[x] move check for 95% HWLW coverage from KWK_process.py to central location
[x] We did compute the number of expected timesteps with help of tstart/tstop here, would be nice to also generate an error if we compute havengetallen for 2021.0 (requiring 10 years of data) and only supply the last four years of the required timeseries >> added note to calc_havengetallen function
[x] replace the absolute min_count also with a argument-defined-percentage and a median(?) frequency derived from the dataframe index. Should not be mean, since large gaps will strongly affect this value
[x] min_count currently does not filter out nans in the yearly/monthly counts, this should be fixed
[x] before computing this freq, drop all nan values. Add tests for datasets with 3hr/1hr/10min/mixed freqs and 12.42/2hr freq
[x] do we want to set years with freq going from 60min to 10min to nan? In that case make sure this happens, but for means it might not be super relevant (it is for tidal analysis)
[x] when deriving tidalindicators, consider already providing nans for missing years, define periodindex with period_index = pd.period_range(data_wl_pd.index.min(), data_wl_pd.index.max(), freq="Y")
Todo:
min_count
also with a argument-defined-percentage and a median(?) frequency derived from the dataframe index. Should not be mean, since large gaps will strongly affect this valuemin_count
currently does not filter out nans in the yearly/monthly counts, this should be fixedperiod_index = pd.period_range(data_wl_pd.index.min(), data_wl_pd.index.max(), freq="Y")
index.diff()
is only possible in pandas 2.2.0