This example demonstrates how to do segment brain tumor using Nvidia's AI-assisted annotation tool in batch mode (without GUI, using qMRMLSegmentEditorWidget) using 3D Slicer

NvidiaAiaaTumorSegmentation.py

# Load/generate input data
################################################

# Load master volume
import SampleData
sampleDataLogic = SampleData.SampleDataLogic()
masterVolumeNode = sampleDataLogic.downloadMRBrainTumor1()

# Define boundary points
import numpy as np
inputPoints = np.array([[  0.4685    ,  49.68717985,  25.36491174],
       [  0.4685    ,  30.92409445,  53.85700439],
       [  0.4685    ,   7.29650543,  34.3989899 ],
       [  0.4685    ,  28.83930719,   6.60182634],
       [ 17.6458556 ,  24.98187703,  30.92434445],
       [-22.80139768,  26.96458552,  30.92434445]])

aiaaModelName = "annotation_mri_brain_tumors_t2_wt"

# Run filter
################################################

# Create segmentation node to store the segmentation result
segmentationNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLSegmentationNode")
segmentationNode.CreateDefaultDisplayNodes()
segmentationNode.GetSegmentation().AddEmptySegment()

# Create segment editor to get access to effects
segmentEditorWidget = slicer.qMRMLSegmentEditorWidget()
segmentEditorWidget.setMRMLScene(slicer.mrmlScene)
segmentEditorNode = slicer.vtkMRMLSegmentEditorNode()
slicer.mrmlScene.AddNode(segmentEditorNode)
segmentEditorWidget.setMRMLSegmentEditorNode(segmentEditorNode)
segmentEditorWidget.setSegmentationNode(segmentationNode)
segmentEditorWidget.setMasterVolumeNode(masterVolumeNode)

# Run segmentation
segmentEditorWidget.setActiveEffectByName("Nvidia AIAA")
effect = segmentEditorWidget.activeEffect()

# optionally, custom server address can be set like this:
#  effect.self().ui.serverComboBox.currentText = serverUrl
#  effect.self().onClickFetchModels()

effect.self().ui.annotationModelSelector.currentText = aiaaModelName

# This is only needed for annotation (not for automatic segmentation)
slicer.util.updateMarkupsControlPointsFromArray(effect.self().annotationFiducialNode, inputPoints)

# for automatic segmentation use: effect.self().onClickSegmentation()
effect.self().onClickAnnotation()

# Clean up and show results
################################################

# Clean up
slicer.mrmlScene.RemoveNode(segmentEditorNode)

# Show result visible in 3D
segmentationNode.CreateClosedSurfaceRepresentation()

you can directly use AIAA python client to get everything done.. you don't need slicer behind the scene.

Reference:
https://github.com/NVIDIA/ai-assisted-annotation-client/blob/master/py_client/client_api.py
https://github.com/NVIDIA/ai-assisted-annotation-client/blob/master/py_client/test_aiaa_server.py#L117-L179

Thanks for the suggestion. This example shows how simple it is to use AIAA Slicer plugin if data is already in 3D Slicer. It is really just a few lines of trivial code (the rest of the code is just setting up the example - downloading sample data, etc.):

segmentEditorWidget.setActiveEffectByName("Nvidia AIAA")
effect = segmentEditorWidget.activeEffect()
effect.self().ui.annotationModelSelector.currentText = aiaaModelName
slicer.util.updateMarkupsControlPointsFromArray(effect.self().annotationFiducialNode, inputPoints)
effect.self().onClickAnnotation()

The underlying low-level AIAA client API is nice, too, but probably you would need 5-10x more code to get all the data from the Slicer scene, run the processing, and put back results into the Slicer scene.