Recent changes in intake-esm already created a 'member_id' dimension on datasets. This PR makes the xmip equivalent process more flexible (allowing to work with both intake-esm and raw zarr stores) and also fully conforms with the official reconstruction method (see under "File name template:") including accounting for sub_experiment_id (only relevant for DCPP experiments I believe).
Recent changes in intake-esm already created a 'member_id' dimension on datasets. This PR makes the xmip equivalent process more flexible (allowing to work with both intake-esm and raw zarr stores) and also fully conforms with the official reconstruction method (see under "File name template:") including accounting for
sub_experiment_id
(only relevant for DCPP experiments I believe).