add access to data completeness via new diagnostics category

where data completeness corresponds to the actual number of time steps
available for each subset/sample, once missing data, mask, and bootstrap
have been applied