Handle missing data gracefully in clipping.levels()

[wfvining]

features.clipping.levels() raises some very difficult to decipher value errors form numpy when there are NAs in the data. We should either raise a more meaningful and useful exception or simply deal with NAs by dropping them, applying the clipping filter, reindexing, and filling missing values with False.

[cwhanse]

Can you link an example?

[wfvining]

Running this:

daytime_power = power.copy()
daytime_power[power < 100] = pd.NA
clip_levels = clipping.levels(daytime_power)

Results in ValueError: autodetected range of [nan, nan] is not finite from numpy.histograms

If you do clipping.levels(daytime_power.dropna()) it works as expected, but you need to reindex the mask if you want to use it on the original data. My suggestion is to do the dropna() and reindex() inside clipping.levels().

Alternatively we can add to the documentation that the data must not contain undefined values and raise an exception with a meaningful/useful error message when undefined values are present.