`env.observation_space` doesn't match the shape of actual `observation` returned during env reset/step

[rejuvyesh]

e.g.

size(reset(antenv), 1) == antenv.observation_space.n

returns false.

I'll add those tests in a PR.

Reason is because although wedefine observation space as https://github.com/dojo-sim/Dojo.jl/blob/main/environments/ant/methods/env.jl#L44-L51 but then get_observation concatenates the contact forces as well: https://github.com/dojo-sim/Dojo.jl/blob/main/environments/ant/methods/env.jl#L175. We need a way to reflect this in the observation space.

[janbruedigam]

I cut down the Environments in the latest version to be very minimal, and observartion_space does no longer exist. I think Dojo's environments should really mostly be about the dynamics. As an example, for RL, you would have a Dojo Environment for the dynamical system, and a ReinforcementLearning Environment for the actual learning including rewards, state spaces, and so on.
But Dojo's Environments definitely need rethinking.

I'll close this for now, but feel free to open a general issue on Environments