Display Wrappers not aligned with Gymnasium
Gymnasium's reset method returns a tuple containing the first observation and the info dict. This is currently not the case for our wrappers. While most algorithms we have don't throw an error, this is wrong and should be fixed.
Consider this example that I have just discovered:
self._last_obs = self.env.reset()
from https://collaborating.tuhh.de/ckv0173/Scilab-RL/-/blob/devel/src/custom_algorithms/cleansac/cleansac.py?ref_type=heads#L200 is actually a tuple because
def reset(self, **kwargs):
observations = self.env.reset(**kwargs)
if not self.displaying and self._display_enabled():
self.start_displayer()
return observations
in https://collaborating.tuhh.de/ckv0173/Scilab-RL/-/blob/devel/src/utils/custom_wrappers.py?ref_type=heads#L58 simply forwards the tuple.