Remove v1 refactor

[reginald-mclean]

This PR re-organizes the Meta-World repository by:

• removing the version 1 environments (we can keep the reward functions by adding a per environment flag about which reward function to use if desired, I think it might be a good idea)
• moving environments from metaworld/envs/mujoco/sawyer_xyz/v2 to metaworld/envs
• also move auxiliary files around that are imported/used by each environment

[pseudo-rnd-thoughts]

Looks good, may only suggestion would be have a metaworld/envs/sawyer_xyz then all the current environments as at a future date, we could consider adding alternative environments and this will be obvious the different types

[pseudo-rnd-thoughts]

I'm surprised that the test time hasn't reduced with this change. I would have thought removing half the environment would have reduced the time by 2.
I know you have answered this before @reginald-mclean but is it just a single test that takes most of the time?

[reginald-mclean]

@pseudo-rnd-thoughts I will run the tests locally to confirm, but the tests in tests/metaworld/envs/mujoco/sawyer_xyz/test_scripted_policies.py are most likely the lengthiest. They run the expert policy in each environment 50 times (one for each of those task objects created at initialization). The expert policy must achieve a success rate of 80% to pass the test.