Fixes wait time in `play.py` by using `env.step_dt` (#2239)

# Description

When running `play.py` with `--real-time`, the dt used to calculate this
is incorrect. It is currently using `env.physics_dt`, which is `sim_dt`.
However, if the decimation is >1, then the effective dt is
`env.step_dt`, which is `sim_dt * decimation`. We are running 1
env.step() per loop, so this should definitely be `env.step_dt`.

This affects all reinforcement_learning/<rl_library>/play.py files. This
updates all of these appropriately

Fixes #2230 

## Checklist

- [x] I have run the [`pre-commit` checks](https://pre-commit.com/) with
`./isaaclab.sh --format`
- [x] I have made corresponding changes to the documentation
- [x] My changes generate no new warnings
- [x] (No need for test for small change in script) I have added tests
that prove my fix is effective or that my feature works
- [x] (No need for changelog in scripts) I have updated the changelog
and the corresponding version in the extension's `config/extension.toml`
file
- [x] I have added my name to the `CONTRIBUTORS.md` or my name already
exists there
Signed-off-by: Tyler Lum <tylergwlum@gmail.com>

CONTRIBUTORS.md

@@ -93,6 +93,7 @@ Guidelines for modifications:
 * Shafeef Omar
 * Shundo Kishi
 * Stephan Pleines
+* Tyler Lum
 * Victor Khaustov
 * Vladimir Fokow
 * Wei Yang

scripts/reinforcement_learning/rl_games/play.py

@@ -152,7 +152,7 @@ def main():
     agent.restore(resume_path)
     agent.reset()
 
-    dt = env.unwrapped.physics_dt
+    dt = env.unwrapped.step_dt
 
     # reset environment
     obs = env.reset()

scripts/reinforcement_learning/rsl_rl/play.py

@@ -140,7 +140,7 @@ def main():
         ppo_runner.alg.policy, normalizer=ppo_runner.obs_normalizer, path=export_model_dir, filename="policy.onnx"
     )
 
-    dt = env.unwrapped.physics_dt
+    dt = env.unwrapped.step_dt
 
     # reset environment
     obs, _ = env.get_observations()

scripts/reinforcement_learning/sb3/play.py

@@ -134,7 +134,7 @@ def main():
     print(f"Loading checkpoint from: {checkpoint_path}")
     agent = PPO.load(checkpoint_path, env, print_system_info=True)
 
-    dt = env.unwrapped.physics_dt
+    dt = env.unwrapped.step_dt
 
     # reset environment
     obs = env.reset()

scripts/reinforcement_learning/skrl/play.py

@@ -137,11 +137,11 @@ def main():
     if isinstance(env.unwrapped, DirectMARLEnv) and algorithm in ["ppo"]:
         env = multi_agent_to_single_agent(env)
 
-    # get environment (physics) dt for real-time evaluation
+    # get environment (step) dt for real-time evaluation
     try:
-        dt = env.physics_dt
+        dt = env.step_dt
     except AttributeError:
-        dt = env.unwrapped.physics_dt
+        dt = env.unwrapped.step_dt
 
     # wrap for video recording
     if args_cli.video: