Richard Edgar

What do you mean by 'proper error handling' @romanlutz ? The error is coming from `sklearn` not `MetricFrame`. We could have `MetricFrame` catch any exceptions from the metric functions, and...

How would `MetricFrame` know to skip any particular metric? As I said, we could certainly catch and wrap the exception, but 'which metric failed to process the input data' is...

The problem is that at this stage, `MetricFrame` is only requiring the inputs to be 1D arrays. The contents of the arrays is the preserve of the metric functions themselves....

Re-reading @hildeweerts comment, that does make sense, so long as we're handling the 'coerce' case by calling the metric function itself, and not trying to figure out in advance what...

The PR I started a long time ago about this: https://github.com/fairlearn/fairlearn/pull/480 This was more about explaining to myself what was going on.

We definitely should be figuring out how to offer metrics for other scenarios. However, I'm a little concerned about a few points. Doing (1) is going to break all existing...

Is there a way of detecting whether an argument was passed by position rather than by keyword?

I think that there are a few different things going on here, and I'd like to separate the different threads. Firstly, I think that the idea behind `shared_sample_params` is a...

> @riedgar-ms : one of the reason to have keyword-only parameters is so that you can skip them when they're not needed. For example, if we went with the signature:...

It would break existing code, and require us to bump the version to v0.7.0. If we do that, I think we should release v0.7.0 as soon as possible after making...