Describe the workflow you want to enable

A parameter accepting custom bin edges as an array.

Describe your proposed solution

Use pd.cut() under the hood or any other computationally efficient method.

Describe alternatives you've considered, if relevant

pandas' cut function.

Additional context

Say if we need to bin ages as 'infant', 'kid', 'teen', 'adult', 'senior citizen' into [0, 1, 13, 20, 60, np.inf] etc.

We can't do it using KBinsDiscretizer currently.

I wonder if there is any reason behind not implementing it?