Are pipelines now supported?

[nxorable]

Per #11

...it appears we are waiting for a pipeline.get_feature_names() implementation to support Pipeline. Is this still the case, when we have something like a pipeline whose transformer expands the dimensionality prior to fitting with a classifier?

[oegedijk]

It seems that this feature is now available in sklearn v1, so should be able to add this yes!

https://scikit-learn.org/stable/modules/generated/sklearn.pipeline.Pipeline.html#sklearn.pipeline.Pipeline.get_feature_names_out

[nxorable]

That's good news! Any estimate on the date of this support enhancement?

[nxorable]

Friendly bump (as opposed to the unfriendly variety)

[oegedijk]

I've had zero time for this in the past two months (moving house with two young children is not exactly OSS development friendly :), but should have some time in the coming weeks so will see if I can do this.

If you'd like to help: basically just need a piece of code that take a Pipeline and extracts the dataframe that goes into the final model (with proper column names) and the final model.

[angelmtenor]

Here you can find a piece of code I use for that:

#213 (comment)