Implementation:Facebookresearch Habitat lab Env integration testing

Knowledge Sources	Habitat-Lab
Domains	Testing, Embodied_AI
Last Updated	2026-02-15 02:00 GMT

Overview

Concrete pattern for testing custom Habitat components using habitat.Env and HabGymWrapper in live simulation environments.

Description

Integration testing uses habitat.Env (for native API testing) and HabGymWrapper / habitat.gym.make_gym_from_config (for Gym API testing) to validate that custom components produce expected outputs. The Env.reset() and Env.step() methods exercise the full component lifecycle.

Usage

Write test scripts that create environments with custom components and verify outputs. Use habitat_test_scenes dataset for lightweight testing.

Code Reference

Source Location

• Repository: habitat-lab
• File: habitat-lab/habitat/core/env.py (Env: L70-137, reset: L236-270, step: L283-315), habitat-lab/habitat/gym/gym_wrapper.py (HabGymWrapper: L190-251)

Signature

class Env:
    def __init__(
        self,
        config: DictConfig,
        dataset: Optional[Dataset] = None,
    ):
        """Create Habitat environment."""

    def reset(self) -> Observations:
        """Reset environment to new episode."""

    def step(
        self, action: Union[int, str, Dict[str, Any]]
    ) -> Observations:
        """Take action, return new observations."""

# Gym wrapper
class HabGymWrapper(gym.Env):
    def __init__(self, env: RLEnv, save_orig_obs: bool = False): ...
    def reset(self): ...
    def step(self, action): ...

Import

import habitat
from habitat.gym.gym_wrapper import HabGymWrapper
# Or: env = habitat.gym.make_gym_from_config(config)

I/O Contract

Inputs

Name	Type	Required	Description
config	DictConfig	Yes	Config with custom components wired in
action	Union[int, str, Dict]	Yes	Action to execute in env.step

Outputs

Name	Type	Description
observations	Dict	Sensor outputs including custom sensor values
info	Dict	Contains metrics dict with custom measure values

Usage Examples

Test Custom Components

import habitat

# Create env with custom components
env = habitat.Env(config=config)
obs = env.reset()

# Verify custom sensor output
assert "my_proximity" in obs
assert obs["my_proximity"].shape == (1,)

# Run a step and check metrics
obs = env.step("move_forward")
metrics = env.get_metrics()
assert "my_success" in metrics

env.close()

Test via Gym Wrapper

import habitat.gym

env = habitat.gym.make_gym_from_config(config)
obs, info = env.reset()
obs, reward, terminated, truncated, info = env.step(env.action_space.sample())
env.close()

Related Pages

Implements Principle

• Principle:Facebookresearch_Habitat_lab_Integration_Testing

Requires Environment

• Environment:Facebookresearch_Habitat_lab_Python_3_9_Core_Dependencies