Target value

---

Air Force Research Laboratory (AFRL) Autonomous Capabilities Team (ACT3) Reinforcement Learning (RL) Core.

This is a US Government Work not subject to copyright protection in the US.

The use, dissemination or disclosure of data in this file is subject to limitation or restriction. See accompanying README and LICENSE for details.

---

LimitConfigValidator (BaseModel) pydantic-model

minimum: defines the min the obs of the glue may produce maximum: defines the max the obs of the glue may produce unit: defines the unit for the limits clip: defines if the glue should clip the obs values to between min/max

Source code in corl/glues/common/target_value.py

class LimitConfigValidator(BaseModel):
    """
    minimum: defines the min the obs of the glue may produce
    maximum: defines the max the obs of the glue may produce
    unit: defines the unit for the limits
    clip: defines if the glue should clip the obs values to between min/max
    """
    minimum: float
    maximum: float
    unit: typing.Optional[str]
    clip: bool = False

TargetValue (BaseAgentPlatformGlue)

Glue class for an observation of some constant value

Source code in corl/glues/common/target_value.py

class TargetValue(BaseAgentPlatformGlue):
    """
    Glue class for an observation of some constant value
    """

    # pylint: disable=too-few-public-methods
    class Fields:
        """
        field data
        """
        TARGET_VALUE = "target_value"

    def __init__(self, **kwargs) -> None:
        self.config: TargetValueValidator
        super().__init__(**kwargs)

    @property
    def get_validator(self) -> typing.Type[TargetValueValidator]:
        return TargetValueValidator

    @lru_cache(maxsize=1)
    def get_unique_name(self) -> str:
        """Provies a unique name of the glue to differentiate it from other glues.
        """
        return f"{self.config.name}TargetValue"

    def invalid_value(self) -> OrderedDict:
        """When invalid return a value of 0

        TODO: this may need to be self.min in the case that the minimum is larger than 0 (i.e. a harddeck)

        Returns:
            OrderedDict -- Dictionary with <FIELD> entry containing 1D array
        """
        d = OrderedDict()
        d[self.Fields.TARGET_VALUE] = np.asarray([self.config.target_value], dtype=np.float32)
        return d

    @lru_cache(maxsize=1)
    def observation_units(self):
        """Return a space dictionary that indicates the units of the observation
        """
        d = gym.spaces.dict.Dict()
        d.spaces[self.Fields.TARGET_VALUE] = [self.config.unit]
        return d

    @lru_cache(maxsize=1)
    def observation_space(self) -> gym.spaces.Space:
        """Describes observation space for a single scalar continuous value

        Returns:
            gym.spaces.Space -- The rllib policy observation space
        """

        d = gym.spaces.dict.Dict()
        d.spaces[self.Fields.TARGET_VALUE
                 ] = gym.spaces.Box(self.config.limit.minimum, self.config.limit.maximum, shape=(1, ), dtype=np.float32)
        return d

    def get_observation(self) -> OrderedDict:
        """Constructs a portion of the observation space

        Arguments:
            platforms {List[BasePlatform]} -- List of current platforms.
                Gauranteed to be length of 1 and set to self.obs_agent

        Returns:
            OrderedDict -- Dictionary containing observation in <FIELD> key
        """

        d = OrderedDict()
        d[self.Fields.TARGET_VALUE] = np.asarray([self.config.target_value], dtype=np.float32)
        return d

    def action_space(self):
        ...

    def apply_action(self, action, observation):
        ...

get_validator: Type[corl.glues.common.target_value.TargetValueValidator] property readonly

returns the validator for this class

Returns:

Type	Description
Type[corl.glues.common.target_value.TargetValueValidator]	BaseAgentGlueValidator -- A pydantic validator to be used to validate kwargs

Fields

field data

Source code in corl/glues/common/target_value.py

class Fields:
    """
    field data
    """
    TARGET_VALUE = "target_value"

action_space(self)

Build the action space for the controller, etc. that defines the action given to apply_action

Returns

gym.spaces.Space The gym Space that defines the action given to the apply_action function

Source code in corl/glues/common/target_value.py

def action_space(self):
    ...

apply_action(self, action, observation)

Apply the action for the controller, etc.

Parameters

action The action for the class to apply to the platform observation The current observations before appling the action

Source code in corl/glues/common/target_value.py

def apply_action(self, action, observation):
    ...

get_observation(self)

Constructs a portion of the observation space

Returns:

Type	Description
OrderedDict	OrderedDict -- Dictionary containing observation in key

Source code in corl/glues/common/target_value.py

def get_observation(self) -> OrderedDict:
    """Constructs a portion of the observation space

    Arguments:
        platforms {List[BasePlatform]} -- List of current platforms.
            Gauranteed to be length of 1 and set to self.obs_agent

    Returns:
        OrderedDict -- Dictionary containing observation in <FIELD> key
    """

    d = OrderedDict()
    d[self.Fields.TARGET_VALUE] = np.asarray([self.config.target_value], dtype=np.float32)
    return d

get_unique_name(self)

Provies a unique name of the glue to differentiate it from other glues.

Source code in corl/glues/common/target_value.py

@lru_cache(maxsize=1)
def get_unique_name(self) -> str:
    """Provies a unique name of the glue to differentiate it from other glues.
    """
    return f"{self.config.name}TargetValue"

invalid_value(self)

When invalid return a value of 0

TODO: this may need to be self.min in the case that the minimum is larger than 0 (i.e. a harddeck)

Returns:

Type	Description
OrderedDict	OrderedDict -- Dictionary with entry containing 1D array

Source code in corl/glues/common/target_value.py

def invalid_value(self) -> OrderedDict:
    """When invalid return a value of 0

    TODO: this may need to be self.min in the case that the minimum is larger than 0 (i.e. a harddeck)

    Returns:
        OrderedDict -- Dictionary with <FIELD> entry containing 1D array
    """
    d = OrderedDict()
    d[self.Fields.TARGET_VALUE] = np.asarray([self.config.target_value], dtype=np.float32)
    return d

observation_space(self)

Describes observation space for a single scalar continuous value

Returns:

Type	Description
Space	gym.spaces.Space -- The rllib policy observation space

Source code in corl/glues/common/target_value.py

@lru_cache(maxsize=1)
def observation_space(self) -> gym.spaces.Space:
    """Describes observation space for a single scalar continuous value

    Returns:
        gym.spaces.Space -- The rllib policy observation space
    """

    d = gym.spaces.dict.Dict()
    d.spaces[self.Fields.TARGET_VALUE
             ] = gym.spaces.Box(self.config.limit.minimum, self.config.limit.maximum, shape=(1, ), dtype=np.float32)
    return d

observation_units(self)

Return a space dictionary that indicates the units of the observation

Source code in corl/glues/common/target_value.py

@lru_cache(maxsize=1)
def observation_units(self):
    """Return a space dictionary that indicates the units of the observation
    """
    d = gym.spaces.dict.Dict()
    d.spaces[self.Fields.TARGET_VALUE] = [self.config.unit]
    return d

TargetValueValidator (BaseAgentPlatformGlueValidator) pydantic-model

target_value: the Value this glue should produce this episode unit: the unit of the value produced

Source code in corl/glues/common/target_value.py

class TargetValueValidator(BaseAgentPlatformGlueValidator):
    """
    target_value: the Value this glue should produce this episode
    unit: the unit of the value produced
    """
    target_value: float
    unit: str
    limit: LimitConfigValidator