingApplet.topLevelOperator.getLane(laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect( opData.ImageGroup[1] )
opTwoLevelThreshold.RawInput.connect( opData.ImageGroup[0] ) # Used for display only
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
opCropSelection.InputImage.connect( opData.ImageGroup[0] )
opCropSelection.PredictionImage.connect( opData.ImageGroup[1] )
opObjExtraction.RawImage.connect( op5Raw.Output )
opObjExtraction.BinaryImage.connect( op5Binary.Output )
opTrackingFeatureExtraction.RawImage.connect( op5Raw.Output )
opTrackingFeatureExtraction.BinaryImage.connect( op5Binary.Output )
# vigra_features = list((set(config.vigra_features)).union(config.selected_features_objectcount[config.features_vigra_name]))
# feature_names_vigra = {}
# feature_names_vigra[config.features_vigra_name] = { name: {} for name in vigra_features }
opTrackingFeatureExtraction.FeatureNamesVigra.setValue(configConservation.allFeaturesObjectCount)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = { name: {} for name in config.division_features }
opTrackingFeatureExtraction.FeatureNamesDivision.setValue(feature_dict_division)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect( op5Binary.Output )
opDivDetection.RawImages.connect( op5Raw.Output )
opDivDetection.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect( op5Binary.Output )
opCellClassification.RawImages.connect( op5Raw.Output )
opCellClassification.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opCellClassification.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesNoDivisions)
opAnnotations.RawImage.connect( op5Raw.Output )
opAnnotations.BinaryImage.connect( op5Binary.Output )
opAnnotations.LabelImage.connect( opObjExtraction.LabelImage )
opAnnotations.ObjectFeatures.connect( opObjExtraction.RegionFeatures )
opAnnotations.ComputedFeatureNames.connect(opObjExtraction.Features)
opAnnotations.Crops.connect( opCropSelection.Crops)
# opDataAnnotationsExport.Inputs.resize(2)
# opDataAnnotationsExport.Inputs[0].connect( opAnnotations.TrackImage )
# opDataAnnotationsExport.Inputs[1].connect( opAnnotations.LabelImage )
# opDataAnnotationsExport.RawData.connect( op5Raw.Output )
# opDataAnnotationsExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
opStructuredTracking.RawImage.connect( op5Raw.Output )
opStructuredTracking.LabelImage.connect( opTrackingFeatureExtraction.LabelImage )
opStructuredTracking.ObjectFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesVigra )
opStructuredTracking.ComputedFeatureNames.connect( opTrackingFeatureExtraction.FeatureNamesVigra )
if self.divisionDetectionApplet:
opStructuredTracking.ObjectFeaturesWithDivFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesAll)
opStructuredTracking.ComputedFeatureNamesWithDivFeatures.connect( opTrackingFeatureExtraction.ComputedFeatureNamesAll )
opStructuredTracking.DivisionProbabilities.connect( opDivDetection.Probabilities )
# configure tracking export settings
settings = {'file path': self.default_tracking_export_filename, 'compression': {}, 'file type': 'csv'}
selected_features = ['Count', 'RegionCenter']
opStructuredTracking.configure_table_export_settings(settings, selected_features)
opStructuredTracking.DetectionProbabilities.connect( opCellClassification.Probabilities )
opStructuredTracking.NumLabels.connect( opCellClassification.NumLabels )
opStructuredTracking.Crops.connect (opCropSelection.Crops)
opStructuredTracking.Annotations.connect (opAnnotations.Annotations)
opStructuredTracking.Labels.connect (opAnnotations.Labels)
opStructuredTracking.Divisions.connect (opAnnotations.Divisions)
opStructuredTracking.MaxNumObj.connect (opCellClassification.MaxNumObj)
opDataTrackingExport.Inputs.resize(3)
opDataTrackingExport.Inputs[0].connect( opStructuredTracking.Output )
opDataTrackingExport.Inputs[1].connect( opStructuredTracking.MergerOutput )
opDataTrackingExport.Inputs[2].connect( opStructuredTracking.LabelImage )
opDataTrackingExport.RawData.connect( op5Raw.Output )
opDataTrackingExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def post_process_lane_export(self, lane_index):
# FIXME: This probably only works for the non-blockwise export slot.
# We should assert that the user isn't using the blockwise slot.
settings, selected_features = self.trackingApplet.topLevelOperator.getLane(lane_index).get_table_export_settings()
from lazyflow.utility import PathComponents, make_absolute, format_known_keys
if settings:
self.dataExportTrackingApplet.progressSignal.emit(-1)
raw_dataset_info = self.dataSelectionApplet.topLevelOperator.DatasetGroup[lane_index][0].value
project_path = self.shell.projectManager.currentProjectPath
project_dir = os.path.dirname(project_path)
dataset_dir = PathComponents(raw_dataset_info.filePath).externalDirectory
abs_dataset_dir = make_absolute(dataset_dir, cwd=project_dir)
known_keys = {}
known_keys['dataset_dir'] = abs_dataset_dir
nickname = raw_dataset_info.nickname.replace('*', '')
if os.path.pathsep in nickname:
nickname = PathComponents(nickname.split(os.path.pathsep)[0]).fileNameBase
known_keys['nickname'] = nickname
# use partial formatting to fill in non-coordinate name fields
name_format = settings['file path']
partially_formatted_name = format_known_keys( name_format, known_keys )
settings['file path'] = partially_formatted_name
req = self.trackingApplet.topLevelOperator.getLane(lane_index).export_object_data(
lane_index,
# FIXME: Even in non-headless mode, we can't show the gui because we're running in a non-main thread.
# That's not a huge deal, because there's still a progress bar for the overall export.
show_gui=False)
req.wait()
self.dataExportTrackingApplet.progressSignal.emit(100)
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and len(thresholdingOutput) > 0
else:
thresholding_ready = input_ready
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator
trackingFeatureExtractionOutput = opTrackingFeatureExtraction.ComputedFeatureNamesAll
tracking_features_ready = thresholding_ready and len(trackingFeatureExtractionOutput) > 0
opCropSelection = self.cropSelectionApplet.topLevelOperator
croppingOutput = opCropSelection.Crops
cropping_ready = thresholding_ready and len(croppingOutput) > 0
objectCountClassifier_ready = tracking_features_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.RegionFeatures
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
opAnnotations = self.annotationsApplet.topLevelOperator
annotations_ready = features_ready and \
len(opAnnotations.Labels) > 0 and \
opAnnotations.Labels.ready() and \
opAnnotations.TrackImage.ready()
opStructuredTracking = self.trackingApplet.topLevelOperator
structured_tracking_ready = objectCountClassifier_ready and \
len(opStructuredTracking.EventsVector) > 0
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.annotationsApplet.busy
# busy |= self.dataExportAnnotationsApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportTrackingApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.trackingFeatureExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.cropSelectionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, not busy)
self._shell.setAppletEnabled(self.annotationsApplet, features_ready and not busy)
# self._shell.setAppletEnabled(self.dataExportAnnotationsApplet, annotations_ready and not busy and \
# self.dataExportAnnotationsApplet.topLevelOperator.Inputs[0][0].ready() )
self._shell.setAppletEnabled(self.trackingApplet, objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportTrackingApplet, structured_tracking_ready and not busy and \
self.dataExportTrackingApplet.topLevelOperator.Inputs[0][0].ready() )
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder='txyzc',
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configStructured.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configStructured.selectedFeaturesObjectCount)
self.cropSelectionApplet = CropSelectionApplet(self,"Crop Selection","CropSelection")
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
# self.default_training_export_filename = '{dataset_dir}/{nickname}-training_exported_data.csv'
# self.dataExportAnnotationsApplet = TrackingBaseDataExportApplet(self, "Training Export",default_export_filename=self.default_training_export_filename)
# opDataExportAnnotations = self.dataExportAnnotationsApplet.topLevelOperator
# opDataExportAnnotations.SelectionNames.setValue( ['User Training for Tracking', 'Object Identities'] )
# opDataExportAnnotations.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# self.dataExportAnnotationsApplet.set_exporting_operator(opAnnotations)
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export",default_export_filename=self.default_tracking_export_filename)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking Result', 'Merger Result', 'Object Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.cropSelectionApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
# self._applets.append(self.dataExportAnnotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configConservation.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount)
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
if SOLVER=="CPLEX" or SOLVER=="GUROBI":
self._solver="ILP"
elif SOLVER=="DPCT":
self._solver="Flow-based"
else:
self._solver=None
opStructuredTracking._solver = self._solver
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_tracking_export_filename,
pluginExportFunc=self._pluginExportFunc
)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking-Result', 'Merger-Result', 'Object-Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.prepare_lane_for_export = self.prepare_lane_for_export
# configure export settings
settings = {'file path': self.default_tracking_export_filename, 'compression': {}, 'file type': 'h5'}
selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
opStructuredTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(self, "Batch Processing", self.dataSelectionApplet, self.dataExportTrackingApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
if workflow_cmdline_args:
if '--testFullAnnotations' in workflow_cmdline_args:
self.testFullAnnotations = True
else:
self.testFullAnnotations = False
self._data_export_args, unused_args = self.dataExportTrackingApplet.parse_known_cmdline_args( workflow_cmdline_args )
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args( workflow_cmdline_args )
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
self.testFullAnnotations = False
if unused_args:
logger.warning("Unused command-line args: {}".format( unused_args ))
class ConservationTrackingWorkflowBase(Workflow):
workflowName = "Automatic Tracking Workflow (Conservation Tracking) BASE"
def __init__(self, shell, headless, workflow_cmdline_args, *args,
**kwargs):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase, self).__init__(shell,
headless,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
batchDataGui=False,
force5d=True,
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue(['Raw Data', 'Binary Image'])
else:
opDataSelection.DatasetRoles.setValue(
['Raw Data', 'Prediction Maps'])
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
if self.withOptTrans:
self.opticalTranslationApplet = OpticalTranslationApplet(
workflow=self)
self.objectExtractionApplet = TrackingFeatureExtractionApplet(
workflow=self,
interactive=False,
name="Object Feature Computation")
self.divisionDetectionApplet = ObjectClassificationApplet(
workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection")
self.cellClassificationApplet = ObjectClassificationApplet(
workflow=self,
name="Object Count Classification (optional)",
projectFileGroupName="CountClassification")
self.trackingApplet = ConservationTrackingApplet(workflow=self)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet = TrackingBaseDataExportApplet(
self, "Tracking Result Export")
self.dataExportApplet.set_exporting_operator(opTracking)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
['Tracking Result', 'Merger Result', 'Object Identities'])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
if self.withOptTrans:
self._applets.append(self.opticalTranslationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(
laneIndex)
if self.withOptTrans:
opOptTranslation = self.opticalTranslationApplet.topLevelOperator.getLane(
laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(
laneIndex)
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(
laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(
laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(
laneIndex)
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(
opData.ImageGroup[0]) # Used for display only
# opTwoLevelThreshold.Channel.setValue(1)
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
if self.withOptTrans:
opOptTranslation.RawImage.connect(op5Raw.Output)
opOptTranslation.BinaryImage.connect(op5Binary.Output)
# # Connect operators ##
vigra_features = list((set(config.vigra_features)).union(
config.selected_features_objectcount[config.features_vigra_name]))
feature_names_vigra = {}
feature_names_vigra[config.features_vigra_name] = {
name: {}
for name in vigra_features
}
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
if self.withOptTrans:
opObjExtraction.TranslationVectors.connect(
opOptTranslation.TranslationVectors)
opObjExtraction.FeatureNamesVigra.setValue(feature_names_vigra)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = {
name: {}
for name in config.division_features
}
opObjExtraction.FeatureNamesDivision.setValue(feature_dict_division)
selected_features_div = {}
for plugin_name in config.selected_features_division.keys():
selected_features_div[plugin_name] = {
name: {}
for name in config.selected_features_division[plugin_name]
}
# FIXME: do not hard code this
for name in [
'SquaredDistances_' + str(i)
for i in range(config.n_best_successors)
]:
selected_features_div[config.features_division_name][name] = {}
opDivDetection.BinaryImages.connect(op5Binary.Output)
opDivDetection.RawImages.connect(op5Raw.Output)
opDivDetection.LabelsAllowedFlags.connect(opData.AllowLabels)
opDivDetection.SegmentationImages.connect(opObjExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(
opObjExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(
opObjExtraction.ComputedFeatureNamesAll)
opDivDetection.SelectedFeatures.setValue(selected_features_div)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
selected_features_objectcount = {}
for plugin_name in config.selected_features_objectcount.keys():
selected_features_objectcount[plugin_name] = {
name: {}
for name in config.selected_features_objectcount[plugin_name]
}
opCellClassification.BinaryImages.connect(op5Binary.Output)
opCellClassification.RawImages.connect(op5Raw.Output)
opCellClassification.LabelsAllowedFlags.connect(opData.AllowLabels)
opCellClassification.SegmentationImages.connect(
opObjExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(
opObjExtraction.RegionFeaturesVigra)
opCellClassification.ComputedFeatureNames.connect(
opObjExtraction.ComputedFeatureNamesVigra)
opCellClassification.SelectedFeatures.setValue(
selected_features_objectcount)
opCellClassification.SuggestedLabelNames.setValue([
'false detection',
] + [str(i) + ' Objects' for i in range(1, 10)])
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
opTracking.RawImage.connect(op5Raw.Output)
opTracking.LabelImage.connect(opObjExtraction.LabelImage)
opTracking.ObjectFeatures.connect(opObjExtraction.RegionFeaturesVigra)
opTracking.ObjectFeaturesWithDivFeatures.connect(
opObjExtraction.RegionFeaturesAll)
opTracking.ComputedFeatureNames.connect(
opObjExtraction.ComputedFeatureNamesVigra)
opTracking.ComputedFeatureNamesWithDivFeatures.connect(
opObjExtraction.ComputedFeatureNamesAll)
opTracking.DivisionProbabilities.connect(opDivDetection.Probabilities)
opTracking.DetectionProbabilities.connect(
opCellClassification.Probabilities)
opTracking.NumLabels.connect(opCellClassification.NumLabels)
opDataExport.Inputs.resize(3)
opDataExport.Inputs[0].connect(opTracking.Output)
opDataExport.Inputs[1].connect(opTracking.MergerOutput)
opDataExport.Inputs[2].connect(opTracking.LabelImage)
opDataExport.RawData.connect(op5Raw.Output)
opDataExport.RawDatasetInfo.connect(opData.DatasetGroup[0])
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
opDataSelection = self.dataSelectionApplet.topLevelOperator
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and \
len(thresholdingOutput) > 0
else:
thresholding_ready = True and input_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.ComputedFeatureNamesAll
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
objectCountClassifier_ready = features_ready
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = objectCountClassifier_ready and \
len(opTracking.EventsVector) > 0
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportApplet.busy
self._shell.enableProjectChanges(not busy)
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet,
input_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet,
thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet,
features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet,
features_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet,
objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
class ManualTrackingWorkflow(Workflow):
workflowName = "Manual Tracking Workflow"
workflowDisplayName = "Manual Tracking Workflow [Inputs: Raw Data, Pixel Prediction Map]"
workflowDescription = "Manual tracking of objects, based on Prediction Maps or (binary) Segmentation Images"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__(self, shell, headless, workflow_cmdline_args,
project_creation_args, *args, **kwargs):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(ManualTrackingWorkflow, self).__init__(shell,
headless,
workflow_cmdline_args,
project_creation_args,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'\
'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
batchDataGui=False,
force5d=True,
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
opDataSelection.DatasetRoles.setValue(['Raw Data', 'Prediction Maps'])
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
self.objectExtractionApplet = ObjectExtractionApplet(
name="Object Feature Computation",
workflow=self,
interactive=False)
self.trackingApplet = ManualTrackingApplet(workflow=self)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet = TrackingBaseDataExportApplet(
self, "Tracking Result Export")
self.dataExportApplet.set_exporting_operator(opTracking)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
['Manual Tracking', 'Object Identities'])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
self._applets = []
self._applets.append(self.dataSelectionApplet)
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(
laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(
laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(
laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(
opData.ImageGroup[0]) # Used for display only
# Use OpReorderAxis for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(opTwoLevelThreshold.CachedOutput)
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
opTracking.RawImage.connect(op5Raw.Output)
opTracking.BinaryImage.connect(op5Binary.Output)
opTracking.LabelImage.connect(opObjExtraction.LabelImage)
opTracking.ObjectFeatures.connect(opObjExtraction.RegionFeatures)
opTracking.ComputedFeatureNames.connect(
opObjExtraction.ComputedFeatureNames)
opDataExport.Inputs.resize(2)
opDataExport.Inputs[0].connect(opTracking.TrackImage)
opDataExport.Inputs[1].connect(opTracking.LabelImage)
opDataExport.RawData.connect(op5Raw.Output)
opDataExport.RawDatasetInfo.connect(opData.DatasetGroup[0])
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and \
len(thresholdingOutput) > 0
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.ComputedFeatureNames
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = features_ready and \
len(opTracking.Labels) > 0 and \
opTracking.Labels.ready() and \
opTracking.TrackImage.ready()
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.dataExportApplet.busy
busy |= self.trackingApplet.busy
self._shell.enableProjectChanges(not busy)
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready
and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet,
thresholding_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet, features_ready
and not busy)
self._shell.setAppletEnabled(self.dataExportApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
def __init__(self, shell, headless, workflow_cmdline_args, *args,
**kwargs):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase, self).__init__(shell,
headless,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
batchDataGui=False,
force5d=True,
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue(['Raw Data', 'Binary Image'])
else:
opDataSelection.DatasetRoles.setValue(
['Raw Data', 'Prediction Maps'])
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
if self.withOptTrans:
self.opticalTranslationApplet = OpticalTranslationApplet(
workflow=self)
self.objectExtractionApplet = TrackingFeatureExtractionApplet(
workflow=self,
interactive=False,
name="Object Feature Computation")
self.divisionDetectionApplet = ObjectClassificationApplet(
workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection")
self.cellClassificationApplet = ObjectClassificationApplet(
workflow=self,
name="Object Count Classification (optional)",
projectFileGroupName="CountClassification")
self.trackingApplet = ConservationTrackingApplet(workflow=self)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet = TrackingBaseDataExportApplet(
self, "Tracking Result Export")
self.dataExportApplet.set_exporting_operator(opTracking)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
['Tracking Result', 'Merger Result', 'Object Identities'])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
if self.withOptTrans:
self._applets.append(self.opticalTranslationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
class StructuredTrackingWorkflowBase( Workflow ):
workflowName = "Structured Learning Tracking Workflow BASE"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configStructured.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configStructured.selectedFeaturesObjectCount)
self.cropSelectionApplet = CropSelectionApplet(self,"Crop Selection","CropSelection")
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
# self.default_training_export_filename = '{dataset_dir}/{nickname}-training_exported_data.csv'
# self.dataExportAnnotationsApplet = TrackingBaseDataExportApplet(self, "Training Export",default_export_filename=self.default_training_export_filename)
# opDataExportAnnotations = self.dataExportAnnotationsApplet.topLevelOperator
# opDataExportAnnotations.SelectionNames.setValue( ['User Training for Tracking', 'Object Identities'] )
# opDataExportAnnotations.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# self.dataExportAnnotationsApplet.set_exporting_operator(opAnnotations)
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export",default_export_filename=self.default_tracking_export_filename)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking Result', 'Merger Result', 'Object Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.cropSelectionApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
# self._applets.append(self.dataExportAnnotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator.getLane(laneIndex)
opAnnotations = self.annotationsApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
# opDataAnnotationsExport = self.dataExportAnnotationsApplet.topLevelOperator.getLane(laneIndex)
opCropSelection = self.cropSelectionApplet.topLevelOperator.getLane(laneIndex)
opStructuredTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataTrackingExport = self.dataExportTrackingApplet.topLevelOperator.getLane(laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect( opData.ImageGroup[1] )
opTwoLevelThreshold.RawInput.connect( opData.ImageGroup[0] ) # Used for display only
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
opCropSelection.InputImage.connect( opData.ImageGroup[0] )
opCropSelection.PredictionImage.connect( opData.ImageGroup[1] )
opObjExtraction.RawImage.connect( op5Raw.Output )
opObjExtraction.BinaryImage.connect( op5Binary.Output )
opTrackingFeatureExtraction.RawImage.connect( op5Raw.Output )
opTrackingFeatureExtraction.BinaryImage.connect( op5Binary.Output )
# vigra_features = list((set(config.vigra_features)).union(config.selected_features_objectcount[config.features_vigra_name]))
# feature_names_vigra = {}
# feature_names_vigra[config.features_vigra_name] = { name: {} for name in vigra_features }
opTrackingFeatureExtraction.FeatureNamesVigra.setValue(configConservation.allFeaturesObjectCount)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = { name: {} for name in config.division_features }
opTrackingFeatureExtraction.FeatureNamesDivision.setValue(feature_dict_division)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect( op5Binary.Output )
opDivDetection.RawImages.connect( op5Raw.Output )
opDivDetection.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect( op5Binary.Output )
opCellClassification.RawImages.connect( op5Raw.Output )
opCellClassification.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opCellClassification.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesNoDivisions)
opAnnotations.RawImage.connect( op5Raw.Output )
opAnnotations.BinaryImage.connect( op5Binary.Output )
opAnnotations.LabelImage.connect( opObjExtraction.LabelImage )
opAnnotations.ObjectFeatures.connect( opObjExtraction.RegionFeatures )
opAnnotations.ComputedFeatureNames.connect(opObjExtraction.Features)
opAnnotations.Crops.connect( opCropSelection.Crops)
# opDataAnnotationsExport.Inputs.resize(2)
# opDataAnnotationsExport.Inputs[0].connect( opAnnotations.TrackImage )
# opDataAnnotationsExport.Inputs[1].connect( opAnnotations.LabelImage )
# opDataAnnotationsExport.RawData.connect( op5Raw.Output )
# opDataAnnotationsExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
opStructuredTracking.RawImage.connect( op5Raw.Output )
opStructuredTracking.LabelImage.connect( opTrackingFeatureExtraction.LabelImage )
opStructuredTracking.ObjectFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesVigra )
opStructuredTracking.ComputedFeatureNames.connect( opTrackingFeatureExtraction.FeatureNamesVigra )
if self.divisionDetectionApplet:
opStructuredTracking.ObjectFeaturesWithDivFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesAll)
opStructuredTracking.ComputedFeatureNamesWithDivFeatures.connect( opTrackingFeatureExtraction.ComputedFeatureNamesAll )
opStructuredTracking.DivisionProbabilities.connect( opDivDetection.Probabilities )
# configure tracking export settings
settings = {'file path': self.default_tracking_export_filename, 'compression': {}, 'file type': 'csv'}
selected_features = ['Count', 'RegionCenter']
opStructuredTracking.configure_table_export_settings(settings, selected_features)
opStructuredTracking.DetectionProbabilities.connect( opCellClassification.Probabilities )
opStructuredTracking.NumLabels.connect( opCellClassification.NumLabels )
opStructuredTracking.Crops.connect (opCropSelection.Crops)
opStructuredTracking.Annotations.connect (opAnnotations.Annotations)
opStructuredTracking.Labels.connect (opAnnotations.Labels)
opStructuredTracking.Divisions.connect (opAnnotations.Divisions)
opStructuredTracking.MaxNumObj.connect (opCellClassification.MaxNumObj)
opDataTrackingExport.Inputs.resize(3)
opDataTrackingExport.Inputs[0].connect( opStructuredTracking.Output )
opDataTrackingExport.Inputs[1].connect( opStructuredTracking.MergerOutput )
opDataTrackingExport.Inputs[2].connect( opStructuredTracking.LabelImage )
opDataTrackingExport.RawData.connect( op5Raw.Output )
opDataTrackingExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def post_process_lane_export(self, lane_index):
# FIXME: This probably only works for the non-blockwise export slot.
# We should assert that the user isn't using the blockwise slot.
settings, selected_features = self.trackingApplet.topLevelOperator.getLane(lane_index).get_table_export_settings()
from lazyflow.utility import PathComponents, make_absolute, format_known_keys
if settings:
self.dataExportTrackingApplet.progressSignal.emit(-1)
raw_dataset_info = self.dataSelectionApplet.topLevelOperator.DatasetGroup[lane_index][0].value
project_path = self.shell.projectManager.currentProjectPath
project_dir = os.path.dirname(project_path)
dataset_dir = PathComponents(raw_dataset_info.filePath).externalDirectory
abs_dataset_dir = make_absolute(dataset_dir, cwd=project_dir)
known_keys = {}
known_keys['dataset_dir'] = abs_dataset_dir
nickname = raw_dataset_info.nickname.replace('*', '')
if os.path.pathsep in nickname:
nickname = PathComponents(nickname.split(os.path.pathsep)[0]).fileNameBase
known_keys['nickname'] = nickname
# use partial formatting to fill in non-coordinate name fields
name_format = settings['file path']
partially_formatted_name = format_known_keys( name_format, known_keys )
settings['file path'] = partially_formatted_name
req = self.trackingApplet.topLevelOperator.getLane(lane_index).export_object_data(
lane_index,
# FIXME: Even in non-headless mode, we can't show the gui because we're running in a non-main thread.
# That's not a huge deal, because there's still a progress bar for the overall export.
show_gui=False)
req.wait()
self.dataExportTrackingApplet.progressSignal.emit(100)
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and len(thresholdingOutput) > 0
else:
thresholding_ready = input_ready
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator
trackingFeatureExtractionOutput = opTrackingFeatureExtraction.ComputedFeatureNamesAll
tracking_features_ready = thresholding_ready and len(trackingFeatureExtractionOutput) > 0
opCropSelection = self.cropSelectionApplet.topLevelOperator
croppingOutput = opCropSelection.Crops
cropping_ready = thresholding_ready and len(croppingOutput) > 0
objectCountClassifier_ready = tracking_features_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.RegionFeatures
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
opAnnotations = self.annotationsApplet.topLevelOperator
annotations_ready = features_ready and \
len(opAnnotations.Labels) > 0 and \
opAnnotations.Labels.ready() and \
opAnnotations.TrackImage.ready()
opStructuredTracking = self.trackingApplet.topLevelOperator
structured_tracking_ready = objectCountClassifier_ready and \
len(opStructuredTracking.EventsVector) > 0
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.annotationsApplet.busy
# busy |= self.dataExportAnnotationsApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportTrackingApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.trackingFeatureExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.cropSelectionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, not busy)
self._shell.setAppletEnabled(self.annotationsApplet, features_ready and not busy)
# self._shell.setAppletEnabled(self.dataExportAnnotationsApplet, annotations_ready and not busy and \
# self.dataExportAnnotationsApplet.topLevelOperator.Inputs[0][0].ready() )
self._shell.setAppletEnabled(self.trackingApplet, objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportTrackingApplet, structured_tracking_ready and not busy and \
self.dataExportTrackingApplet.topLevelOperator.Inputs[0][0].ready() )
class ConservationTrackingWorkflowBase(Workflow):
workflowName = "Automatic Tracking Workflow (Conservation Tracking) BASE"
def __init__(self, shell, headless, workflow_cmdline_args,
project_creation_args, *args, **kwargs):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase,
self).__init__(shell,
headless,
workflow_cmdline_args,
project_creation_args,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Variables to store division and cell classifiers to prevent retraining every-time batch processing runs
self.stored_division_classifier = None
self.stored_cell_classifier = None
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=None)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue(
['Raw Data', 'Segmentation Image'])
else:
opDataSelection.DatasetRoles.setValue(
['Raw Data', 'Prediction Maps'])
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
if self.withOptTrans:
self.opticalTranslationApplet = OpticalTranslationApplet(
workflow=self)
self.objectExtractionApplet = TrackingFeatureExtractionApplet(
workflow=self,
interactive=False,
name="Object Feature Computation")
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
opObjectExtraction.FeatureNamesVigra.setValue(
configConservation.allFeaturesObjectCount)
self.divisionDetectionApplet = self._createDivisionDetectionApplet(
configConservation.selectedFeaturesDiv) # Might be None
if self.divisionDetectionApplet:
feature_dict_division = {}
feature_dict_division[config.features_division_name] = {
name: {}
for name in config.division_features
}
opObjectExtraction.FeatureNamesDivision.setValue(
feature_dict_division)
selected_features_div = {}
for plugin_name in config.selected_features_division.keys():
selected_features_div[plugin_name] = {
name: {}
for name in config.selected_features_division[plugin_name]
}
# FIXME: do not hard code this
for name in [
'SquaredDistances_' + str(i)
for i in range(config.n_best_successors)
]:
selected_features_div[config.features_division_name][name] = {}
opDivisionDetection = self.divisionDetectionApplet.topLevelOperator
opDivisionDetection.SelectedFeatures.setValue(
configConservation.selectedFeaturesDiv)
opDivisionDetection.LabelNames.setValue(
['Not Dividing', 'Dividing'])
opDivisionDetection.AllowDeleteLabels.setValue(False)
opDivisionDetection.AllowAddLabel.setValue(False)
opDivisionDetection.EnableLabelTransfer.setValue(False)
self.cellClassificationApplet = ObjectClassificationApplet(
workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount)
selected_features_objectcount = {}
for plugin_name in config.selected_features_objectcount.keys():
selected_features_objectcount[plugin_name] = {
name: {}
for name in config.selected_features_objectcount[plugin_name]
}
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(
configConservation.selectedFeaturesObjectCount)
opCellClassification.SuggestedLabelNames.setValue([
'False Detection',
] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2, 10)])
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
self.trackingApplet = ConservationTrackingApplet(workflow=self)
self.default_export_filename = '{dataset_dir}/{nickname}-exported_data.csv'
self.dataExportApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_export_filename)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
['Object-Identities', 'Tracking-Result', 'Merger-Result'])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.prepare_lane_for_export = self.prepare_lane_for_export
self.dataExportApplet.post_process_lane_export = self.post_process_lane_export
# table only export is just available for the pgmlink backend, hytra uses the CSV plugin instead
try:
import hytra
except ImportError:
self.dataExportApplet.includeTableOnlyOption(
) # Export table only, without volumes
# configure export settings
settings = {
'file path': self.default_export_filename,
'compression': {},
'file type': 'csv'
}
selected_features = [
'Count', 'RegionCenter', 'RegionRadii', 'RegionAxes'
]
opTracking.ExportSettings.setValue((settings, selected_features))
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
if self.withOptTrans:
self._applets.append(self.opticalTranslationApplet)
self._applets.append(self.objectExtractionApplet)
if self.divisionDetectionApplet:
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(
self, "Batch Processing", self.dataSelectionApplet,
self.dataExportApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
self._applets.append(self.batchProcessingApplet)
# Parse export and batch command-line arguments for headless mode
if workflow_cmdline_args:
self._data_export_args, unused_args = self.dataExportApplet.parse_known_cmdline_args(
workflow_cmdline_args)
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args(
workflow_cmdline_args)
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
if unused_args:
logger.warn("Unused command-line args: {}".format(unused_args))
@property
def applets(self):
return self._applets
def _createDivisionDetectionApplet(self, selectedFeatures=dict()):
return ObjectClassificationApplet(
workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=selectedFeatures)
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def prepareForNewLane(self, laneIndex):
# Store division and cell classifiers
if self.divisionDetectionApplet:
opDivisionClassification = self.divisionDetectionApplet.topLevelOperator
if opDivisionClassification.classifier_cache.Output.ready() and \
not opDivisionClassification.classifier_cache._dirty:
self.stored_division_classifier = opDivisionClassification.classifier_cache.Output.value
else:
self.stored_division_classifier = None
opCellClassification = self.cellClassificationApplet.topLevelOperator
if opCellClassification.classifier_cache.Output.ready() and \
not opCellClassification.classifier_cache._dirty:
self.stored_cell_classifier = opCellClassification.classifier_cache.Output.value
else:
self.stored_cell_classifier = None
def handleNewLanesAdded(self):
"""
If new lanes were added, then we invalidated our classifiers unecessarily.
Here, we can restore the classifier so it doesn't need to be retrained.
"""
# If we have stored division and cell classifiers, restore them into the workflow now.
if self.stored_division_classifier:
opDivisionClassification = self.divisionDetectionApplet.topLevelOperator
opDivisionClassification.classifier_cache.forceValue(
self.stored_division_classifier)
# Release reference
self.stored_division_classifier = None
# If we have stored division and cell classifiers, restore them into the workflow now.
if self.stored_cell_classifier:
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.classifier_cache.forceValue(
self.stored_cell_classifier)
# Release reference
self.stored_cell_classifier = None
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(
laneIndex)
if self.withOptTrans:
opOptTranslation = self.opticalTranslationApplet.topLevelOperator.getLane(
laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(
laneIndex)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(
laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(
laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(
laneIndex)
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(
opData.ImageGroup[0]) # Used for display only
# opTwoLevelThreshold.Channel.setValue(1)
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
if self.withOptTrans:
opOptTranslation.RawImage.connect(op5Raw.Output)
opOptTranslation.BinaryImage.connect(op5Binary.Output)
# # Connect operators ##
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
if self.withOptTrans:
opObjExtraction.TranslationVectors.connect(
opOptTranslation.TranslationVectors)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect(op5Binary.Output)
opDivDetection.RawImages.connect(op5Raw.Output)
opDivDetection.SegmentationImages.connect(
opObjExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(
opObjExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(
opObjExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect(op5Binary.Output)
opCellClassification.RawImages.connect(op5Raw.Output)
opCellClassification.SegmentationImages.connect(
opObjExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(
opObjExtraction.RegionFeaturesVigra)
opCellClassification.ComputedFeatureNames.connect(
opObjExtraction.FeatureNamesVigra)
if self.divisionDetectionApplet:
opTracking.ObjectFeaturesWithDivFeatures.connect(
opObjExtraction.RegionFeaturesAll)
opTracking.ComputedFeatureNamesWithDivFeatures.connect(
opObjExtraction.ComputedFeatureNamesAll)
opTracking.DivisionProbabilities.connect(
opDivDetection.Probabilities)
opTracking.RawImage.connect(op5Raw.Output)
opTracking.LabelImage.connect(opObjExtraction.LabelImage)
opTracking.ObjectFeatures.connect(opObjExtraction.RegionFeaturesVigra)
opTracking.ComputedFeatureNames.connect(
opObjExtraction.FeatureNamesVigra)
opTracking.DetectionProbabilities.connect(
opCellClassification.Probabilities)
opTracking.NumLabels.connect(opCellClassification.NumLabels)
opDataExport.Inputs.resize(3)
opDataExport.Inputs[0].connect(opTracking.RelabeledImage)
opDataExport.Inputs[1].connect(opTracking.Output)
opDataExport.Inputs[2].connect(opTracking.MergerOutput)
opDataExport.RawData.connect(op5Raw.Output)
opDataExport.RawDatasetInfo.connect(opData.DatasetGroup[0])
def prepare_lane_for_export(self, lane_index):
# Bypass cache on headless mode and batch processing mode
self.objectExtractionApplet.topLevelOperator[
lane_index].BypassModeEnabled.setValue(True)
# Get axes info
maxt = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[0]
maxx = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[1]
maxy = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[2]
maxz = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[3]
time_enum = range(maxt)
x_range = (0, maxx)
y_range = (0, maxy)
z_range = (0, maxz)
ndim = 2
if (z_range[1] - z_range[0]) > 1:
ndim = 3
parameters = self.trackingApplet.topLevelOperator.Parameters.value
# Save state of axis ranges
if 'time_range' in parameters:
self.prev_time_range = parameters['time_range']
else:
self.prev_time_range = time_enum
if 'x_range' in parameters:
self.prev_x_range = parameters['x_range']
else:
self.prev_x_range = x_range
if 'y_range' in parameters:
self.prev_y_range = parameters['y_range']
else:
self.prev_y_range = y_range
if 'z_range' in parameters:
self.prev_z_range = parameters['z_range']
else:
self.prev_z_range = z_range
self.trackingApplet.topLevelOperator[lane_index].track(
time_range=time_enum,
x_range=x_range,
y_range=y_range,
z_range=z_range,
size_range=parameters['size_range'],
x_scale=parameters['scales'][0],
y_scale=parameters['scales'][1],
z_scale=parameters['scales'][2],
maxDist=parameters['maxDist'],
maxObj=parameters['maxObj'],
divThreshold=parameters['divThreshold'],
avgSize=parameters['avgSize'],
withTracklets=parameters['withTracklets'],
sizeDependent=parameters['sizeDependent'],
divWeight=parameters['divWeight'],
transWeight=parameters['transWeight'],
withDivisions=parameters['withDivisions'],
withOpticalCorrection=parameters['withOpticalCorrection'],
withClassifierPrior=parameters['withClassifierPrior'],
ndim=ndim,
withMergerResolution=parameters['withMergerResolution'],
borderAwareWidth=parameters['borderAwareWidth'],
withArmaCoordinates=parameters['withArmaCoordinates'],
cplex_timeout=parameters['cplex_timeout'],
appearance_cost=parameters['appearanceCost'],
disappearance_cost=parameters['disappearanceCost'],
max_nearest_neighbors=parameters['max_nearest_neighbors'],
numFramesPerSplit=parameters['numFramesPerSplit'],
force_build_hypotheses_graph=False,
withBatchProcessing=True)
def post_process_lane_export(self, lane_index, checkOverwriteFiles=False):
# `time` parameter ensures we check only once for files that could be overwritten, pop up
# the MessageBox and then don't export (time=0). For the next round we click the export button,
# we really want it to export, so time=1. The default parameter is 1, so everything but not 0,
# in order to ensure writing out even in headless mode.
# FIXME: This probably only works for the non-blockwise export slot.
# We should assert that the user isn't using the blockwise slot.
# Plugin export if selected
logger.info(
"Export source is: " +
self.dataExportApplet.topLevelOperator.SelectedExportSource.value)
if self.dataExportApplet.topLevelOperator.SelectedExportSource.value == OpTrackingBaseDataExport.PluginOnlyName:
logger.info("Export source plugin selected!")
selectedPlugin = self.dataExportApplet.topLevelOperator.SelectedPlugin.value
exportPluginInfo = pluginManager.getPluginByName(
selectedPlugin, category="TrackingExportFormats")
if exportPluginInfo is None:
logger.error("Could not find selected plugin %s" %
exportPluginInfo)
else:
exportPlugin = exportPluginInfo.plugin_object
logger.info("Exporting tracking result using %s" %
selectedPlugin)
name_format = self.dataExportApplet.topLevelOperator.getLane(
lane_index).OutputFilenameFormat.value
partially_formatted_name = self.getPartiallyFormattedName(
lane_index, name_format)
if exportPlugin.exportsToFile:
filename = partially_formatted_name
if os.path.basename(filename) == '':
filename = os.path.join(filename, 'pluginExport.txt')
else:
filename = os.path.dirname(partially_formatted_name)
if filename is None or len(str(filename)) == 0:
logger.error(
"Cannot export from plugin with empty output filename")
return
exportStatus = self.trackingApplet.topLevelOperator.getLane(
lane_index).exportPlugin(filename, exportPlugin,
checkOverwriteFiles)
if not exportStatus:
return False
logger.info("Export done")
return
# CSV Table export (only if plugin was not selected)
settings, selected_features = self.trackingApplet.topLevelOperator.getLane(
lane_index).get_table_export_settings()
if settings:
self.dataExportApplet.progressSignal.emit(0)
name_format = settings['file path']
partially_formatted_name = self.getPartiallyFormattedName(
lane_index, name_format)
settings['file path'] = partially_formatted_name
req = self.trackingApplet.topLevelOperator.getLane(
lane_index
).export_object_data(
lane_index,
# FIXME: Even in non-headless mode, we can't show the gui because we're running in a non-main thread.
# That's not a huge deal, because there's still a progress bar for the overall export.
show_gui=False)
req.wait()
self.dataExportApplet.progressSignal.emit(100)
# Restore option to bypass cache to false
self.objectExtractionApplet.topLevelOperator[
lane_index].BypassModeEnabled.setValue(False)
# Restore state of axis ranges
parameters = self.trackingApplet.topLevelOperator.Parameters.value
parameters['time_range'] = self.prev_time_range
parameters['x_range'] = self.prev_x_range
parameters['y_range'] = self.prev_y_range
parameters['z_range'] = self.prev_z_range
def getPartiallyFormattedName(self, lane_index, path_format_string):
''' Takes the format string for the output file, fills in the most important placeholders, and returns it '''
raw_dataset_info = self.dataSelectionApplet.topLevelOperator.DatasetGroup[
lane_index][0].value
project_path = self.shell.projectManager.currentProjectPath
project_dir = os.path.dirname(project_path)
dataset_dir = PathComponents(
raw_dataset_info.filePath).externalDirectory
abs_dataset_dir = make_absolute(dataset_dir, cwd=project_dir)
known_keys = {}
known_keys['dataset_dir'] = abs_dataset_dir
nickname = raw_dataset_info.nickname.replace('*', '')
if os.path.pathsep in nickname:
nickname = PathComponents(nickname.split(
os.path.pathsep)[0]).fileNameBase
known_keys['nickname'] = nickname
opDataExport = self.dataExportApplet.topLevelOperator.getLane(
lane_index)
known_keys[
'result_type'] = self.dataExportApplet.topLevelOperator.SelectedPlugin._value
# use partial formatting to fill in non-coordinate name fields
partially_formatted_name = format_known_keys(path_format_string,
known_keys)
return partially_formatted_name
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def onProjectLoaded(self, projectManager):
"""
Overridden from Workflow base class. Called by the Project Manager.
If the user provided command-line arguments, use them to configure
the workflow inputs and output settings.
"""
# Configure the data export operator.
if self._data_export_args:
self.dataExportApplet.configure_operator_with_parsed_args(
self._data_export_args)
# Configure headless mode.
if self._headless and self._batch_input_args and self._data_export_args:
logger.info("Beginning Batch Processing")
self.batchProcessingApplet.run_export_from_parsed_args(
self._batch_input_args)
logger.info("Completed Batch Processing")
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
opDataSelection = self.dataSelectionApplet.topLevelOperator
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and \
len(thresholdingOutput) > 0
else:
thresholding_ready = True and input_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.ComputedFeatureNamesAll
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
objectCountClassifier_ready = features_ready
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = objectCountClassifier_ready
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportApplet.busy
busy |= self.batchProcessingApplet.busy
self._shell.enableProjectChanges(not busy)
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet,
input_ready and not busy)
if self.divisionDetectionApplet:
self._shell.setAppletEnabled(self.divisionDetectionApplet,
features_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet,
thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet,
features_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet,
objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
self._shell.setAppletEnabled(self.batchProcessingApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
class ConservationTrackingWorkflowBase( Workflow ):
workflowName = "Automatic Tracking Workflow (Conservation Tracking) BASE"
def __init__( self, shell, headless, workflow_cmdline_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase, self).__init__(shell, headless, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
forceAxisOrder='txyzc',
instructionText=data_instructions,
max_lanes=1
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,
"Threshold and Size Filter",
"ThresholdTwoLevels" )
if self.withOptTrans:
self.opticalTranslationApplet = OpticalTranslationApplet(workflow=self)
self.objectExtractionApplet = TrackingFeatureExtractionApplet(workflow=self, interactive=False,
name="Object Feature Computation")
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection")
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification (optional)",
projectFileGroupName="CountClassification")
self.trackingApplet = ConservationTrackingApplet( workflow=self )
self.dataExportApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export")
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue( ['Tracking Result', 'Merger Result', 'Object Identities'] )
opDataExport.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
if self.withOptTrans:
self._applets.append(self.opticalTranslationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
if self.withOptTrans:
opOptTranslation = self.opticalTranslationApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(laneIndex)
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(opData.ImageGroup[0]) # Used for display only
# opTwoLevelThreshold.Channel.setValue(1)
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
if self.withOptTrans:
opOptTranslation.RawImage.connect(op5Raw.Output)
opOptTranslation.BinaryImage.connect(op5Binary.Output)
# # Connect operators ##
vigra_features = list((set(config.vigra_features)).union(config.selected_features_objectcount[config.features_vigra_name]))
feature_names_vigra = {}
feature_names_vigra[config.features_vigra_name] = { name: {} for name in vigra_features }
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
if self.withOptTrans:
opObjExtraction.TranslationVectors.connect(opOptTranslation.TranslationVectors)
opObjExtraction.FeatureNamesVigra.setValue(feature_names_vigra)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = { name: {} for name in config.division_features }
opObjExtraction.FeatureNamesDivision.setValue(feature_dict_division)
selected_features_div = {}
for plugin_name in config.selected_features_division.keys():
selected_features_div[plugin_name] = { name: {} for name in config.selected_features_division[plugin_name] }
# FIXME: do not hard code this
for name in [ 'SquaredDistances_' + str(i) for i in range(config.n_best_successors) ]:
selected_features_div[config.features_division_name][name] = {}
opDivDetection.BinaryImages.connect( op5Binary.Output )
opDivDetection.RawImages.connect( op5Raw.Output )
opDivDetection.LabelsAllowedFlags.connect(opData.AllowLabels)
opDivDetection.SegmentationImages.connect(opObjExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(opObjExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(opObjExtraction.ComputedFeatureNamesAll)
opDivDetection.SelectedFeatures.setValue(selected_features_div)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
selected_features_objectcount = {}
for plugin_name in config.selected_features_objectcount.keys():
selected_features_objectcount[plugin_name] = { name: {} for name in config.selected_features_objectcount[plugin_name] }
opCellClassification.BinaryImages.connect( op5Binary.Output )
opCellClassification.RawImages.connect( op5Raw.Output )
opCellClassification.LabelsAllowedFlags.connect(opData.AllowLabels)
opCellClassification.SegmentationImages.connect(opObjExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(opObjExtraction.RegionFeaturesVigra)
opCellClassification.ComputedFeatureNames.connect(opObjExtraction.ComputedFeatureNamesVigra)
opCellClassification.SelectedFeatures.setValue( selected_features_objectcount )
opCellClassification.SuggestedLabelNames.setValue( ['false detection',] + [str(i) + ' Objects' for i in range(1,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
opTracking.RawImage.connect( op5Raw.Output )
opTracking.LabelImage.connect( opObjExtraction.LabelImage )
opTracking.ObjectFeatures.connect( opObjExtraction.RegionFeaturesVigra )
opTracking.ObjectFeaturesWithDivFeatures.connect( opObjExtraction.RegionFeaturesAll)
opTracking.ComputedFeatureNames.connect( opObjExtraction.ComputedFeatureNamesVigra )
opTracking.ComputedFeatureNamesWithDivFeatures.connect( opObjExtraction.ComputedFeatureNamesAll )
opTracking.DivisionProbabilities.connect( opDivDetection.Probabilities )
opTracking.DetectionProbabilities.connect( opCellClassification.Probabilities )
opTracking.NumLabels.connect( opCellClassification.NumLabels )
opDataExport.Inputs.resize(3)
opDataExport.Inputs[0].connect( opTracking.Output )
opDataExport.Inputs[1].connect( opTracking.MergerOutput )
opDataExport.Inputs[2].connect( opTracking.LabelImage )
opDataExport.RawData.connect( op5Raw.Output )
opDataExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def post_process_lane_export(self, lane_index):
# FIXME: This probably only works for the non-blockwise export slot.
# We should assert that the user isn't using the blockwise slot.
settings, selected_features = self.trackingApplet.topLevelOperator.getLane(lane_index).get_table_export_settings()
if settings:
raw_dataset_info = self.dataSelectionApplet.topLevelOperator.DatasetGroup[lane_index][0].value
if raw_dataset_info.location == DatasetInfo.Location.FileSystem:
filename_suffix = raw_dataset_info.nickname
else:
filename_suffix = str(lane_index)
req = self.trackingApplet.topLevelOperator.getLane(lane_index).export_object_data(
lane_index,
# FIXME: Even in non-headless mode, we can't show the gui because we're running in a non-main thread.
# That's not a huge deal, because there's still a progress bar for the overall export.
show_gui=False,
filename_suffix=filename_suffix)
req.wait()
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
opDataSelection = self.dataSelectionApplet.topLevelOperator
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and \
len(thresholdingOutput) > 0
else:
thresholding_ready = True and input_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.ComputedFeatureNamesAll
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
objectCountClassifier_ready = features_ready
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = objectCountClassifier_ready and \
len(opTracking.EventsVector) > 0
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet, features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet, features_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet, objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
class ConservationTrackingWorkflowBase(Workflow):
workflowName = "Automatic Tracking Workflow (Conservation Tracking) BASE"
def __init__(self, shell, headless, workflow_cmdline_args,
project_creation_args, *args, **kwargs):
graph = kwargs["graph"] if "graph" in kwargs else Graph()
if "graph" in kwargs:
del kwargs["graph"]
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase,
self).__init__(shell,
headless,
workflow_cmdline_args,
project_creation_args,
graph=graph,
*args,
**kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Variables to store division and cell classifiers to prevent retraining every-time batch processing runs
self.stored_division_classifier = None
self.stored_cell_classifier = None
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
forceAxisOrder=["txyzc"],
instructionText=data_instructions,
max_lanes=None,
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue(
["Raw Data", "Segmentation Image"])
else:
opDataSelection.DatasetRoles.setValue(
["Raw Data", "Prediction Maps"])
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels")
self.objectExtractionApplet = TrackingFeatureExtractionApplet(
workflow=self,
interactive=False,
name="Object Feature Computation")
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
self.divisionDetectionApplet = self._createDivisionDetectionApplet(
configConservation.selectedFeaturesDiv) # Might be None
if self.divisionDetectionApplet:
feature_dict_division = {}
feature_dict_division[config.features_division_name] = {
name: {}
for name in config.division_features
}
opObjectExtraction.FeatureNamesDivision.setValue(
feature_dict_division)
selected_features_div = {}
for plugin_name in list(config.selected_features_division.keys()):
selected_features_div[plugin_name] = {
name: {}
for name in config.selected_features_division[plugin_name]
}
# FIXME: do not hard code this
for name in [
"SquaredDistances_" + str(i)
for i in range(config.n_best_successors)
]:
selected_features_div[config.features_division_name][name] = {}
opDivisionDetection = self.divisionDetectionApplet.topLevelOperator
opDivisionDetection.SelectedFeatures.setValue(
configConservation.selectedFeaturesDiv)
opDivisionDetection.LabelNames.setValue(
["Not Dividing", "Dividing"])
opDivisionDetection.AllowDeleteLabels.setValue(False)
opDivisionDetection.AllowAddLabel.setValue(False)
opDivisionDetection.EnableLabelTransfer.setValue(False)
self.cellClassificationApplet = ObjectClassificationApplet(
workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount,
)
selected_features_objectcount = {}
for plugin_name in list(config.selected_features_objectcount.keys()):
selected_features_objectcount[plugin_name] = {
name: {}
for name in config.selected_features_objectcount[plugin_name]
}
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(
configConservation.selectedFeaturesObjectCount)
opCellClassification.SuggestedLabelNames.setValue(
["False Detection"] + [str(1) + " Object"] +
[str(i) + " Objects" for i in range(2, 10)])
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
self.trackingApplet = ConservationTrackingApplet(workflow=self)
self.default_export_filename = "{dataset_dir}/{nickname}-exported_data.csv"
self.dataExportApplet = TrackingBaseDataExportApplet(
self,
"Tracking Result Export",
default_export_filename=self.default_export_filename,
pluginExportFunc=self._pluginExportFunc,
)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(
["Object-Identities", "Tracking-Result", "Merger-Result"])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.prepare_lane_for_export = self.prepare_lane_for_export
# configure export settings
# settings = {'file path': self.default_export_filename, 'compression': {}, 'file type': 'csv'}
# selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
# opTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
if self.divisionDetectionApplet:
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(
self, "Batch Processing", self.dataSelectionApplet,
self.dataExportApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
self._applets.append(self.batchProcessingApplet)
# Parse export and batch command-line arguments for headless mode
if workflow_cmdline_args:
self._data_export_args, unused_args = self.dataExportApplet.parse_known_cmdline_args(
workflow_cmdline_args)
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args(
workflow_cmdline_args)
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
if unused_args:
logger.warning("Unused command-line args: {}".format(unused_args))
@property
def applets(self):
return self._applets
def _createDivisionDetectionApplet(self, selectedFeatures=dict()):
return ObjectClassificationApplet(
workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=selectedFeatures,
)
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def prepareForNewLane(self, laneIndex):
# Store division and cell classifiers
if self.divisionDetectionApplet:
opDivisionClassification = self.divisionDetectionApplet.topLevelOperator
if (opDivisionClassification.classifier_cache.Output.ready()
and not opDivisionClassification.classifier_cache._dirty):
self.stored_division_classifier = opDivisionClassification.classifier_cache.Output.value
else:
self.stored_division_classifier = None
opCellClassification = self.cellClassificationApplet.topLevelOperator
if opCellClassification.classifier_cache.Output.ready(
) and not opCellClassification.classifier_cache._dirty:
self.stored_cell_classifier = opCellClassification.classifier_cache.Output.value
else:
self.stored_cell_classifier = None
def handleNewLanesAdded(self):
"""
If new lanes were added, then we invalidated our classifiers unecessarily.
Here, we can restore the classifier so it doesn't need to be retrained.
"""
# If we have stored division and cell classifiers, restore them into the workflow now.
if self.stored_division_classifier:
opDivisionClassification = self.divisionDetectionApplet.topLevelOperator
opDivisionClassification.classifier_cache.forceValue(
self.stored_division_classifier)
# Release reference
self.stored_division_classifier = None
# If we have stored division and cell classifiers, restore them into the workflow now.
if self.stored_cell_classifier:
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.classifier_cache.forceValue(
self.stored_cell_classifier)
# Release reference
self.stored_cell_classifier = None
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(
laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(
laneIndex)
opObjExtraction.setDefaultFeatures(
configConservation.allFeaturesObjectCount)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(
laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(
laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(
laneIndex)
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(
opData.ImageGroup[0]) # Used for display only
# opTwoLevelThreshold.Channel.setValue(1)
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
# # Connect operators ##
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect(op5Binary.Output)
opDivDetection.RawImages.connect(op5Raw.Output)
opDivDetection.SegmentationImages.connect(
opObjExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(
opObjExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(
opObjExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect(op5Binary.Output)
opCellClassification.RawImages.connect(op5Raw.Output)
opCellClassification.SegmentationImages.connect(
opObjExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(
opObjExtraction.RegionFeaturesVigra)
opCellClassification.ComputedFeatureNames.connect(
opObjExtraction.FeatureNamesVigra)
if self.divisionDetectionApplet:
opTracking.ObjectFeaturesWithDivFeatures.connect(
opObjExtraction.RegionFeaturesAll)
opTracking.ComputedFeatureNamesWithDivFeatures.connect(
opObjExtraction.ComputedFeatureNamesAll)
opTracking.DivisionProbabilities.connect(
opDivDetection.Probabilities)
opTracking.RawImage.connect(op5Raw.Output)
opTracking.LabelImage.connect(opObjExtraction.LabelImage)
opTracking.ObjectFeatures.connect(opObjExtraction.RegionFeaturesVigra)
opTracking.ComputedFeatureNames.connect(
opObjExtraction.FeatureNamesVigra)
opTracking.DetectionProbabilities.connect(
opCellClassification.Probabilities)
opTracking.NumLabels.connect(opCellClassification.NumLabels)
opDataExport.Inputs.resize(3)
opDataExport.Inputs[0].connect(opTracking.RelabeledImage)
opDataExport.Inputs[1].connect(opTracking.Output)
opDataExport.Inputs[2].connect(opTracking.MergerOutput)
opDataExport.RawData.connect(op5Raw.Output)
opDataExport.RawDatasetInfo.connect(opData.DatasetGroup[0])
def prepare_lane_for_export(self, lane_index):
# Bypass cache on headless mode and batch processing mode
self.objectExtractionApplet.topLevelOperator[
lane_index].BypassModeEnabled.setValue(True)
if not self.fromBinary:
self.thresholdTwoLevelsApplet.topLevelOperator[
lane_index].opCache.BypassModeEnabled.setValue(True)
self.thresholdTwoLevelsApplet.topLevelOperator[
lane_index].opSmootherCache.BypassModeEnabled.setValue(True)
# Get axes info
maxt = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[0]
maxx = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[1]
maxy = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[2]
maxz = self.trackingApplet.topLevelOperator[
lane_index].RawImage.meta.shape[3]
time_enum = list(range(maxt))
x_range = (0, maxx)
y_range = (0, maxy)
z_range = (0, maxz)
ndim = 2
if (z_range[1] - z_range[0]) > 1:
ndim = 3
parameters = self.trackingApplet.topLevelOperator.Parameters.value
# Save state of axis ranges
if "time_range" in parameters:
self.prev_time_range = parameters["time_range"]
else:
self.prev_time_range = time_enum
if "x_range" in parameters:
self.prev_x_range = parameters["x_range"]
else:
self.prev_x_range = x_range
if "y_range" in parameters:
self.prev_y_range = parameters["y_range"]
else:
self.prev_y_range = y_range
if "z_range" in parameters:
self.prev_z_range = parameters["z_range"]
else:
self.prev_z_range = z_range
if "numFramesPerSplit" in parameters:
numFramesPerSplit = parameters["numFramesPerSplit"]
else:
numFramesPerSplit = 0
self.trackingApplet.topLevelOperator[lane_index].track(
time_range=time_enum,
x_range=x_range,
y_range=y_range,
z_range=z_range,
size_range=parameters["size_range"],
x_scale=parameters["scales"][0],
y_scale=parameters["scales"][1],
z_scale=parameters["scales"][2],
maxDist=parameters["maxDist"],
maxObj=parameters["maxObj"],
divThreshold=parameters["divThreshold"],
avgSize=parameters["avgSize"],
withTracklets=parameters["withTracklets"],
sizeDependent=parameters["sizeDependent"],
divWeight=parameters["divWeight"],
transWeight=parameters["transWeight"],
withDivisions=parameters["withDivisions"],
withOpticalCorrection=parameters["withOpticalCorrection"],
withClassifierPrior=parameters["withClassifierPrior"],
ndim=ndim,
withMergerResolution=parameters["withMergerResolution"],
borderAwareWidth=parameters["borderAwareWidth"],
withArmaCoordinates=parameters["withArmaCoordinates"],
cplex_timeout=parameters["cplex_timeout"],
appearance_cost=parameters["appearanceCost"],
disappearance_cost=parameters["disappearanceCost"],
max_nearest_neighbors=parameters["max_nearest_neighbors"],
numFramesPerSplit=numFramesPerSplit,
force_build_hypotheses_graph=False,
withBatchProcessing=True,
)
def _pluginExportFunc(self, lane_index, filename, exportPlugin,
checkOverwriteFiles, plugArgsSlot) -> int:
return self.trackingApplet.topLevelOperator.getLane(
lane_index).exportPlugin(filename, exportPlugin,
checkOverwriteFiles, plugArgsSlot)
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and all(
[sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def onProjectLoaded(self, projectManager):
"""
Overridden from Workflow base class. Called by the Project Manager.
If the user provided command-line arguments, use them to configure
the workflow inputs and output settings.
"""
# Configure the data export operator.
if self._data_export_args:
self.dataExportApplet.configure_operator_with_parsed_args(
self._data_export_args)
# Configure headless mode.
if self._headless and self._batch_input_args and self._data_export_args:
logger.info("Beginning Batch Processing")
self.batchProcessingApplet.run_export_from_parsed_args(
self._batch_input_args)
logger.info("Completed Batch Processing")
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
opDataSelection = self.dataSelectionApplet.topLevelOperator
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and len(thresholdingOutput) > 0
else:
thresholding_ready = True and input_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.ComputedFeatureNamesAll
features_ready = thresholding_ready and len(objectExtractionOutput) > 0
objectCountClassifier_ready = features_ready
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = objectCountClassifier_ready
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportApplet.busy
busy |= self.batchProcessingApplet.busy
self._shell.enableProjectChanges(not busy)
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet,
input_ready and not busy)
if self.divisionDetectionApplet:
self._shell.setAppletEnabled(self.divisionDetectionApplet,
features_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet,
thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet,
features_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet,
objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(
self.dataExportApplet,
tracking_ready and not busy
and self.dataExportApplet.topLevelOperator.Inputs[0][0].ready(),
)
self._shell.setAppletEnabled(
self.batchProcessingApplet,
tracking_ready and not busy
and self.dataExportApplet.topLevelOperator.Inputs[0][0].ready(),
)
class ConservationTrackingWorkflowBase(Workflow):
workflowName = "Automatic Tracking Workflow (Conservation Tracking) BASE"
def __init__(self, shell, headless, workflow_cmdline_args, *args, **kwargs):
graph = kwargs["graph"] if "graph" in kwargs else Graph()
if "graph" in kwargs:
del kwargs["graph"]
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase, self).__init__(shell, headless, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
batchDataGui=False,
force5d=True,
instructionText=data_instructions,
max_lanes=1,
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue(["Raw Data", "Binary Image"])
else:
opDataSelection.DatasetRoles.setValue(["Raw Data", "Prediction Maps"])
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels"
)
if self.withOptTrans:
self.opticalTranslationApplet = OpticalTranslationApplet(workflow=self)
self.objectExtractionApplet = TrackingFeatureExtractionApplet(
workflow=self, interactive=False, name="Object Feature Computation"
)
self.divisionDetectionApplet = ObjectClassificationApplet(
workflow=self, name="Division Detection (optional)", projectFileGroupName="DivisionDetection"
)
self.cellClassificationApplet = ObjectClassificationApplet(
workflow=self, name="Object Count Classification (optional)", projectFileGroupName="CountClassification"
)
self.trackingApplet = ConservationTrackingApplet(workflow=self)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export")
self.dataExportApplet.set_exporting_operator(opTracking)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(["Tracking Result", "Merger Result", "Object Identities"])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
if self.withOptTrans:
self._applets.append(self.opticalTranslationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
if self.withOptTrans:
opOptTranslation = self.opticalTranslationApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(laneIndex)
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect(opData.ImageGroup[1])
opTwoLevelThreshold.RawInput.connect(opData.ImageGroup[0]) # Used for display only
# opTwoLevelThreshold.Channel.setValue(1)
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
if self.withOptTrans:
opOptTranslation.RawImage.connect(op5Raw.Output)
opOptTranslation.BinaryImage.connect(op5Binary.Output)
# # Connect operators ##
vigra_features = list(
(set(config.vigra_features)).union(config.selected_features_objectcount[config.features_vigra_name])
)
feature_names_vigra = {}
feature_names_vigra[config.features_vigra_name] = {name: {} for name in vigra_features}
opObjExtraction.RawImage.connect(op5Raw.Output)
opObjExtraction.BinaryImage.connect(op5Binary.Output)
if self.withOptTrans:
opObjExtraction.TranslationVectors.connect(opOptTranslation.TranslationVectors)
opObjExtraction.FeatureNamesVigra.setValue(feature_names_vigra)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = {name: {} for name in config.division_features}
opObjExtraction.FeatureNamesDivision.setValue(feature_dict_division)
selected_features_div = {}
for plugin_name in config.selected_features_division.keys():
selected_features_div[plugin_name] = {name: {} for name in config.selected_features_division[plugin_name]}
# FIXME: do not hard code this
for name in ["SquaredDistances_" + str(i) for i in range(config.n_best_successors)]:
selected_features_div[config.features_division_name][name] = {}
opDivDetection.BinaryImages.connect(op5Binary.Output)
opDivDetection.RawImages.connect(op5Raw.Output)
opDivDetection.LabelsAllowedFlags.connect(opData.AllowLabels)
opDivDetection.SegmentationImages.connect(opObjExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(opObjExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(opObjExtraction.ComputedFeatureNamesAll)
opDivDetection.SelectedFeatures.setValue(selected_features_div)
opDivDetection.LabelNames.setValue(["Not Dividing", "Dividing"])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
selected_features_objectcount = {}
for plugin_name in config.selected_features_objectcount.keys():
selected_features_objectcount[plugin_name] = {
name: {} for name in config.selected_features_objectcount[plugin_name]
}
opCellClassification.BinaryImages.connect(op5Binary.Output)
opCellClassification.RawImages.connect(op5Raw.Output)
opCellClassification.LabelsAllowedFlags.connect(opData.AllowLabels)
opCellClassification.SegmentationImages.connect(opObjExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(opObjExtraction.RegionFeaturesVigra)
opCellClassification.ComputedFeatureNames.connect(opObjExtraction.ComputedFeatureNamesVigra)
opCellClassification.SelectedFeatures.setValue(selected_features_objectcount)
opCellClassification.SuggestedLabelNames.setValue(
["false detection"] + [str(i) + " Objects" for i in range(1, 10)]
)
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
opTracking.RawImage.connect(op5Raw.Output)
opTracking.LabelImage.connect(opObjExtraction.LabelImage)
opTracking.ObjectFeatures.connect(opObjExtraction.RegionFeaturesVigra)
opTracking.ObjectFeaturesWithDivFeatures.connect(opObjExtraction.RegionFeaturesAll)
opTracking.ComputedFeatureNames.connect(opObjExtraction.ComputedFeatureNamesVigra)
opTracking.ComputedFeatureNamesWithDivFeatures.connect(opObjExtraction.ComputedFeatureNamesAll)
opTracking.DivisionProbabilities.connect(opDivDetection.Probabilities)
opTracking.DetectionProbabilities.connect(opCellClassification.Probabilities)
opTracking.NumLabels.connect(opCellClassification.NumLabels)
opDataExport.Inputs.resize(3)
opDataExport.Inputs[0].connect(opTracking.Output)
opDataExport.Inputs[1].connect(opTracking.MergerOutput)
opDataExport.Inputs[2].connect(opTracking.LabelImage)
opDataExport.RawData.connect(op5Raw.Output)
opDataExport.RawDatasetInfo.connect(opData.DatasetGroup[0])
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
opDataSelection = self.dataSelectionApplet.topLevelOperator
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and len(thresholdingOutput) > 0
else:
thresholding_ready = True and input_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.ComputedFeatureNamesAll
features_ready = thresholding_ready and len(objectExtractionOutput) > 0
objectCountClassifier_ready = features_ready
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = objectCountClassifier_ready and len(opTracking.EventsVector) > 0
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportApplet.busy
self._shell.enableProjectChanges(not busy)
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet, features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet, features_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet, objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(
self.dataExportApplet,
tracking_ready and not busy and self.dataExportApplet.topLevelOperator.Inputs[0][0].ready(),
)
class ManualTrackingWorkflow( Workflow ):
workflowName = "Manual Tracking Workflow"
workflowDisplayName = "Manual Tracking Workflow [Inputs: Raw Data, Pixel Prediction Map]"
workflowDescription = "Manual tracking of objects, based on Prediction Maps or (binary) Segmentation Images"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(ManualTrackingWorkflow, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'\
'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
forceAxisOrder='txyzc',
instructionText=data_instructions,
max_lanes=1
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,
"Threshold and Size Filter",
"ThresholdTwoLevels" )
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",
workflow=self, interactive=False)
self.trackingApplet = ManualTrackingApplet( workflow=self )
self.default_export_filename = '{dataset_dir}/{nickname}-exported_data.csv'
self.dataExportApplet = TrackingBaseDataExportApplet(self,
"Tracking Result Export",
default_export_filename=self.default_export_filename)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue( ['Manual Tracking', 'Object Identities'] )
opDataExport.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# Extra configuration for object export table (as CSV table or HDF5 table)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet.set_exporting_operator(opTracking)
self.dataExportApplet.post_process_lane_export = self.post_process_lane_export
self._applets = []
self._applets.append(self.dataSelectionApplet)
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
opDataExport = self.dataExportApplet.topLevelOperator.getLane(laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opTwoLevelThreshold.InputImage.connect( opData.ImageGroup[1] )
opTwoLevelThreshold.RawInput.connect( opData.ImageGroup[0] ) # Used for display only
# Use OpReorderAxis for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes( parent=self )
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect( opTwoLevelThreshold.CachedOutput )
opObjExtraction.RawImage.connect( op5Raw.Output )
opObjExtraction.BinaryImage.connect( op5Binary.Output )
opTracking.RawImage.connect( op5Raw.Output )
opTracking.BinaryImage.connect( op5Binary.Output )
opTracking.LabelImage.connect( opObjExtraction.LabelImage )
opTracking.ObjectFeatures.connect( opObjExtraction.RegionFeatures )
opTracking.ComputedFeatureNames.connect(opObjExtraction.Features)
opDataExport.Inputs.resize(2)
opDataExport.Inputs[0].connect( opTracking.TrackImage )
opDataExport.Inputs[1].connect( opTracking.LabelImage )
opDataExport.RawData.connect( op5Raw.Output )
opDataExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def post_process_lane_export(self, lane_index):
# FIXME: This probably only works for the non-blockwise export slot.
# We should assert that the user isn't using the blockwise slot.
settings, selected_features = self.trackingApplet.topLevelOperator.getLane(lane_index).get_table_export_settings()
if settings:
raw_dataset_info = self.dataSelectionApplet.topLevelOperator.DatasetGroup[lane_index][0].value
if raw_dataset_info.location == DatasetInfo.Location.FileSystem:
filename_suffix = raw_dataset_info.nickname
else:
filename_suffix = str(lane_index)
req = self.trackingApplet.topLevelOperator.getLane(lane_index).export_object_data(
lane_index,
# FIXME: Even in non-headless mode, we can't show the gui because we're running in a non-main thread.
# That's not a huge deal, because there's still a progress bar for the overall export.
show_gui=False,
filename_suffix=filename_suffix)
req.wait()
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and \
len(thresholdingOutput) > 0
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
features_ready = thresholding_ready
opTracking = self.trackingApplet.topLevelOperator
tracking_ready = features_ready and \
len(opTracking.Labels) > 0 and \
opTracking.Labels.ready() and \
opTracking.TrackImage.ready()
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.dataExportApplet.busy
busy |= self.trackingApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.trackingApplet, features_ready and not busy)
self._shell.setAppletEnabled(self.dataExportApplet, tracking_ready and not busy and \
self.dataExportApplet.topLevelOperator.Inputs[0][0].ready() )
def __init__(self, shell, headless, workflow_cmdline_args, *args, **kwargs):
graph = kwargs["graph"] if "graph" in kwargs else Graph()
if "graph" in kwargs:
del kwargs["graph"]
# if 'withOptTrans' in kwargs:
# self.withOptTrans = kwargs['withOptTrans']
# if 'fromBinary' in kwargs:
# self.fromBinary = kwargs['fromBinary']
super(ConservationTrackingWorkflowBase, self).__init__(shell, headless, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
## Create applets
self.dataSelectionApplet = DataSelectionApplet(
self,
"Input Data",
"Input Data",
batchDataGui=False,
force5d=True,
instructionText=data_instructions,
max_lanes=1,
)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue(["Raw Data", "Binary Image"])
else:
opDataSelection.DatasetRoles.setValue(["Raw Data", "Prediction Maps"])
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet(
self, "Threshold and Size Filter", "ThresholdTwoLevels"
)
if self.withOptTrans:
self.opticalTranslationApplet = OpticalTranslationApplet(workflow=self)
self.objectExtractionApplet = TrackingFeatureExtractionApplet(
workflow=self, interactive=False, name="Object Feature Computation"
)
self.divisionDetectionApplet = ObjectClassificationApplet(
workflow=self, name="Division Detection (optional)", projectFileGroupName="DivisionDetection"
)
self.cellClassificationApplet = ObjectClassificationApplet(
workflow=self, name="Object Count Classification (optional)", projectFileGroupName="CountClassification"
)
self.trackingApplet = ConservationTrackingApplet(workflow=self)
opTracking = self.trackingApplet.topLevelOperator
self.dataExportApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export")
self.dataExportApplet.set_exporting_operator(opTracking)
opDataExport = self.dataExportApplet.topLevelOperator
opDataExport.SelectionNames.setValue(["Tracking Result", "Merger Result", "Object Identities"])
opDataExport.WorkingDirectory.connect(opDataSelection.WorkingDirectory)
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
if self.withOptTrans:
self._applets.append(self.opticalTranslationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportApplet)
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configConservation.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configConservation.selectedFeaturesObjectCount)
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
if SOLVER=="CPLEX" or SOLVER=="GUROBI":
self._solver="ILP"
elif SOLVER=="DPCT":
self._solver="Flow-based"
else:
self._solver=None
opStructuredTracking._solver = self._solver
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export",default_export_filename=self.default_tracking_export_filename)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking-Result', 'Merger-Result', 'Object-Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.prepare_lane_for_export = self.prepare_lane_for_export
self.dataExportTrackingApplet.post_process_lane_export = self.post_process_lane_export
# configure export settings
settings = {'file path': self.default_tracking_export_filename, 'compression': {}, 'file type': 'h5'}
selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
opStructuredTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(self, "Batch Processing", self.dataSelectionApplet, self.dataExportTrackingApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
if workflow_cmdline_args:
if '--testFullAnnotations' in workflow_cmdline_args:
self.testFullAnnotations = True
else:
self.testFullAnnotations = False
self._data_export_args, unused_args = self.dataExportTrackingApplet.parse_known_cmdline_args( workflow_cmdline_args )
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args( workflow_cmdline_args )
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
self.testFullAnnotations = False
if unused_args:
logger.warning("Unused command-line args: {}".format( unused_args ))
class StructuredTrackingWorkflowBase( Workflow ):
workflowName = "Structured Learning Tracking Workflow BASE"
@property
def applets(self):
return self._applets
@property
def imageNameListSlot(self):
return self.dataSelectionApplet.topLevelOperator.ImageName
def __init__( self, shell, headless, workflow_cmdline_args, project_creation_args, *args, **kwargs ):
graph = kwargs['graph'] if 'graph' in kwargs else Graph()
if 'graph' in kwargs: del kwargs['graph']
super(StructuredTrackingWorkflowBase, self).__init__(shell, headless, workflow_cmdline_args, project_creation_args, graph=graph, *args, **kwargs)
data_instructions = 'Use the "Raw Data" tab to load your intensity image(s).\n\n'
if self.fromBinary:
data_instructions += 'Use the "Binary Image" tab to load your segmentation image(s).'
else:
data_instructions += 'Use the "Prediction Maps" tab to load your pixel-wise probability image(s).'
# Create applets
self.dataSelectionApplet = DataSelectionApplet(self,
"Input Data",
"Input Data",
batchDataGui=False,
forceAxisOrder=['txyzc'],
instructionText=data_instructions,
max_lanes=1)
opDataSelection = self.dataSelectionApplet.topLevelOperator
if self.fromBinary:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Binary Image'] )
else:
opDataSelection.DatasetRoles.setValue( ['Raw Data', 'Prediction Maps'] )
if not self.fromBinary:
self.thresholdTwoLevelsApplet = ThresholdTwoLevelsApplet( self,"Threshold and Size Filter","ThresholdTwoLevels" )
self.divisionDetectionApplet = ObjectClassificationApplet(workflow=self,
name="Division Detection (optional)",
projectFileGroupName="DivisionDetection",
selectedFeatures=configStructured.selectedFeaturesDiv)
self.cellClassificationApplet = ObjectClassificationApplet(workflow=self,
name="Object Count Classification",
projectFileGroupName="CountClassification",
selectedFeatures=configStructured.selectedFeaturesObjectCount)
self.cropSelectionApplet = CropSelectionApplet(self,"Crop Selection","CropSelection")
self.trackingFeatureExtractionApplet = TrackingFeatureExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.objectExtractionApplet = ObjectExtractionApplet(name="Object Feature Computation",workflow=self, interactive=False)
self.annotationsApplet = AnnotationsApplet( name="Training", workflow=self )
opAnnotations = self.annotationsApplet.topLevelOperator
# self.default_training_export_filename = '{dataset_dir}/{nickname}-training_exported_data.csv'
# self.dataExportAnnotationsApplet = TrackingBaseDataExportApplet(self, "Training Export",default_export_filename=self.default_training_export_filename)
# opDataExportAnnotations = self.dataExportAnnotationsApplet.topLevelOperator
# opDataExportAnnotations.SelectionNames.setValue( ['User Training for Tracking', 'Object Identities'] )
# opDataExportAnnotations.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
# self.dataExportAnnotationsApplet.set_exporting_operator(opAnnotations)
self.trackingApplet = StructuredTrackingApplet( name="Tracking - Structured Learning", workflow=self )
opStructuredTracking = self.trackingApplet.topLevelOperator
self.default_tracking_export_filename = '{dataset_dir}/{nickname}-tracking_exported_data.csv'
self.dataExportTrackingApplet = TrackingBaseDataExportApplet(self, "Tracking Result Export",default_export_filename=self.default_tracking_export_filename)
opDataExportTracking = self.dataExportTrackingApplet.topLevelOperator
opDataExportTracking.SelectionNames.setValue( ['Tracking-Result', 'Merger-Result', 'Object-Identities'] )
opDataExportTracking.WorkingDirectory.connect( opDataSelection.WorkingDirectory )
self.dataExportTrackingApplet.set_exporting_operator(opStructuredTracking)
self.dataExportTrackingApplet.prepare_lane_for_export = self.prepare_lane_for_export
self.dataExportTrackingApplet.post_process_lane_export = self.post_process_lane_export
# configure export settings
settings = {'file path': self.default_tracking_export_filename, 'compression': {}, 'file type': 'h5'}
selected_features = ['Count', 'RegionCenter', 'RegionRadii', 'RegionAxes']
opStructuredTracking.ExportSettings.setValue( (settings, selected_features) )
self._applets = []
self._applets.append(self.dataSelectionApplet)
if not self.fromBinary:
self._applets.append(self.thresholdTwoLevelsApplet)
self._applets.append(self.trackingFeatureExtractionApplet)
self._applets.append(self.divisionDetectionApplet)
self.batchProcessingApplet = BatchProcessingApplet(self, "Batch Processing", self.dataSelectionApplet, self.dataExportTrackingApplet)
self._applets.append(self.cellClassificationApplet)
self._applets.append(self.cropSelectionApplet)
self._applets.append(self.objectExtractionApplet)
self._applets.append(self.annotationsApplet)
# self._applets.append(self.dataExportAnnotationsApplet)
self._applets.append(self.trackingApplet)
self._applets.append(self.dataExportTrackingApplet)
if self.divisionDetectionApplet:
opDivDetection = self.divisionDetectionApplet.topLevelOperator
opDivDetection.SelectedFeatures.setValue(configConservation.selectedFeaturesDiv)
opDivDetection.LabelNames.setValue(['Not Dividing', 'Dividing'])
opDivDetection.AllowDeleteLabels.setValue(False)
opDivDetection.AllowAddLabel.setValue(False)
opDivDetection.EnableLabelTransfer.setValue(False)
opCellClassification = self.cellClassificationApplet.topLevelOperator
opCellClassification.SelectedFeatures.setValue(configConservation.selectedFeaturesObjectCount )
opCellClassification.SuggestedLabelNames.setValue( ['False Detection',] + [str(1) + ' Object'] + [str(i) + ' Objects' for i in range(2,10) ] )
opCellClassification.AllowDeleteLastLabelOnly.setValue(True)
opCellClassification.EnableLabelTransfer.setValue(False)
if workflow_cmdline_args:
self._data_export_args, unused_args = self.dataExportTrackingApplet.parse_known_cmdline_args( workflow_cmdline_args )
self._batch_input_args, unused_args = self.batchProcessingApplet.parse_known_cmdline_args( workflow_cmdline_args )
else:
unused_args = None
self._data_export_args = None
self._batch_input_args = None
if unused_args:
logger.warn("Unused command-line args: {}".format( unused_args ))
def connectLane(self, laneIndex):
opData = self.dataSelectionApplet.topLevelOperator.getLane(laneIndex)
opObjExtraction = self.objectExtractionApplet.topLevelOperator.getLane(laneIndex)
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator.getLane(laneIndex)
opAnnotations = self.annotationsApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold = self.thresholdTwoLevelsApplet.topLevelOperator.getLane(laneIndex)
# opDataAnnotationsExport = self.dataExportAnnotationsApplet.topLevelOperator.getLane(laneIndex)
opCropSelection = self.cropSelectionApplet.topLevelOperator.getLane(laneIndex)
opStructuredTracking = self.trackingApplet.topLevelOperator.getLane(laneIndex)
opDataTrackingExport = self.dataExportTrackingApplet.topLevelOperator.getLane(laneIndex)
## Connect operators ##
op5Raw = OpReorderAxes(parent=self)
op5Raw.AxisOrder.setValue("txyzc")
op5Raw.Input.connect(opData.ImageGroup[0])
opDivDetection = self.divisionDetectionApplet.topLevelOperator.getLane(laneIndex)
opCellClassification = self.cellClassificationApplet.topLevelOperator.getLane(laneIndex)
if not self.fromBinary:
opTwoLevelThreshold.InputImage.connect( opData.ImageGroup[1] )
opTwoLevelThreshold.RawInput.connect( opData.ImageGroup[0] ) # Used for display only
binarySrc = opTwoLevelThreshold.CachedOutput
else:
binarySrc = opData.ImageGroup[1]
# Use Op5ifyers for both input datasets such that they are guaranteed to
# have the same axis order after thresholding
op5Binary = OpReorderAxes(parent=self)
op5Binary.AxisOrder.setValue("txyzc")
op5Binary.Input.connect(binarySrc)
opCropSelection.InputImage.connect( opData.ImageGroup[0] )
opCropSelection.PredictionImage.connect( opData.ImageGroup[1] )
opObjExtraction.RawImage.connect( op5Raw.Output )
opObjExtraction.BinaryImage.connect( op5Binary.Output )
opTrackingFeatureExtraction.RawImage.connect( op5Raw.Output )
opTrackingFeatureExtraction.BinaryImage.connect( op5Binary.Output )
# vigra_features = list((set(config.vigra_features)).union(config.selected_features_objectcount[config.features_vigra_name]))
# feature_names_vigra = {}
# feature_names_vigra[config.features_vigra_name] = { name: {} for name in vigra_features }
opTrackingFeatureExtraction.FeatureNamesVigra.setValue(configConservation.allFeaturesObjectCount)
feature_dict_division = {}
feature_dict_division[config.features_division_name] = { name: {} for name in config.division_features }
opTrackingFeatureExtraction.FeatureNamesDivision.setValue(feature_dict_division)
if self.divisionDetectionApplet:
opDivDetection.BinaryImages.connect( op5Binary.Output )
opDivDetection.RawImages.connect( op5Raw.Output )
opDivDetection.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opDivDetection.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opDivDetection.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesAll)
opCellClassification.BinaryImages.connect( op5Binary.Output )
opCellClassification.RawImages.connect( op5Raw.Output )
opCellClassification.SegmentationImages.connect(opTrackingFeatureExtraction.LabelImage)
opCellClassification.ObjectFeatures.connect(opTrackingFeatureExtraction.RegionFeaturesAll)
opCellClassification.ComputedFeatureNames.connect(opTrackingFeatureExtraction.ComputedFeatureNamesNoDivisions)
opAnnotations.RawImage.connect( op5Raw.Output )
opAnnotations.BinaryImage.connect( op5Binary.Output )
opAnnotations.LabelImage.connect( opObjExtraction.LabelImage )
opAnnotations.ObjectFeatures.connect( opObjExtraction.RegionFeatures )
opAnnotations.ComputedFeatureNames.connect(opObjExtraction.Features)
opAnnotations.Crops.connect( opCropSelection.Crops)
opAnnotations.DivisionProbabilities.connect( opDivDetection.Probabilities )
opAnnotations.DetectionProbabilities.connect( opCellClassification.Probabilities )
opAnnotations.MaxNumObj.connect (opCellClassification.MaxNumObj)
# opDataAnnotationsExport.Inputs.resize(2)
# opDataAnnotationsExport.Inputs[0].connect( opAnnotations.TrackImage )
# opDataAnnotationsExport.Inputs[1].connect( opAnnotations.LabelImage )
# opDataAnnotationsExport.RawData.connect( op5Raw.Output )
# opDataAnnotationsExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
opStructuredTracking.RawImage.connect( op5Raw.Output )
opStructuredTracking.LabelImage.connect( opTrackingFeatureExtraction.LabelImage )
opStructuredTracking.ObjectFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesVigra )
opStructuredTracking.ComputedFeatureNames.connect( opTrackingFeatureExtraction.FeatureNamesVigra )
if self.divisionDetectionApplet:
opStructuredTracking.ObjectFeaturesWithDivFeatures.connect( opTrackingFeatureExtraction.RegionFeaturesAll)
opStructuredTracking.ComputedFeatureNamesWithDivFeatures.connect( opTrackingFeatureExtraction.ComputedFeatureNamesAll )
opStructuredTracking.DivisionProbabilities.connect( opDivDetection.Probabilities )
# configure tracking export settings
settings = {'file path': self.default_tracking_export_filename, 'compression': {}, 'file type': 'csv'}
selected_features = ['Count', 'RegionCenter']
opStructuredTracking.configure_table_export_settings(settings, selected_features)
opStructuredTracking.DetectionProbabilities.connect( opCellClassification.Probabilities )
opStructuredTracking.NumLabels.connect( opCellClassification.NumLabels )
opStructuredTracking.Crops.connect (opCropSelection.Crops)
opStructuredTracking.Annotations.connect (opAnnotations.Annotations)
opStructuredTracking.Labels.connect (opAnnotations.Labels)
opStructuredTracking.Divisions.connect (opAnnotations.Divisions)
opStructuredTracking.MaxNumObj.connect (opCellClassification.MaxNumObj)
opDataTrackingExport.Inputs.resize(3)
opDataTrackingExport.Inputs[0].connect( opStructuredTracking.RelabeledImage )
opDataTrackingExport.Inputs[1].connect( opStructuredTracking.MergerOutput )
opDataTrackingExport.Inputs[2].connect( opStructuredTracking.LabelImage )
opDataTrackingExport.RawData.connect( op5Raw.Output )
opDataTrackingExport.RawDatasetInfo.connect( opData.DatasetGroup[0] )
def prepare_lane_for_export(self, lane_index):
import logging
logger = logging.getLogger(__name__)
maxt = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[0]
maxx = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[1]
maxy = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[2]
maxz = self.trackingApplet.topLevelOperator[lane_index].RawImage.meta.shape[3]
time_enum = range(maxt)
x_range = (0, maxx)
y_range = (0, maxy)
z_range = (0, maxz)
ndim = 2
if ( z_range[1] - z_range[0] ) > 1:
ndim = 3
parameters = self.trackingApplet.topLevelOperator.Parameters.value
# Save state of axis ranges
if 'time_range' in parameters:
self.prev_time_range = parameters['time_range']
else:
self.prev_time_range = time_enum
if 'x_range' in parameters:
self.prev_x_range = parameters['x_range']
else:
self.prev_x_range = x_range
if 'y_range' in parameters:
self.prev_y_range = parameters['y_range']
else:
self.prev_y_range = y_range
if 'z_range' in parameters:
self.prev_z_range = parameters['z_range']
else:
self.prev_z_range = z_range
# batch processing starts a new lane, so training data needs to be copied from the lane that loaded the project
loaded_project_lane_index=0
self.annotationsApplet.topLevelOperator[lane_index].Annotations.setValue(
self.trackingApplet.topLevelOperator[loaded_project_lane_index].Annotations.value)
self.cropSelectionApplet.topLevelOperator[lane_index].Crops.setValue(
self.trackingApplet.topLevelOperator[loaded_project_lane_index].Crops.value)
logger.info("Test: Structured Learning")
weights = self.trackingApplet.topLevelOperator[lane_index]._runStructuredLearning(
z_range,
parameters['maxObj'],
parameters['max_nearest_neighbors'],
parameters['maxDist'],
parameters['divThreshold'],
[parameters['scales'][0],parameters['scales'][1],parameters['scales'][2]],
parameters['size_range'],
parameters['withDivisions'],
parameters['borderAwareWidth'],
parameters['withClassifierPrior'],
withBatchProcessing=True)
logger.info("weights: {}".format(weights))
logger.info("Test: Tracking")
self.trackingApplet.topLevelOperator[lane_index].track(
time_range = time_enum,
x_range = x_range,
y_range = y_range,
z_range = z_range,
size_range = parameters['size_range'],
x_scale = parameters['scales'][0],
y_scale = parameters['scales'][1],
z_scale = parameters['scales'][2],
maxDist=parameters['maxDist'],
maxObj = parameters['maxObj'],
divThreshold=parameters['divThreshold'],
avgSize=parameters['avgSize'],
withTracklets=parameters['withTracklets'],
sizeDependent=parameters['sizeDependent'],
detWeight=parameters['detWeight'],
divWeight=parameters['divWeight'],
transWeight=parameters['transWeight'],
withDivisions=parameters['withDivisions'],
withOpticalCorrection=parameters['withOpticalCorrection'],
withClassifierPrior=parameters['withClassifierPrior'],
ndim=ndim,
withMergerResolution=parameters['withMergerResolution'],
borderAwareWidth = parameters['borderAwareWidth'],
withArmaCoordinates = parameters['withArmaCoordinates'],
cplex_timeout = parameters['cplex_timeout'],
appearance_cost = parameters['appearanceCost'],
disappearance_cost = parameters['disappearanceCost'],
force_build_hypotheses_graph = False,
withBatchProcessing = True
)
def post_process_lane_export(self, lane_index):
# FIXME: This probably only works for the non-blockwise export slot.
# We should assert that the user isn't using the blockwise slot.
settings, selected_features = self.trackingApplet.topLevelOperator.getLane(lane_index).get_table_export_settings()
from lazyflow.utility import PathComponents, make_absolute, format_known_keys
if settings:
self.dataExportTrackingApplet.progressSignal.emit(-1)
raw_dataset_info = self.dataSelectionApplet.topLevelOperator.DatasetGroup[lane_index][0].value
project_path = self.shell.projectManager.currentProjectPath
project_dir = os.path.dirname(project_path)
dataset_dir = PathComponents(raw_dataset_info.filePath).externalDirectory
abs_dataset_dir = make_absolute(dataset_dir, cwd=project_dir)
known_keys = {}
known_keys['dataset_dir'] = abs_dataset_dir
nickname = raw_dataset_info.nickname.replace('*', '')
if os.path.pathsep in nickname:
nickname = PathComponents(nickname.split(os.path.pathsep)[0]).fileNameBase
known_keys['nickname'] = nickname
# use partial formatting to fill in non-coordinate name fields
name_format = settings['file path']
partially_formatted_name = format_known_keys( name_format, known_keys )
settings['file path'] = partially_formatted_name
req = self.trackingApplet.topLevelOperator.getLane(lane_index).export_object_data(
lane_index,
# FIXME: Even in non-headless mode, we can't show the gui because we're running in a non-main thread.
# That's not a huge deal, because there's still a progress bar for the overall export.
show_gui=False)
req.wait()
self.dataExportTrackingApplet.progressSignal.emit(100)
def _inputReady(self, nRoles):
slot = self.dataSelectionApplet.topLevelOperator.ImageGroup
if len(slot) > 0:
input_ready = True
for sub in slot:
input_ready = input_ready and \
all([sub[i].ready() for i in range(nRoles)])
else:
input_ready = False
return input_ready
def onProjectLoaded(self, projectManager):
"""
Overridden from Workflow base class. Called by the Project Manager.
If the user provided command-line arguments, use them to configure
the workflow inputs and output settings.
"""
# Configure the data export operator.
if self._data_export_args:
self.dataExportTrackingApplet.configure_operator_with_parsed_args( self._data_export_args )
# Configure headless mode.
if self._headless and self._batch_input_args and self._data_export_args:
logger.info("Beginning Batch Processing")
self.batchProcessingApplet.run_export_from_parsed_args(self._batch_input_args)
logger.info("Completed Batch Processing")
def handleAppletStateUpdateRequested(self):
"""
Overridden from Workflow base class
Called when an applet has fired the :py:attr:`Applet.statusUpdateSignal`
"""
# If no data, nothing else is ready.
input_ready = self._inputReady(2) and not self.dataSelectionApplet.busy
if not self.fromBinary:
opThresholding = self.thresholdTwoLevelsApplet.topLevelOperator
thresholdingOutput = opThresholding.CachedOutput
thresholding_ready = input_ready and len(thresholdingOutput) > 0
else:
thresholding_ready = input_ready
opTrackingFeatureExtraction = self.trackingFeatureExtractionApplet.topLevelOperator
trackingFeatureExtractionOutput = opTrackingFeatureExtraction.ComputedFeatureNamesAll
tracking_features_ready = thresholding_ready and len(trackingFeatureExtractionOutput) > 0
opCropSelection = self.cropSelectionApplet.topLevelOperator
croppingOutput = opCropSelection.Crops
cropping_ready = thresholding_ready and len(croppingOutput) > 0
objectCountClassifier_ready = tracking_features_ready
opObjectExtraction = self.objectExtractionApplet.topLevelOperator
objectExtractionOutput = opObjectExtraction.RegionFeatures
features_ready = thresholding_ready and \
len(objectExtractionOutput) > 0
opAnnotations = self.annotationsApplet.topLevelOperator
annotations_ready = features_ready and \
len(opAnnotations.Labels) > 0 and \
opAnnotations.Labels.ready() and \
opAnnotations.TrackImage.ready()
opStructuredTracking = self.trackingApplet.topLevelOperator
structured_tracking_ready = objectCountClassifier_ready and \
len(opStructuredTracking.EventsVector) > 0
busy = False
busy |= self.dataSelectionApplet.busy
busy |= self.annotationsApplet.busy
# busy |= self.dataExportAnnotationsApplet.busy
busy |= self.trackingApplet.busy
busy |= self.dataExportTrackingApplet.busy
self._shell.enableProjectChanges( not busy )
self._shell.setAppletEnabled(self.dataSelectionApplet, not busy)
if not self.fromBinary:
self._shell.setAppletEnabled(self.thresholdTwoLevelsApplet, input_ready and not busy)
self._shell.setAppletEnabled(self.trackingFeatureExtractionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.cellClassificationApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.divisionDetectionApplet, tracking_features_ready and not busy)
self._shell.setAppletEnabled(self.cropSelectionApplet, thresholding_ready and not busy)
self._shell.setAppletEnabled(self.objectExtractionApplet, not busy)
self._shell.setAppletEnabled(self.annotationsApplet, features_ready and not busy)
# self._shell.setAppletEnabled(self.dataExportAnnotationsApplet, annotations_ready and not busy and \
# self.dataExportAnnotationsApplet.topLevelOperator.Inputs[0][0].ready() )
self._shell.setAppletEnabled(self.trackingApplet, objectCountClassifier_ready and not busy)
self._shell.setAppletEnabled(self.dataExportTrackingApplet, structured_tracking_ready and not busy and \
self.dataExportTrackingApplet.topLevelOperator.Inputs[0][0].ready() )