Ibicus interface, xclim computation

[aulemahal]

This branch is an experiment of building an Ibicus interface to debiasing methods, but using xclim code for the computation. It is also a place where we can takes notes of the major (and minor!) caveats that make integrating xclim into ibicus difficult.

We could another such PR on xclim to test the opposite (a xclim interface using ibicus code for computation).

Caveats

• No xarray. Xclim mid-level adjustment function expect xarray objects. Rebuilding DataArrays from numpy arrays (that I had opened with xarray) feels weird, but it works.

[FussyDuck]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[aulemahal]

Woops, I meant to open that PR on the Ouranosinc fork, as so not to pollute this repo too much. Well, too late, I'll keep it here. I've just begun my exploration.

[coxipi]

@jakobwes Hi, we now have a separate module for bias adjustment and downscaling, xsdba, https://github.com/Ouranosinc/xsdba. Is a coordination xsdba/ibicus still of interest on your side? In this specific PR, we can simply replace xclim.sdba by xsdba, and add the package in dependencies.