Add datacube variables support as MLM input and output dimensions

[fmigneault]

Description

• Add variables properties to Model Input Object to allow specifying the relevant data variables used by the model, with cross-references to the datacube extension (relates to #90).
• Add bands and variables properties to Model Output Object to allow specifying the relevant bands or variables produced by the model if any applies.
• Add downscaling to Tasks as common operation for climate variable models.
• Refactor the JSON schema to check for bands and variables references within both mlm:input and mlm:output. If either location detects that either bands or variables is provided, their corresponding sets of extensions providing relevant descriptions are verified.
• Refactor the JSON schema mlm:output property to employ a ModelOutput object definition rather than directly provided properties nested under the array.
• Refactor the JSON schema to allow the omission of bands under mlm:input if the variables property is provided.
• Fix missing encoding="utf-8" parameters in open calls leading to failing parsing of example JSON STAC Item when they contain non-ASCII characters.
• Fix parsing of PYTHON_PATH in Makefile for Windows OS.

Related Issue

• Fixes Support CF-variables as alternate I/O references #90

Type of Change

• 📚 Examples, docs, tutorials or dependencies update;
• 🔧 Bug fix (non-breaking change which fixes an issue);
• 🥂 Improvement (non-breaking change which improves an existing feature);
• 🚀 New feature (non-breaking change which adds functionality);
• 💥 Breaking change (fix or feature that would cause existing functionality to change);
• 🔐 Security fix.

Checklist

• I've read the CONTRIBUTING.md guide;
• I've updated the code style using make check;
• I've written tests for all new methods and classes that I created;
• I've written the docstring in Google format for all the methods and classes that I used.

[fmigneault]

@huard @Nazim-crim FYI

Initial MLM definition using datacube variables for describing a climate/ML model.
I have found a ERA5 downscaling model on GitHub code that should resemble the Ouranos UNet.
I used that reference for creating an example in the meantime. Only difference should be the actual variables definition using the RDPS/HRDPS instead of ERA5.

[rbavery]

Refactor the JSON schema to allow the omission of bands under mlm:input if the variables property is provided.

Is the reverse allowed, specifying bands but omitting variables ? I think the README might need to be updated, it currently says both are REQUIRED

[fmigneault]

@rbavery All comments should be addressed. Let me know if there is something else to adjust.