Potential bug in agent actions

[thumsing]

Sometimes, it looks like both red and blue agents are taking actions on the same drones.

E.g., after 48 time steps in one of my episodes, drones 1, 8, and 14 all have both red and blue agents executing actions on them.

{key: value for key, value in wrapped_cyborg.env.environment_controller.action.items() if key.split("_")[0] == "blue" or (key.split("_")[0] == "red" and str(value) not in ["InvalidAction", "Sleep"])}

{'blue_agent_0': BlockTraffic,
 'blue_agent_2': RemoveOtherSessions,
 'blue_agent_3': RetakeControl,
 'blue_agent_4': RetakeControl,
 'blue_agent_5': AllowTraffic,
 'blue_agent_6': AllowTraffic,
 'blue_agent_7': RetakeControl,
 'blue_agent_9': RetakeControl,
 'blue_agent_10': BlockTraffic,
 'blue_agent_11': RetakeControl,
 'blue_agent_12': RemoveOtherSessions,
 'blue_agent_13': BlockTraffic,
 'blue_agent_14': RetakeControl,
 'blue_agent_15': BlockTraffic,
 'blue_agent_16': AllowTraffic,
 'blue_agent_17': RetakeControl,
 'red_agent_1': BlockTraffic,
 'blue_agent_1': Sleep,
 'red_agent_8': BlockTraffic,
 'blue_agent_8': Sleep,
 'red_agent_14': FloodBandwidth}

In another episode, after 186 time steps, on drone 3, red agent is flooding bandwidth and blue agent is blocking traffic.

{key: value for key, value in wrapped_cyborg.env.environment_controller.action.items() if key.split("_")[0] == "blue" or (key.split("_")[0] == "red" and str(value) not in ["InvalidAction", "Sleep"])}

{'blue_agent_0': RetakeControl,
 'blue_agent_1': BlockTraffic,
 'blue_agent_2': AllowTraffic,
 'blue_agent_3': BlockTraffic,
 'blue_agent_5': RetakeControl,
 'blue_agent_6': BlockTraffic,
 'blue_agent_8': AllowTraffic,
 'blue_agent_9': RetakeControl,
 'blue_agent_10': BlockTraffic,
 'blue_agent_11': AllowTraffic,
 'blue_agent_12': BlockTraffic,
 'blue_agent_13': AllowTraffic,
 'blue_agent_14': AllowTraffic,
 'blue_agent_15': BlockTraffic,
 'blue_agent_17': AllowTraffic,
 'red_agent_3': FloodBandwidth,
 'red_agent_4': FloodBandwidth,
 'blue_agent_4': Sleep,
 'blue_agent_7': Sleep,
 'red_agent_16': FloodBandwidth,
 'blue_agent_16': Sleep}