Explore consistent python environments for users that create explainers

[axsaucedo]

As a precursor for #2825 (and parallel to #2584) one of the pieces to explore will be to identify what the best tool to enable our users to ensure they have replicated enviornments when they train an explainer or an model, whas this is one of the biggest bottlenecks when model binaries (pickles) are deployed - main issues when the library versions don't match (especially on exaplainers as the python version also takes precedence).

The two tools to explore initially will be conda.lock and poetry.lock - the former seems to have further support for environment components like Java and python version, whereas the latter seems to have a smoother interface.