Should LabelEncoder class have `inverse` as a parameter? about dstoolbox

[benjamin-work]

That's a reasonable requirement. Honestly I would drop the inverse=True parameter because of the reasons such as the one you mentioned.

I could change this, though probably not this week, but would also accept a PR :)

Interestingly, sklearn seems to have a similar feature in their pipeline:

https://github.com/scikit-learn/scikit-learn/blob/755c7dda3885e52c3439fc5a50903408d642bc58/sklearn/compose/_target.py#L17

I'm not sure whether this covers all the use cases of PipelineY but in general, it's probably better to de-couple the target transformation from the Pipeline.