Datasets:

openclimatefix

/

dwd-icon-eu

like 9

@peterdudfield
Hi,
I wanted to take a moment to thank you and your team for consistently providing prompt and detailed responses to my queries—it's much appreciated.

As I was reviewing the code repository, I came across a section that deals with "accumulated data channels" within the NWP variables.
https://github.com/openclimatefix/ocf_datapipes/blob/9ec252eeee44937c12ab52699579bdcace76e72f/ocf_datapipes/select/select_time_slice_nwp.py#L55

Could you please point me to where I can find a list or dictionary that defines which variables are accumulated and which are not?
Also, how would you handle the difference in length between accumulated and non-accumulated? (I noticed you used .diff() which is expected to make accum. array one-length shorter than non-accum)

            xr_non_accum = xr_min.sel({self.channel_dim_name: non_accum_channels})
            xr_sel_non_accum = xr_non_accum.sel(
                step=step_indexer, init_time_utc=init_time_indexer
            )

            # Slice out the channels which need to be diffed
            xr_accum = xr_min.sel({self.channel_dim_name: accum_channels})

            # Take the diff and slice requested data
            xr_accum = xr_accum.diff(dim="step", label="lower")
            xr_sel_accum = xr_accum.sel(step=step_indexer, init_time_utc=init_time_indexer)

            # Join diffed and non-diffed variables
            xr_sel = xr.concat([xr_sel_non_accum, xr_sel_accum], dim=self.channel_dim_name)`

Thanks in advance for your help.