Add wrapper checker utility for validating wrapper implementations

[Jatin-Shihora]

Description

Add a check_wrapper() utility function for validating that wrapper implementations follow Gymnasium's Wrapper API correctly, similar to the existing check_env() for environments.

The checker verifies:

• The wrapper is a valid gymnasium.Wrapper instance with an inner environment
• Observation and action spaces exist and are valid
• unwrapped returns the base environment
• reset() returns (obs, info) with correct types and obs in space
• step() returns (obs, reward, terminated, truncated, info) with correct types
• Reward is a valid number (warns on NaN/infinity)
• render() and close() don't crash

Fixes #6

Type of change

• New feature (non-breaking change which adds functionality)

Checklist:

• I have run the pre-commit checks with pre-commit run --all-files (see CONTRIBUTING.md instructions to set it up)
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation ()
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[pseudo-rnd-thoughts]

Thanks for the PR, this overall looks good.
Could you add a test to check every wrapper that we implement to check that they all work as expected

[Jatin-Shihora]

Thanks for the PR, this overall looks good. Could you add a test to check every wrapper that we implement to check that they all work as expected

Hi @pseudo-rnd-thoughts thanks for reviewing the PR. I just added tests that run check_wrapper() against 27 of built-in wrappers to verify they all pass the checker.

I split the wrappers into three groups based on the environment they need:

1. CartPole (Discrete actions, Box obs): 19 wrappers including TimeLimit, Autoreset, FlattenObservation, NormalizeObservation, ClipReward, StickyAction, etc.
2. MountainCarContinuous (Box actions): 3 wrappers that require continuous action spaces: ClipAction, RescaleAction, TransformAction
3. Special environments via GenericTestEnv: 5 wrappers that need custom setups
   • GrayscaleObservation: using a Box(0, 255, shape=(25,25,3), dtype=uint8) image env
   • DiscretizeObservation: using a finite-bounded Box obs env
   • DiscretizeAction: using a finite-bounded Box action env
   • RenderCollection and AddRenderObservation: using render_mode="rgb_array" with a mock render function

The remaining wrappers I couldn't find a clean way to test:

• ResizeObservation: imports cv2 (OpenCV) at init time, so it raises DependencyNotInstalled if opencv isn't available. Would need gymnasium[other] installed in the test environment.
• AtariPreprocessing: requires Atari ROM files to construct the base environment.
• RecordVideo: requires moviepy to be installed for video encoding.
• HumanRendering, AddWhiteNoise, ObstructView: these rendering wrappers need a real display backend (pygame/OpenGL), which doesn't work in headless CI.
• JaxToNumpy, JaxToTorch, NumpyToTorch: lazy-loaded and require jax/torch to be installed.

If there's a preferred way to handle these (maybe pytest.importorskip for the ones that just need optional deps?), happy to add those too.

total tests: 41 tests (14 original + 27 new ones)