Howto RL-WP-002: MLPro to PettingZoo

Prerequisites

Please install the following packages to run this examples properly:

Executable code

## -------------------------------------------------------------------------------------------------
## -- Project : MLPro - A Synoptic Framework for Standardized Machine Learning Tasks
## -- Package : mlpro.rl.examples
## -- Module  : howto_rl_wp_002_mlpro_environment_to_petting_zoo_environment.py
## -------------------------------------------------------------------------------------------------
## -- History :
## -- yyyy-mm-dd  Ver.      Auth.    Description
## -- 2021-10-02  0.0.0     SY       Creation
## -- 2021-10-02  1.0.0     SY       Released first version
## -- 2021-10-04  1.0.1     DA       Minor fix
## -- 2021-11-15  1.0.2     DA       Refactoring
## -- 2021-12-03  1.0.3     DA       Refactoring
## -- 2022-10-14  1.0.4     SY       Refactoring 
## -- 2022-11-02  1.0.5     SY       Unable logging in unit test model
## -- 2023-03-02  1.0.6     LSB      Refactoring
## -------------------------------------------------------------------------------------------------

"""
Ver. 1.0.6 (2023-03-02)

This module shows how to wrap mlpro's Environment class to petting zoo compatible.

1. How to setup an MLPro environment.

2. How to wrap MLPro's native envrionment to a Petting Zoo environment.
"""


from mlpro.bf.various import Log
from mlpro.wrappers.pettingzoo import WrEnvMLPro2PZoo
from mlpro.rl.pool.envs.bglp import BGLP
from pettingzoo.test import api_test


if __name__ == "__main__":
    logging = Log.C_LOG_ALL
else:
    logging = Log.C_LOG_NOTHING

# 1. Set up MLPro native environment
mlpro_env = BGLP(p_logging=logging)

# 2. Wrap the MLPro environment to PettingZoo compatible environment
env = WrEnvMLPro2PZoo(mlpro_env,
                      p_num_agents=5,
                      p_state_space=None,
                      p_action_space=None,
                      p_logging=logging).pzoo_env

# 3. Check whether the environment is valid
try:
    api_test(env, num_cycles=10, verbose_progress=False)
    print("test completed")
    assert True
except:
    print("test failed")  
    assert False

Results

The Bulk Good Laboratory Plant (BGLP) environment will be wrapped to a PettingZoo compliant environment.

YYYY-MM-DD  HH:MM:SS.SSSSSS  I  Environment BGLP: Instantiated
YYYY-MM-DD  HH:MM:SS.SSSSSS  I  Environment BGLP: Instantiated
YYYY-MM-DD  HH:MM:SS.SSSSSS  I  Environment BGLP: Operation mode set to 0
YYYY-MM-DD  HH:MM:SS.SSSSSS  I  Environment BGLP: Reset
Starting API test
...
Passed API test
test completed

There are several lines of action processing logs due to the API tests. When there is no detected failure, the environment is successfully wrapped.

Cross Reference