Allow arbitrary kwargs for preprocessors that use iris.cube.Cube.collapsed or iris.cube.Cube.aggregated_by

[schlunma]

Is your feature request related to a problem? Please describe.

Many iris operators used for collapsing/aggregating allow arbitrary kwargs (see example). We use iris.cube.Cube.collapsed/ iris.cube.Cube.aggregated_by in many preprocessors, so it might make sense to allow arbitrary kwargs in the recipes.

Affected preprocessors:

• multi_model_statistics
• ensemble_statistics
• climate_statistics
• area_statistics
• zonal_statistics
• meridional_statistics
• volume_statistics
• axis_statistics
• hourly_statistics
• daily_statistics
• monthly_statistics
• seasonal_statistics
• annual_statistics
• decadal_statistics

[valeriupredoi]

good idea! Do we know which of those are lazy? Also, do we know which are slower/faster? Do we care about performance - well, indirect performance, via iris using internal or external methods, but directly affecting our resources.

[schlunma]

Since basically all of them use iris, I guess that all of them are lazy. There's a list here, but not sure how up-to-date that is.

Regarding performance, I'm not quite sure how that is relevant to this issue, it's just about allowing additional options (which as far as I can tell are not about performance) 😄