Update get_agg_dates to work with multi-dim data, remove depreciated loffset parameter from resample and add unit test. Speed up min_tsteps option in temporal_aggregation by reducing the values loaded to 2 and removing the xr.where search. This assumes only the first and last aggregation period needs to be checked and they will be consistent across all other dimensions (mainly used to ensure full time periods are aggregated when target_freq != YE-DEC). Add unit tests for min_tsteps option in temporal_aggregation. Fix mask.values = error in select_time_period.
Update get_agg_dates to work with multi-dim data, remove depreciated loffset parameter from resample and add unit test. Speed up min_tsteps option in temporal_aggregation by reducing the values loaded to 2 and removing the xr.where search. This assumes only the first and last aggregation period needs to be checked and they will be consistent across all other dimensions (mainly used to ensure full time periods are aggregated when target_freq != YE-DEC). Add unit tests for min_tsteps option in temporal_aggregation. Fix mask.values = error in select_time_period.