CDO-like convenience methods to select times

I feel like the time selecting features of xray can be improved. Currently, some common operations are too involved or verbose, like selecting the data in a group of months that are not a standard season (e.g. the monsoon season in india JJAS), or in non consecutive years (e.g. El Niño years). I think it would be great to implement (and easy), some methods inspired in the widely used Climate Data Operators https://code.zmaw.de/projects/cdo For example: selyear, selmon, selday and selhour.

Then we could easily do a composite of JJAS seasons in El Niño years like this:

pr_dataset = xray.open(my_precipitation_dataset)
elnino_years = [year list here...]
pr_dataset.selyear(elnino_years).selmon([6, 7, 8, 9]).mean('time')

This would make me very happy. The way to go would be to write methods that call grouby, then select the years/months, merge them, and return the corresponding dataset/dataarray, but I am not sure about what is the most efficient way to do this.

No files or tests are named. Start by reading xarray's existing groupby and time-selection entry points, then compare the proposed selyear, selmon, selday, and selhour behavior with the CDO operations. Done means agreeing on the API and semantics for non-consecutive years and month groups, implementing the methods, and covering the El Niño and JJAS examples with tests.