ENH: Add a user_index operator to aggregate values between spatially varying dates (1 date per pixel)

[bzah]

• icclim version: 5.2.1

Description

Originally posted by @tlogan2000 and @pagecp in #162 (comment)

xclim provides a generic function to aggregate data between two ndarray of dates on simple reducers {‘min’, ‘max’, ‘sum’, ‘mean’, ‘std’}.

aggregate_between_dates allows for an array of start, end dates (i.e. that vary spatially over a large territory) to be used as inputs. A specific example would be something like calculating accumulated precipitation within a growing season.

Scientifically it would be quite useful to have this feature of having different start and end date possible depending on the gridpoint, and also the same for thresholds on field values (not percentiles).

In this ENH request we will focus on "different dates depending on the gridpoint".
For this we we should implement a new calc_operator to be used with user_index. To ease user experience I propose to stick with xclim's name "aggregate_between_dates".

Ideally, I would also like to come up with proper metadata generation for this calc_operator. Until now, metadata were quite poor when using user_index operators and it would be nice to come up with a generalist solution here to improve this situation.

[bzah]

Another option could be to enable a tuple[DataArray, DataArray] + a freq as a valid input for slice_mode (or add parameter for the tuple).

In that case it would enable this feature for all indices at once.
We could then compute, say su, on varying summer at the same time.

This could really go hand in hand with the generification done in #189.