AgentInterface

The agent interfaces for the BaseAgent instances.

Attributes:

Name	Type	Description
`actions`	`List[Action]`	list of possible actions
`action_space`	`ActionSpace`	ActionSpace instance
`active`	`bool`	flag of active (currently performing actions) agent
`agent`	`BaseAgent`	agent object instance
`agent_name`	`str`	name of the agent
`allowed_subnets`	`List[str]`	list of allowed subnets
`file`	`dict`
`group_name`	`dict`
`hostname`	`dict`
`interface_name`	`dict`
`internal_only`	`bool`	flag for the agent being an internal only agent
`last_action = None`
`messages`	`list`	list of messages
`path`	`dict`
`password`	`dict`
`password_hash`	`dict`
`process_name`	`dict`
`scenario`	`Scenario`	current scenario instance
`username = dict`

Functions

__init__(agent_obj, agent_name, actions, allowed_subnets, scenario, active = True, internal_only = False)

Parameters:

Name	Type	Default
`agent_obj`	`BaseAgent`	required
`agent_name`	`str`	required
`actions`	`List[Action]`	required
`allowed_subnets`	`List[str]`	required
`scenario`	`Scenario`	required
`active`	`bool`	`True`
`internal_only`	`bool`	`False`

create_reward_calculator(reward_calculator: str, agent_name: str, scenario: Scenario) -> RewardCalculator

Creates a reward calculator based on the name of the calculator to be used.

Parameters:

Name	Type	Description	Default
`reward_calculator`	`str`	name of reward calculator	required
`agent_name`	`str`	name of agent	required
`scenario`	`Scenario`	current scenario object	required

Returns:

Type	Description
`RewardCalculator`	created reward calculator

end_episode()

perform agent end of episode functionality and reset the interface

get_action(observation: Observation, action_space: dict = None)

Gets an action from the agent to perform on the environment

Parameters:

Name	Type	Description	Default
`observation`	`Observation`	agent observation space	required
`action_space`	`dict`	agent action space	`None`

Returns:

Name	Type	Description
`last_action`	`Action`	last action performed

reset()

resets the interface with empty dictionaries

set_init_obs(init_obs, true_obs)

set and update the true and initial observations

update(obs: dict, known: bool = True)

update the action space with the observation

update_allowed_subnets(allowed_subnets)

Updates allowed_subnets for agent

Attributes:

Name	Type	Description
`allowed_subnets`	`List(str)`	agent's allowed_subnets for mission phase