feat(imitation): integrate native OpenYAM Quest collection

dimos/imitation/collection/README.md

@@ -6,7 +6,7 @@ controls through ordinary `autoconnect()` and `remappings()`.
 
 ## Declare the recorded features
 
-This example declares measured joint state, accepted joint targets, and one RGB
+This example declares measured joint state, commanded joint targets, and one RGB
 image. Constructing the profile does not connect to hardware.
 
 ```python session=profile no-result
@@ -29,13 +29,21 @@ profile = CollectionProfile(
             names=["height", "width", "channels"],
         ),
     },
-    actions={"action": state.model_copy(update={
-        "stream": "accepted", "source_kind": "joint_position_updates",
-    }, deep=True)},
+    actions={
+        "action": state.model_copy(
+            update={
+                "stream": "commanded",
+                "source_kind": "joint_position_updates",
+            },
+            deep=True,
+        )
+    },
     metadata={"name": "custom-arm", "robot_type": "custom-arm-2dof"},
 )
 assert profile.input_types() == {
-    "measured": JointState, "wrist_image": Image, "accepted": JointState,
+    "measured": JointState,
+    "wrist_image": Image,
+    "commanded": JointState,
 }
 ```
 
@@ -66,7 +74,7 @@ recorder = collect(profile=profile, recording=Path("recordings/session-001"))
 Include `recorder`, your robot, camera producers, and an
 `EpisodeMonitorModule.blueprint(instance_name="episodes")` in `autoconnect()`.
 The monitor supplies `status: In[String]` containing versioned JSON episode events.
-Robot-specific readiness rules belong on that monitor. The recorder must precede
+The monitor only controls episode start, save, and discard. The recorder must precede
 producers in the Blueprint so it stops after them.
 
 The Python graph supports any number of coordinators. Connect each recorded stream
@@ -131,8 +139,8 @@ add an explicit depth feature and connect its producer.
 Capture stores asynchronous streams at their native rates. The saved schema fixes
 the alignment to a 30 Hz grid, anchored at the first video observation, or the first
 observation when there is no video. Snapshot matching uses a 20 ms tolerance.
-Accepted joint updates use causal hold: only commands at or before the target time
-contribute, and omitted joints retain their last accepted target. Missing accepted
+Commanded joint updates use causal hold: only commands at or before the target time
+contribute, and omitted joints retain their last commanded target. Missing commanded
 targets never fall back to measured state or future commands.
 
 The collection API exposes no `sync` or `quality` options. Existing low-level

dimos/imitation/dataprep/README.md

@@ -85,7 +85,7 @@ config = DataPrepConfig(
         },
         action={
             "action": FeatureSpec(
-                stream="applied_joint_position_command",
+                stream="coordinator_joint_command",
                 field="position",
                 dtype="float32",
                 shape=(2,),

dimos/imitation/dataprep/test_mcap_source.py

@@ -82,8 +82,8 @@ def _write_collection(path: Path) -> None:
             "coordinator_joint_state": _register_channel(
                 writer, "coordinator_joint_state", JointState
             ),
-            "applied_joint_position_command": _register_channel(
-                writer, "applied_joint_position_command", JointState
+            "coordinator_joint_command": _register_channel(
+                writer, "coordinator_joint_command", JointState
             ),
             "status": _register_channel(writer, "status", String, "json"),
         }
@@ -130,7 +130,7 @@ def _write_collection(path: Path) -> None:
             )
             _write_message(
                 writer,
-                channels["applied_joint_position_command"],
+                channels["coordinator_joint_command"],
                 ts,
                 JointState(
                     ts=ts,
@@ -179,7 +179,7 @@ def _config(source: Path, output: Path) -> DataPrepConfig:
             },
             action={
                 "action": FeatureSpec(
-                    stream="applied_joint_position_command",
+                    stream="coordinator_joint_command",
                     field="position",
                     dtype="float32",
                     shape=(2,),
@@ -213,7 +213,7 @@ def writer(samples: Iterator[Sample], selected_output: OutputConfig) -> Path:
     dataset_path = run_dataprep(config, writer=writer)
 
     assert info["streams"] == {
-        "applied_joint_position_command": 3,
+        "coordinator_joint_command": 3,
         "color_image": 3,
         "coordinator_joint_state": 3,
         "status": 2,

dimos/robot/all_blueprints.py

@@ -109,6 +109,7 @@
     "mid360-realsense-record-with-pcap": "dimos.robot.assembly.mid360_realsense_30:mid360_realsense_record_with_pcap",
     "openarm-planner-coordinator": "dimos.robot.manipulators.openarm.blueprints.basic:openarm_planner_coordinator",
     "openyam-planner-coordinator": "dimos.robot.manipulators.openyam.blueprints.basic:openyam_planner_coordinator",
+    "openyam-quest-collection": "dimos.robot.manipulators.openyam.blueprints.learning_quest_collection:openyam_quest_collection",
     "r1pro-coordinator": "dimos.robot.galaxea.r1pro.blueprints.basic.r1pro_coordinator:r1pro_coordinator",
     "r1pro-manipulation": "dimos.robot.galaxea.r1pro.blueprints.manipulation.r1pro_manipulation:r1pro_manipulation",
     "r1pro-nav": "dimos.robot.galaxea.r1pro.blueprints.navigation.r1pro_nav:r1pro_nav",

dimos/robot/manipulators/openyam/blueprints/learning_quest_collection.py

@@ -0,0 +1,76 @@
+# Copyright 2026 Dimensional Inc.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+"""Quest-guided OpenYAM collection."""
+
+from dimos.core.coordination.blueprints import autoconnect
+from dimos.hardware.sensors.camera.module import WebcamModule
+from dimos.hardware.sensors.camera.webcam import WebcamConfig
+from dimos.imitation.collection.episode_monitor import EpisodeMonitorModule
+from dimos.imitation.collection.profile import CollectionProfile
+from dimos.imitation.collection.recorder import collect
+from dimos.imitation.dataprep.schema import FeatureSpec
+from dimos.msgs.sensor_msgs.Image import Image
+from dimos.msgs.sensor_msgs.JointState import JointState
+from dimos.robot.manipulators.openyam.blueprints.teleop import teleop_webxr_openyam
+from dimos.robot.manipulators.openyam.joints import OPENYAM_JOINTS
+
+_state = FeatureSpec(
+    stream="coordinator_joint_state",
+    message_type=JointState,
+    field="position",
+    dtype="float32",
+    shape=(len(OPENYAM_JOINTS),),
+    names=list(OPENYAM_JOINTS),
+)
+_profile = CollectionProfile(
+    observations={
+        "observation.images.wrist": FeatureSpec(
+            stream="wrist_image",
+            message_type=Image,
+            field="data",
+            dtype="video",
+            shape=(480, 640, 3),
+            names=["height", "width", "channels"],
+        ),
+        "observation.state": _state,
+    },
+    actions={
+        "action": _state.model_copy(
+            update={
+                "stream": "coordinator_joint_command",
+                "source_kind": "joint_position_updates",
+            },
+            deep=True,
+        )
+    },
+    metadata={"name": "openyam-quest", "robot_type": "openyam"},
+)
+
+openyam_quest_collection = autoconnect(
+    collect(profile=_profile),
+    EpisodeMonitorModule.blueprint(instance_name="episodes"),
+    teleop_webxr_openyam,
+    WebcamModule.blueprint(
+        instance_name="wrist",
+        hardware=WebcamConfig(camera_index=0, fps=30, frame_id_prefix="wrist_image"),
+        frame_id="wrist_camera_link",
+    ),
+).remappings(
+    [
+        ("wrist", "color_image", "wrist_image"),
+        ("wrist", "camera_info", "wrist_camera_info"),
+        ("wrist", "tf", "wrist_tf"),
+    ]
+)