Jump to content

Connect SuperML | Leeroopedia MCP: Equip your AI agents with best practices, code verification, and debugging knowledge. Powered by Leeroo — building Organizational Superintelligence. Contact us at founders@leeroo.com.

Principle:Facebookresearch Habitat lab Component Class Definition

From Leeroopedia
Knowledge Sources
Domains Software_Architecture, Embodied_AI
Last Updated 2026-02-15 02:00 GMT

Overview

The abstract base class pattern for defining custom sensors, measures, and actions that extend the Habitat task framework.

Description

Component Class Definition is the first step in extending Habitat-Lab with custom functionality. Habitat provides three base classes that users subclass:

  • Sensor: Produces observations (e.g., custom image transforms, semantic labels, task-specific signals)
  • Measure: Computes evaluation metrics (e.g., custom success criteria, progress metrics)
  • SimulatorTaskAction: Defines new agent actions (e.g., custom movement, interaction)

Each base class defines abstract methods that the subclass must implement to specify the component's UUID (unique identifier), observation/metric space, and the core computation logic.

Usage

Use when the built-in Habitat sensors, measures, or actions do not cover your task requirements. This is the standard extension mechanism for adding custom task logic.

Theoretical Basis

The component system follows the Template Method pattern:

  1. UUID: Unique string identifier used in config and registry lookup
  2. Space/Type: Defines the shape and type of the output (observations or metrics)
  3. Compute: The core logic that produces the output given simulator state and episode context
# Abstract component interface
class Component(ABC):
    def _get_uuid(self) -> str: ...          # Unique identifier
    def _get_observation_space(self) -> Space: ...  # Output shape/type
    def get_observation(self, **kwargs) -> Any: ... # Core logic

Related Pages

Implemented By

Page Connections

Double-click a node to navigate. Hold to expand connections.
Principle
Implementation
Heuristic
Environment