Number of cases aggregated in title is incorrect.

[jwarner8]

Describe the bug

When running aggregate over 7 forecasts, the histograms report there are 2 forecasts being aggregated. This is the same for other aggregations (i.e. profiles).
In the code, it uses ndim = seq_coord.ndim in the plotting routine to set the title.

How to reproduce

Run with more than 2 cases and observe that the title does not match the number of cases used.

Expected behaviour

It to report the correct number of cases being aggregated

[jfrost-mo]

seq_coord.ndim is the number of dimension. We instead want the length of the sequence coordinate. seq_coord.size or len(seq_coord.points) is probably what we want.

[jwarner8]

Hmm, using seq_coord.points gives the forecast length, not the number of forecasts aggregated. I am debugging what the seq_coord gives us, and whether the collapsed cube retains information about number of cases (the forecast_reference_time is collapsed and will have bounds, but this isn't passed to _set_title_and_filename

[jfrost-mo]

That's true. If the data has been aggregated I don't think we retain the information on the number of cases. We could feed it through as an extra argument if necessary.

[jwarner8]

That's true. If the data has been aggregated I don't think we retain the information on the number of cases. We could feed it through as an extra argument if necessary.

OK. I am putting this on hold, possibly for a CSET surgery, as its not clear where we collapse and how that information should be passed through recipes etc.