Reports can only report missing observations if those observations were uploaded with nan values. Missing values due to discontinuous observations are not caught by discard before resample procedure. _calc_discard_before_resample might need to resample to the observation frequency (not the forecast frequency!) or compute a difference with an expected index to determine the number of missing values. Or maybe that happens in _resample_obs since it has knowledge of the metadata and we probably don't want to mess up the bools returned by _calc_discard_before_resample.
Reports also only report the number of missing resampled observations in combination with a quality flag. Would be good to report missing resampled observations on their own.
Reports can only report missing observations if those observations were uploaded with nan values. Missing values due to discontinuous observations are not caught by discard before resample procedure.
_calc_discard_before_resample
might need to resample to the observation frequency (not the forecast frequency!) or compute a difference with an expected index to determine the number of missing values. Or maybe that happens in_resample_obs
since it has knowledge of the metadata and we probably don't want to mess up the bools returned by_calc_discard_before_resample
.Reports also only report the number of missing resampled observations in combination with a quality flag. Would be good to report missing resampled observations on their own.