tbp.monty.sensor_modules#

tbp.monty.sensor_modules.sensor_module#

class Payload(observation: SensorObservation, percept: Message | None, goals: list[Goal], region: AttentionRegion | None, telemetry: dict[str, Any]) → None[source]#

Bases: object

__init__(observation: SensorObservation, percept: Message | None, goals: list[Goal], region: AttentionRegion | None, telemetry: dict[str, Any]) → None#
goals: list[Goal]#
observation: SensorObservation#
percept: Message | None#
region: AttentionRegion | None#
telemetry: dict[str, Any]#
class RuntimeSensorModule(*args, **kwargs)[source]#

Bases: Protocol

Monty runtime interface to a Sensor Module.

__init__(*args, **kwargs)#
propose_goals() → Collection[Goal][source]#

Return the goals proposed by this Sensor Module.

Return type:

Collection[Goal]

Returns:

A collection of proposed Goals.

propose_region() → AttentionRegion[source]#

Return the region proposed by this Sensor Module.

Return type:

AttentionRegion

Returns:

An AttentionRegion.

step(ctx: RuntimeContext, observation: SensorObservation, motor_only_step: bool = False) → Message | None[source]#

Execute a time-step for the Sensor Module.

Parameters:

  • ctx (RuntimeContext) – The runtime context.

  • observation (SensorObservation) – Sensor observation.

  • motor_only_step (bool) – Whether the current step is a motor-only step.

Return type:

Message | None

Returns:

An optional percept with features and morphological features.

update_state(agent: AgentState) → None[source]#

Update the proprioceptive state for this Sensor Module.

Parameters:

agent (AgentState) – The proprioceptive state of this sensor module’s Agent.

Return type:

None

class SensorModule(sensor_module_id: str, sensor_id: SensorID, transforms: Sequence[tbp.monty.sensor_modules.sensor_module.Transform][Sequence[Transform]]) → None[source]#

Bases: RuntimeSensorModule, Snapshotable

A sensor module implemented as a pipeline of configurable transforms.

This sensor module does not implement a specific observation processing itself. Instead, it is a generic mechanism for processing a SensorObservation through a configurable sequence of transforms to produce sensor module outputs.

__init__(sensor_module_id: str, sensor_id: SensorID, transforms: Sequence[tbp.monty.sensor_modules.sensor_module.Transform][Sequence[Transform]]) → None[source]#

Initialize the sensor module.

Parameters:

  • sensor_module_id (str) – The unique identifier for this sensor module.

  • sensor_id (NewType()(SensorID, str)) – The unique identifier for the sensor associated with this module.

  • transforms (Sequence[Sequence[Transform]]) – A sequence of transforms to process the Payload containing the sensor observation and produce sensor module outputs.

load_state_dict(memento: Mapping[str, Any]) → None[source]#

Load the internal object state from a previously snapshot memento.

Parameters:

memento (Mapping[str, Any]) – State dict to load.

Return type:

None

propose_goals() → Collection[Goal][source]#

Return the goals proposed by this Sensor Module.

Return type:

Collection[Goal]

Returns:

A collection of proposed Goals.

propose_region() → AttentionRegion[source]#

Return the region proposed by this Sensor Module.

Return type:

AttentionRegion

Returns:

An AttentionRegion.

state_dict() → Mapping[str, Any][source]#

Snapshot a memento representing the internal state of this object.

Return type:

Mapping[str, Any]

Returns:

State dict for logging and saving.

step(ctx: RuntimeContext, observation: SensorObservation, motor_only_step: bool = False) → Message | None[source]#

Execute a time-step for the Sensor Module.

Parameters:

  • ctx (RuntimeContext) – The runtime context.

  • observation (SensorObservation) – Sensor observation.

  • motor_only_step (bool) – Whether the current step is a motor-only step.

Return type:

Message | None

Returns:

An optional percept with features and morphological features.

update_state(agent: AgentState) → None[source]#

Update the proprioceptive state for this Sensor Module.

Parameters:

agent (AgentState) – The proprioceptive state of this sensor module’s Agent.

Return type:

None

property sensor_module_id: str#
class Transform(*args, **kwargs)[source]#

Bases: Protocol

A transform that can be applied to a payload.

__init__(*args, **kwargs)#
class TransformContext(rng: numpy.random.RandomState, agent_state: AgentState, sensor_state: SensorState, motor_only_step: bool = False, suppress_runtime_errors: bool = False) → None[source]#

Bases: object

__init__(rng: numpy.random.RandomState, agent_state: AgentState, sensor_state: SensorState, motor_only_step: bool = False, suppress_runtime_errors: bool = False) → None#
agent_state: AgentState#
motor_only_step: bool = False#
rng: RandomState#
sensor_state: SensorState#
suppress_runtime_errors: bool = False#