Possible Bug - Error when creating numpy array out of list of lists having different length

[samiul123]

When labels are hierarchical and represented as a list of lists and all inner lists don't have equal length, numpy array constructor gives the following error

setting an array element with a sequence. The requested array has an inhomogeneous shape after 1 dimensions. The detected shape was (1781,) + inhomogeneous part.

This happens in HierarchicalClassifier.py at line 164.

Seeing the line at 171, where you are leveling the labels, I wonder if it should be done before converting to numpy array.

For Sample data, where Y_train_modifed = [['a'], ['b', 'c']]

This code gives the above error.
lcppn2 = LocalClassifierPerParentNode(local_classifier=model, verbose=1, bert=True) lcppn2.fit(X_train, Y_train_modified)

hi-class version: 4.12.1
numpy version: 1.26.4

[mirand863]

Hi @samiul123,

Thank you for reporting this issue. I believe I fixed it in PR #138, but please let me know if the issue persists in the new version 4.13.1, which will be released shortly.

[samiul123]

Hi @mirand863, Thank you so much for fixing the issue. I'll let you know if I find any more problems.