Is imbalanced-learn supported?

[xrml]

My pipeline has a random under sampler from imbalanced-learn to correct for imbalance in the classification dataset. When I try to convert to onnx, I get the following error:

File "\site-packages\skl2onnx\common_topology.py", line 628, in infer_types
raise MissingShapeCalculator(
skl2onnx.common.exceptions.MissingShapeCalculator: Unable to find a shape calculator for type '<class 'imblearn.under_sampling._prototype_selection._random_under_sampler.RandomUnderSampler'>'.

Has a converter for imbalanced-learn methods been implemented in skl2onnx? Or is this an issue to raise with imbalanced-learn?

[xadupre]

It is not supported but you can extend skl2onnx with your own converters. See https://onnx.ai/sklearn-onnx/tutorial_2_new_converter.html.