chore(processor): rename converters function names (#1853)

* chore(processor): rename to_transition_teleop_action -> action_to_transition

* chore(processor): rename to_transition_robot_observation -> observation_to_transition

* chore(processor): rename to_output_robot_action -> transition_to_robot_action

examples/phone_to_so100/evaluate.py

@@ -23,8 +23,8 @@ from lerobot.policies.act.modeling_act import ACTPolicy
 from lerobot.policies.factory import make_pre_post_processors
 from lerobot.processor import DataProcessorPipeline
 from lerobot.processor.converters import (
-    to_output_robot_action,
+    observation_to_transition,
-    to_transition_robot_observation,
+    transition_to_robot_action,
 )
 from lerobot.record import record_loop
 from lerobot.robots.so100_follower.config_so100_follower import SO100FollowerConfig
@@ -75,7 +75,7 @@ robot_ee_to_joints_processor = DataProcessorPipeline(
         ),
     ],
     to_transition=lambda tr: tr,
-    to_output=to_output_robot_action,
+    to_output=transition_to_robot_action,
 )
 
 # Build pipeline to convert joint observation to ee pose observation
@@ -83,7 +83,7 @@ robot_joints_to_ee_pose_processor = DataProcessorPipeline(
     steps=[
         ForwardKinematicsJointsToEE(kinematics=kinematics_solver, motor_names=list(robot.bus.motors.keys()))
     ],
-    to_transition=to_transition_robot_observation,
+    to_transition=observation_to_transition,
     to_output=lambda tr: tr,
 )
 

examples/phone_to_so100/record.py

@@ -22,9 +22,9 @@ from lerobot.datasets.utils import merge_features
 from lerobot.model.kinematics import RobotKinematics
 from lerobot.processor import DataProcessorPipeline
 from lerobot.processor.converters import (
-    to_output_robot_action,
+    action_to_transition,
-    to_transition_robot_observation,
+    observation_to_transition,
-    to_transition_teleop_action,
+    transition_to_robot_action,
 )
 from lerobot.record import record_loop
 from lerobot.robots.so100_follower.config_so100_follower import SO100FollowerConfig
@@ -88,7 +88,7 @@ phone_to_robot_ee_pose_processor = DataProcessorPipeline(
             max_ee_twist_step_rad=0.50,
         ),
     ],
-    to_transition=to_transition_teleop_action,
+    to_transition=action_to_transition,
     to_output=lambda tr: tr,
 )
 
@@ -106,7 +106,7 @@ robot_ee_to_joints_processor = DataProcessorPipeline(
         ),
     ],
     to_transition=lambda tr: tr,
-    to_output=to_output_robot_action,
+    to_output=transition_to_robot_action,
 )
 
 # Build pipeline to convert joint observation to ee pose observation
@@ -114,7 +114,7 @@ robot_joints_to_ee_pose = DataProcessorPipeline(
     steps=[
         ForwardKinematicsJointsToEE(kinematics=kinematics_solver, motor_names=list(robot.bus.motors.keys()))
     ],
-    to_transition=to_transition_robot_observation,
+    to_transition=observation_to_transition,
     to_output=lambda tr: tr,
 )
 

examples/phone_to_so100/replay.py

@@ -20,7 +20,7 @@ import time
 from lerobot.datasets.lerobot_dataset import LeRobotDataset
 from lerobot.model.kinematics import RobotKinematics
 from lerobot.processor import DataProcessorPipeline
-from lerobot.processor.converters import to_output_robot_action, to_transition_teleop_action
+from lerobot.processor.converters import action_to_transition, transition_to_robot_action
 from lerobot.robots.so100_follower.config_so100_follower import SO100FollowerConfig
 from lerobot.robots.so100_follower.robot_kinematic_processor import (
     AddRobotObservationAsComplimentaryData,
@@ -59,8 +59,8 @@ robot_ee_to_joints_processor = DataProcessorPipeline(
             initial_guess_current_joints=False,  # Because replay is open loop
         ),
     ],
-    to_transition=to_transition_teleop_action,
+    to_transition=action_to_transition,
-    to_output=to_output_robot_action,
+    to_output=transition_to_robot_action,
 )
 
 robot_ee_to_joints_processor.reset()

examples/phone_to_so100/teleoperate.py

@@ -17,7 +17,7 @@ import time
 
 from lerobot.model.kinematics import RobotKinematics
 from lerobot.processor import DataProcessorPipeline
-from lerobot.processor.converters import to_output_robot_action, to_transition_teleop_action
+from lerobot.processor.converters import action_to_transition, transition_to_robot_action
 from lerobot.robots.so100_follower.config_so100_follower import SO100FollowerConfig
 from lerobot.robots.so100_follower.robot_kinematic_processor import (
     AddRobotObservationAsComplimentaryData,
@@ -72,8 +72,8 @@ phone_to_robot_joints_processor = DataProcessorPipeline(
             speed_factor=20.0,
         ),
     ],
-    to_transition=to_transition_teleop_action,
+    to_transition=action_to_transition,
-    to_output=to_output_robot_action,
+    to_output=transition_to_robot_action,
 )
 
 robot.connect()

src/lerobot/processor/converters.py

@@ -233,7 +233,7 @@ def create_transition(
     }
 
 
-def to_transition_teleop_action(action: dict[str, Any]) -> EnvTransition:
+def action_to_transition(action: dict[str, Any]) -> EnvTransition:  # action_to_transition
     """
     Convert a raw teleop action dict into an EnvTransition under the ACTION TransitionKey.
     """
@@ -251,7 +251,7 @@ def to_transition_teleop_action(action: dict[str, Any]) -> EnvTransition:
 
 
 # TODO(Adil, Pepijn): Overtime we can maybe add these converters to pipeline.py itself
-def to_transition_robot_observation(observation: dict[str, Any]) -> EnvTransition:
+def observation_to_transition(observation: dict[str, Any]) -> EnvTransition:
     """
     Convert a raw robot observation dict into an EnvTransition under the OBSERVATION TransitionKey.
     """
@@ -268,7 +268,7 @@ def to_transition_robot_observation(observation: dict[str, Any]) -> EnvTransitio
     return create_transition(observation=obs_dict, action={})
 
 
-def to_output_robot_action(transition: EnvTransition) -> dict[str, Any]:
+def transition_to_robot_action(transition: EnvTransition) -> dict[str, Any]:
     """
     Converts a EnvTransition under the ACTION TransitionKey to a dict with keys ending in '.pos' for raw robot actions.
     """

src/lerobot/record.py

@@ -78,10 +78,10 @@ from lerobot.policies.factory import make_policy, make_pre_post_processors
 from lerobot.policies.pretrained import PreTrainedPolicy
 from lerobot.processor import DataProcessorPipeline, IdentityProcessorStep, TransitionKey
 from lerobot.processor.converters import (
-    to_output_robot_action,
+    action_to_transition,
-    to_transition_robot_observation,
+    observation_to_transition,
-    to_transition_teleop_action,
     transition_to_dataset_frame,
+    transition_to_robot_action,
 )
 from lerobot.processor.rename_processor import rename_stats
 from lerobot.robots import (  # noqa: F401
@@ -245,14 +245,14 @@ def record_loop(
     display_data: bool = False,
 ):
     teleop_action_processor = teleop_action_processor or DataProcessorPipeline(
-        steps=[IdentityProcessorStep()], to_transition=to_transition_teleop_action, to_output=lambda tr: tr
+        steps=[IdentityProcessorStep()], to_transition=action_to_transition, to_output=lambda tr: tr
     )
     robot_action_processor = robot_action_processor or DataProcessorPipeline(
-        steps=[IdentityProcessorStep()], to_transition=lambda tr: tr, to_output=to_output_robot_action
+        steps=[IdentityProcessorStep()], to_transition=lambda tr: tr, to_output=transition_to_robot_action
     )
     robot_observation_processor = robot_observation_processor or DataProcessorPipeline(
         steps=[IdentityProcessorStep()],
-        to_transition=to_transition_robot_observation,
+        to_transition=observation_to_transition,
         to_output=lambda tr: tr,
     )
 

src/lerobot/replay.py

@@ -48,7 +48,7 @@ from pprint import pformat
 from lerobot.configs import parser
 from lerobot.datasets.lerobot_dataset import LeRobotDataset
 from lerobot.processor import DataProcessorPipeline, IdentityProcessorStep
-from lerobot.processor.converters import to_output_robot_action, to_transition_teleop_action
+from lerobot.processor.converters import action_to_transition, transition_to_robot_action
 from lerobot.robots import (  # noqa: F401
     Robot,
     RobotConfig,
@@ -96,8 +96,8 @@ def replay(cfg: ReplayConfig):
     # Initialize robot action processor with default if not provided
     robot_action_processor = cfg.robot_action_processor or DataProcessorPipeline(
         steps=[IdentityProcessorStep()],
-        to_transition=to_transition_teleop_action,
+        to_transition=action_to_transition,
-        to_output=to_output_robot_action,  # type: ignore[arg-type]
+        to_output=transition_to_robot_action,  # type: ignore[arg-type]
     )
 
     # Reset processor

src/lerobot/teleoperate.py

@@ -63,9 +63,9 @@ from lerobot.cameras.realsense.configuration_realsense import RealSenseCameraCon
 from lerobot.configs import parser
 from lerobot.processor import DataProcessorPipeline, IdentityProcessorStep
 from lerobot.processor.converters import (
-    to_output_robot_action,
+    action_to_transition,
-    to_transition_robot_observation,
+    observation_to_transition,
-    to_transition_teleop_action,
+    transition_to_robot_action,
 )
 from lerobot.robots import (  # noqa: F401
     Robot,
@@ -121,16 +121,16 @@ def teleop_loop(
 ):
     # Initialize processors with defaults if not provided
     teleop_action_processor = teleop_action_processor or DataProcessorPipeline(
-        steps=[IdentityProcessorStep()], to_transition=to_transition_teleop_action, to_output=lambda tr: tr
+        steps=[IdentityProcessorStep()], to_transition=action_to_transition, to_output=lambda tr: tr
     )
     robot_action_processor = robot_action_processor or DataProcessorPipeline(
         steps=[IdentityProcessorStep()],
         to_transition=lambda tr: tr,
-        to_output=to_output_robot_action,  # type: ignore[arg-type]
+        to_output=transition_to_robot_action,  # type: ignore[arg-type]
     )
     robot_observation_processor = robot_observation_processor or DataProcessorPipeline(
         steps=[IdentityProcessorStep()],
-        to_transition=to_transition_robot_observation,
+        to_transition=observation_to_transition,
         to_output=lambda tr: tr,
     )
 

tests/processor/test_converters.py

@@ -4,13 +4,13 @@ import torch
 
 from lerobot.processor import TransitionKey
 from lerobot.processor.converters import (
+    action_to_transition,
     batch_to_transition,
-    to_output_robot_action,
+    observation_to_transition,
     to_tensor,
-    to_transition_robot_observation,
-    to_transition_teleop_action,
     transition_to_batch,
     transition_to_dataset_frame,
+    transition_to_robot_action,
 )
 
 
@@ -23,7 +23,7 @@ def test_to_transition_teleop_action_prefix_and_tensor_conversion():
         "raw_img": img,  # uint8 HWC to torch tensor
     }
 
-    tr = to_transition_teleop_action(act)
+    tr = action_to_transition(act)
 
     # Should be an EnvTransition-like dict with ACTION populated
     assert isinstance(tr, dict)
@@ -61,7 +61,7 @@ def test_to_transition_robot_observation_state_vs_images_split():
         "arr": np.array([1.5, 2.5]),  # vector to state to torch tensor
     }
 
-    tr = to_transition_robot_observation(obs)
+    tr = observation_to_transition(obs)
     assert isinstance(tr, dict)
     assert TransitionKey.OBSERVATION in tr
 
@@ -99,7 +99,7 @@ def test_to_output_robot_action_strips_prefix_and_filters_pos_keys_only():
         }
     }
 
-    out = to_output_robot_action(tr)
+    out = transition_to_robot_action(tr)
     # Only ".pos" keys with "action." prefix are retained and stripped to base names
     assert set(out.keys()) == {"j1.pos", "gripper.pos"}
     # Values converted to float