update to version 1.2.3

[BDonnot]

• [ADDED] l2rpn-baselines package dependency in the "binder" environment.
• [FIXED] an issue in the sampling of redispatching action (ramp up and ramp down were inverted)
• [FIXED] an issue causing errors when using action_space.change_bus and action_space.set_bus
• [FIXED] an issue in the sampling: redispatching and "change_bus" where always performed at the
  same time
• [FIXED] rte-france#144: typo that could lead to not
  display some error messages in some cases.
• [FIXED] rte-france#146 : akward behaviour that lead to not calling
  the reward function when the episode was over.
• [FIXED] rte-france#147 : un consistency between step and simulate
  when cooldowns where applied (rule checking was not using the right method).
• [FIXED] An error preventing the loading of an Ambiguous Action (in case an agent took such action, the EpisodeData
  would not load it properly.
• [IMPROVED] overall documentation of BaseEnv and Environment
• [IMPROVED] rationalize the public and private part of the API for Environment and BaseEnv.
  Some members have been moved to private attribute (their modification would largely alterate the
  behaviour of grid2op).
• [IMPROVED] internal functions are tagged as "Internal, do not use" in the documentation.
• [IMPROVED] Improved documentation for the Environment and MultiMixEnvironment.