devagent-physical-engine 0.12.0__tar.gz → 0.12.2__tar.gz

{devagent_physical_engine-0.12.0/src/devagent_physical_engine.egg-info → devagent_physical_engine-0.12.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: devagent-physical-engine
-Version: 0.12.0
+Version: 0.12.2
 Summary: Verification-first agentic engineering, simulation, and evidence runtime for robotics and industrial automation.
 Author: Tom Ha
 License: All Rights Reserved
@@ -189,6 +189,10 @@ measured motion/evidence report
 
 On a successful run, Gazebo / MoveIt remains open by default so the engineer can inspect the final workcell. Add `--close-after` to close it after success.
 
+v0.12.1 hardens the initial ROS joint-state snapshot by parsing `ros2 topic echo --once` output as a YAML document stream. This accepts the normal ROS 2 trailing `---` message separator instead of incorrectly classifying an otherwise valid `/joint_states` sample as `joint_state_snapshot_invalid`.
+
+v0.12.2 additionally hardens Gazebo model-pose read-back across Harmonic CLI renderings. One shared parser now accepts both pipe-delimited and whitespace-delimited finite XYZ/RPY vectors and is used by initial world verification and dynamic workpiece pose read-back, so task simulation does not depend on a single human-readable `gz model --pose` delimiter style.
+
 The packaged reference workcell is **not customer-site geometry**. Arbitrary plant layouts require an imported or user-declared physical Twin with trusted poses, collision geometry, tool/TCP information, and workpiece properties. DevAgent does not invent those site facts from a natural-language prompt.
 
 The v0.12 implementation remains evidence-gated: source/CI success does not prove the ROS target campaign has passed on a particular workstation. Use the executable qualification below before treating the task-specific path as qualified on that target.
@@ -416,9 +420,9 @@ Detailed engineering documentation lives in the repository:
 
 ## Project status
 
-**v0.12.0 — Beta software / evidence-gated task-specific physical simulation.**
+**v0.12.2 — Beta software / evidence-gated task-specific physical simulation.**
 
-The software and release pipeline are production-oriented, but physical readiness is scope-specific and evidence-driven. The new UR5e task-specific reference-workcell path must pass the executable v0.12 ROS/Gazebo/MoveIt campaign on the target workstation before that scope is considered simulation-qualified there. Real robot execution remains locked. Functional-safety certification and site qualification are not claimed.
+The software and release pipeline are production-oriented, but physical readiness is scope-specific and evidence-driven. The UR5e task-specific reference-workcell path must pass the executable v0.12 ROS/Gazebo/MoveIt campaign on the target workstation before that scope is considered simulation-qualified there. Real robot execution remains locked. Functional-safety certification and site qualification are not claimed.
 
 ## Ownership
 

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2}/README.md

@@ -148,6 +148,10 @@ measured motion/evidence report
 
 On a successful run, Gazebo / MoveIt remains open by default so the engineer can inspect the final workcell. Add `--close-after` to close it after success.
 
+v0.12.1 hardens the initial ROS joint-state snapshot by parsing `ros2 topic echo --once` output as a YAML document stream. This accepts the normal ROS 2 trailing `---` message separator instead of incorrectly classifying an otherwise valid `/joint_states` sample as `joint_state_snapshot_invalid`.
+
+v0.12.2 additionally hardens Gazebo model-pose read-back across Harmonic CLI renderings. One shared parser now accepts both pipe-delimited and whitespace-delimited finite XYZ/RPY vectors and is used by initial world verification and dynamic workpiece pose read-back, so task simulation does not depend on a single human-readable `gz model --pose` delimiter style.
+
 The packaged reference workcell is **not customer-site geometry**. Arbitrary plant layouts require an imported or user-declared physical Twin with trusted poses, collision geometry, tool/TCP information, and workpiece properties. DevAgent does not invent those site facts from a natural-language prompt.
 
 The v0.12 implementation remains evidence-gated: source/CI success does not prove the ROS target campaign has passed on a particular workstation. Use the executable qualification below before treating the task-specific path as qualified on that target.
@@ -375,9 +379,9 @@ Detailed engineering documentation lives in the repository:
 
 ## Project status
 
-**v0.12.0 — Beta software / evidence-gated task-specific physical simulation.**
+**v0.12.2 — Beta software / evidence-gated task-specific physical simulation.**
 
-The software and release pipeline are production-oriented, but physical readiness is scope-specific and evidence-driven. The new UR5e task-specific reference-workcell path must pass the executable v0.12 ROS/Gazebo/MoveIt campaign on the target workstation before that scope is considered simulation-qualified there. Real robot execution remains locked. Functional-safety certification and site qualification are not claimed.
+The software and release pipeline are production-oriented, but physical readiness is scope-specific and evidence-driven. The UR5e task-specific reference-workcell path must pass the executable v0.12 ROS/Gazebo/MoveIt campaign on the target workstation before that scope is considered simulation-qualified there. Real robot execution remains locked. Functional-safety certification and site qualification are not claimed.
 
 ## Ownership
 

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "devagent-physical-engine"
-version = "0.12.0"
+version = "0.12.2"
 description = "Verification-first agentic engineering, simulation, and evidence runtime for robotics and industrial automation."
 readme = { file = "README.md", content-type = "text/markdown" }
 requires-python = ">=3.11"

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2}/src/devagent_physical_engine/__init__.py

@@ -41,4 +41,4 @@ __all__ = [
     "TwinValidator",
 ]
 
-__version__ = "0.12.0"
+__version__ = "0.12.2"

devagent_physical_engine-0.12.2/src/devagent_physical_engine/ros2/gazebo_cli.py

@@ -0,0 +1,77 @@
+from __future__ import annotations
+
+from math import isfinite
+import re
+
+from ..simulation_platform import SimulationPlatformError
+
+
+_ANSI_ESCAPE = re.compile(r"\x1b\[[0-?]*[ -/]*[@-~]")
+_BRACKET = re.compile(r"\[([^\[\]\r\n]+)\]")
+_POSE_HEADER = re.compile(r"(?im)^\s*[-*]?\s*Pose\s*\[\s*XYZ(?:\s*\(m\))?\s*\].*$")
+
+
+def _vector3(payload: str) -> tuple[float, float, float] | None:
+    """Parse one Gazebo bracket vector across supported CLI renderings.
+
+    Gazebo Sim releases have rendered vectors both as ``[x | y | z]`` and
+    ``[x y z]``. Some downstream packages also emit comma-separated values.
+    Accept those presentation differences while still requiring exactly three
+    finite numeric values.
+    """
+
+    parts = payload.replace("|", " ").replace(",", " ").split()
+    if len(parts) != 3:
+        return None
+    try:
+        values = tuple(float(value) for value in parts)
+    except ValueError:
+        return None
+    if any(not isfinite(value) for value in values):
+        return None
+    return values[0], values[1], values[2]
+
+
+def parse_gz_model_pose(
+    text: str,
+) -> tuple[tuple[float, float, float], tuple[float, float, float]]:
+    """Parse ``gz model -m NAME --pose`` output deterministically.
+
+    The human-readable Gazebo CLI is not stable in its vector delimiter style.
+    Anchor on the model Pose header and parse the first two valid 3-vectors
+    beneath it as XYZ metres and RPY radians. ANSI terminal escapes are ignored.
+    A legacy plain six-number pose line is accepted as a bounded compatibility
+    fallback.
+    """
+
+    cleaned = _ANSI_ESCAPE.sub("", text)
+    header = _POSE_HEADER.search(cleaned)
+    section = cleaned[header.end() :] if header is not None else cleaned
+
+    vectors: list[tuple[float, float, float]] = []
+    for match in _BRACKET.finditer(section):
+        vector = _vector3(match.group(1))
+        if vector is None:
+            continue
+        vectors.append(vector)
+        if len(vectors) == 2:
+            return vectors[0], vectors[1]
+
+    # Compatibility with older Gazebo model tools that rendered
+    # ``x y z roll pitch yaw`` as one line instead of bracket vectors.
+    for line in section.splitlines():
+        parts = line.replace(",", " ").split()
+        if len(parts) != 6:
+            continue
+        try:
+            values = tuple(float(value) for value in parts)
+        except ValueError:
+            continue
+        if any(not isfinite(value) for value in values):
+            continue
+        return (
+            (values[0], values[1], values[2]),
+            (values[3], values[4], values[5]),
+        )
+
+    raise SimulationPlatformError("gazebo_model_pose_output_invalid")

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2}/src/devagent_physical_engine/ros2/gazebo_pose.py

@@ -2,10 +2,10 @@ from __future__ import annotations
 
 from dataclasses import asdict, dataclass
 from math import isfinite
-import re
 
 from ..simulation_platform import SimulationPlatformError
 from .commands import RosCommandRunner
+from .gazebo_cli import parse_gz_model_pose
 
 
 @dataclass(frozen=True, slots=True)
@@ -21,11 +21,11 @@ class GazeboPoseUpdateReceipt:
 
 
 def _parse_position(text: str) -> tuple[float, float, float]:
-    pattern = re.compile(r"\[\s*([-+0-9.eE]+)\s*\|\s*([-+0-9.eE]+)\s*\|\s*([-+0-9.eE]+)\s*\]")
-    match = pattern.search(text)
-    if match is None:
-        raise SimulationPlatformError("gazebo_pose_readback_invalid")
-    return tuple(float(match.group(index)) for index in range(1, 4))
+    try:
+        position, _orientation = parse_gz_model_pose(text)
+    except SimulationPlatformError as exc:
+        raise SimulationPlatformError("gazebo_pose_readback_invalid") from exc
+    return position
 
 
 class GazeboPoseBridge:

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2}/src/devagent_physical_engine/ros2/gazebo_world.py

@@ -8,6 +8,7 @@ from typing import Any
 from ..simulation_platform import SimulationPlatformError
 from ..twin_materialization import CanonicalTwinMaterialization
 from .commands import RosCommandRunner
+from .gazebo_cli import parse_gz_model_pose
 
 
 @dataclass(frozen=True, slots=True)
@@ -48,13 +49,7 @@ def _parse_model_names(text: str) -> set[str]:
 
 
 def _parse_pose(text: str):
-    triples: list[tuple[float, float, float]] = []
-    pattern = re.compile(r"\[\s*([-+0-9.eE]+)\s*\|\s*([-+0-9.eE]+)\s*\|\s*([-+0-9.eE]+)\s*\]")
-    for match in pattern.finditer(text):
-        triples.append(tuple(float(match.group(index)) for index in range(1, 4)))
-    if len(triples) < 2:
-        raise SimulationPlatformError("gazebo_model_pose_output_invalid")
-    return triples[0], triples[1]
+    return parse_gz_model_pose(text)
 
 
 class GazeboWorldVerifier:

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2}/src/devagent_physical_engine/ros2/trajectory.py

@@ -39,6 +39,43 @@ class JointStateSnapshot:
         return tuple(values[name] for name in joint_names)
 
 
+def _parse_joint_state_echo(stdout: str) -> JointStateSnapshot:
+    """Parse ``ros2 topic echo --once`` output as a YAML document stream.
+
+    ROS 2 topic echo normally terminates a message with ``---``.  PyYAML's
+    ``safe_load`` rejects that otherwise-valid output because the separator starts a
+    second (empty) YAML document.  Treat the CLI output as the stream it actually is,
+    ignore empty separator documents, and accept exactly one joint-state mapping.
+    """
+
+    try:
+        import yaml
+
+        documents = tuple(
+            document for document in yaml.safe_load_all(stdout) if document is not None
+        )
+    except Exception as exc:
+        raise RosTrajectoryError("joint_state_yaml_invalid") from exc
+
+    payloads = tuple(document for document in documents if isinstance(document, Mapping))
+    if len(payloads) != 1:
+        raise RosTrajectoryError("joint_state_document_count_invalid")
+
+    payload = payloads[0]
+    names_raw = payload.get("name")
+    positions_raw = payload.get("position")
+    if not isinstance(names_raw, list) or not isinstance(positions_raw, list):
+        raise RosTrajectoryError("joint_state_fields_invalid")
+
+    try:
+        return JointStateSnapshot(
+            tuple(str(item) for item in names_raw),
+            tuple(float(item) for item in positions_raw),
+        )
+    except (TypeError, ValueError) as exc:
+        raise RosTrajectoryError("joint_state_values_invalid") from exc
+
+
 def read_joint_state(
     runner: RosCommandRunner,
     *,
@@ -55,21 +92,9 @@ def read_joint_state(
             + (result.stderr.strip()[-300:] or "ros2_topic_echo_failed")
         )
     try:
-        import yaml
-
-        payload = yaml.safe_load(result.stdout)
-        if not isinstance(payload, Mapping):
-            raise TypeError("joint state payload is not a mapping")
-        names_raw = payload.get("name")
-        positions_raw = payload.get("position")
-        if not isinstance(names_raw, list) or not isinstance(positions_raw, list):
-            raise TypeError("name/position fields are not arrays")
-        return JointStateSnapshot(
-            tuple(str(item) for item in names_raw),
-            tuple(float(item) for item in positions_raw),
-        )
-    except Exception as exc:
-        raise RosTrajectoryError("joint_state_snapshot_invalid") from exc
+        return _parse_joint_state_echo(result.stdout)
+    except RosTrajectoryError as exc:
+        raise RosTrajectoryError(f"joint_state_snapshot_invalid:{exc}") from exc
 
 
 def _duration_payload(seconds: float) -> dict[str, int]:

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2/src/devagent_physical_engine.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: devagent-physical-engine
-Version: 0.12.0
+Version: 0.12.2
 Summary: Verification-first agentic engineering, simulation, and evidence runtime for robotics and industrial automation.
 Author: Tom Ha
 License: All Rights Reserved
@@ -189,6 +189,10 @@ measured motion/evidence report
 
 On a successful run, Gazebo / MoveIt remains open by default so the engineer can inspect the final workcell. Add `--close-after` to close it after success.
 
+v0.12.1 hardens the initial ROS joint-state snapshot by parsing `ros2 topic echo --once` output as a YAML document stream. This accepts the normal ROS 2 trailing `---` message separator instead of incorrectly classifying an otherwise valid `/joint_states` sample as `joint_state_snapshot_invalid`.
+
+v0.12.2 additionally hardens Gazebo model-pose read-back across Harmonic CLI renderings. One shared parser now accepts both pipe-delimited and whitespace-delimited finite XYZ/RPY vectors and is used by initial world verification and dynamic workpiece pose read-back, so task simulation does not depend on a single human-readable `gz model --pose` delimiter style.
+
 The packaged reference workcell is **not customer-site geometry**. Arbitrary plant layouts require an imported or user-declared physical Twin with trusted poses, collision geometry, tool/TCP information, and workpiece properties. DevAgent does not invent those site facts from a natural-language prompt.
 
 The v0.12 implementation remains evidence-gated: source/CI success does not prove the ROS target campaign has passed on a particular workstation. Use the executable qualification below before treating the task-specific path as qualified on that target.
@@ -416,9 +420,9 @@ Detailed engineering documentation lives in the repository:
 
 ## Project status
 
-**v0.12.0 — Beta software / evidence-gated task-specific physical simulation.**
+**v0.12.2 — Beta software / evidence-gated task-specific physical simulation.**
 
-The software and release pipeline are production-oriented, but physical readiness is scope-specific and evidence-driven. The new UR5e task-specific reference-workcell path must pass the executable v0.12 ROS/Gazebo/MoveIt campaign on the target workstation before that scope is considered simulation-qualified there. Real robot execution remains locked. Functional-safety certification and site qualification are not claimed.
+The software and release pipeline are production-oriented, but physical readiness is scope-specific and evidence-driven. The UR5e task-specific reference-workcell path must pass the executable v0.12 ROS/Gazebo/MoveIt campaign on the target workstation before that scope is considered simulation-qualified there. Real robot execution remains locked. Functional-safety certification and site qualification are not claimed.
 
 ## Ownership
 

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2}/src/devagent_physical_engine.egg-info/SOURCES.txt

@@ -78,6 +78,7 @@ src/devagent_physical_engine/ros2/doctor.py
 src/devagent_physical_engine/ros2/fk_pose.py
 src/devagent_physical_engine/ros2/fk_probe.py
 src/devagent_physical_engine/ros2/frame_alignment.py
+src/devagent_physical_engine/ros2/gazebo_cli.py
 src/devagent_physical_engine/ros2/gazebo_pose.py
 src/devagent_physical_engine/ros2/gazebo_world.py
 src/devagent_physical_engine/ros2/joint_state_recorder.py
@@ -112,6 +113,7 @@ tests/test_ai_cli_v07.py
 tests/test_continuous_clearance_v011.py
 tests/test_core.py
 tests/test_engineering_request_v07.py
+tests/test_gazebo_cli_v0122.py
 tests/test_interpreter_profiles_v08.py
 tests/test_live_provider_qualification.py
 tests/test_measured_motion_v10.py

devagent_physical_engine-0.12.2/tests/test_gazebo_cli_v0122.py

@@ -0,0 +1,88 @@
+from __future__ import annotations
+
+import unittest
+
+from devagent_physical_engine.ros2.gazebo_cli import parse_gz_model_pose
+from devagent_physical_engine.ros2.gazebo_pose import _parse_position
+from devagent_physical_engine.ros2.gazebo_world import _parse_pose
+from devagent_physical_engine.simulation_platform import SimulationPlatformError
+
+
+PIPE_OUTPUT = """Requesting state for world [default]...
+
+Model: [8]
+
+- Name: BOX_101
+
+- Pose [ XYZ (m) ] [ RPY (rad) ]:
+
+[0.550000 | -0.350000 | 0.820000]
+
+[0.000000 | 0.000000 | 1.570796]
+"""
+
+HARMONIC_WHITESPACE_OUTPUT = """Requesting state for world [default]...
+
+* Model [8]
+  * Name: BOX_101
+  * Pose [ XYZ (m) ] [ RPY (rad) ]:
+[0.550000 -0.350000 0.820000]
+[0.000000 -0.000000 1.570796]
+"""
+
+
+class GazeboCliPoseV0122Tests(unittest.TestCase):
+    def test_accepts_documented_pipe_delimited_pose(self):
+        position, orientation = parse_gz_model_pose(PIPE_OUTPUT)
+        self.assertEqual(position, (0.55, -0.35, 0.82))
+        self.assertEqual(orientation, (0.0, 0.0, 1.570796))
+
+    def test_accepts_harmonic_whitespace_delimited_pose(self):
+        position, orientation = parse_gz_model_pose(HARMONIC_WHITESPACE_OUTPUT)
+        self.assertEqual(position, (0.55, -0.35, 0.82))
+        self.assertEqual(orientation, (0.0, -0.0, 1.570796))
+
+    def test_world_verifier_parser_uses_harmonic_compatible_parser(self):
+        self.assertEqual(
+            _parse_pose(HARMONIC_WHITESPACE_OUTPUT),
+            ((0.55, -0.35, 0.82), (0.0, -0.0, 1.570796)),
+        )
+
+    def test_dynamic_workpiece_readback_uses_same_parser(self):
+        self.assertEqual(
+            _parse_position(HARMONIC_WHITESPACE_OUTPUT),
+            (0.55, -0.35, 0.82),
+        )
+
+    def test_ansi_terminal_escape_does_not_break_pose(self):
+        payload = "\x1b[32m" + HARMONIC_WHITESPACE_OUTPUT + "\x1b[0m"
+        self.assertEqual(parse_gz_model_pose(payload)[0], (0.55, -0.35, 0.82))
+
+    def test_legacy_plain_six_number_pose_is_supported(self):
+        self.assertEqual(
+            parse_gz_model_pose("0.55 -0.35 0.82 0.0 0.0 1.570796\n"),
+            ((0.55, -0.35, 0.82), (0.0, 0.0, 1.570796)),
+        )
+
+    def test_incomplete_pose_fails_closed(self):
+        with self.assertRaisesRegex(
+            SimulationPlatformError,
+            "gazebo_model_pose_output_invalid",
+        ):
+            parse_gz_model_pose(
+                "- Pose [ XYZ (m) ] [ RPY (rad) ]:\n[0.55 -0.35 0.82]\n"
+            )
+
+    def test_non_finite_pose_fails_closed(self):
+        with self.assertRaisesRegex(
+            SimulationPlatformError,
+            "gazebo_model_pose_output_invalid",
+        ):
+            parse_gz_model_pose(
+                "- Pose [ XYZ (m) ] [ RPY (rad) ]:\n"
+                "[0.55 -0.35 nan]\n[0.0 0.0 0.0]\n"
+            )
+
+
+if __name__ == "__main__":
+    unittest.main()

{devagent_physical_engine-0.12.0 → devagent_physical_engine-0.12.2}/tests/test_ros_trajectory_v09.py

@@ -4,7 +4,11 @@ import unittest
 
 from devagent_physical_engine.physical_motion import JointTrajectoryPoint, PhysicalMotionPlan
 from devagent_physical_engine.ros2.commands import CommandResult
-from devagent_physical_engine.ros2.trajectory import RosTrajectoryError, execute_joint_trajectory
+from devagent_physical_engine.ros2.trajectory import (
+    RosTrajectoryError,
+    execute_joint_trajectory,
+    read_joint_state,
+)
 
 
 class FakeRunner:
@@ -29,11 +33,64 @@ def motion():
     )
 
 
-def result(stdout, *, code=0):
-    return CommandResult(("ros2",), code, stdout, "", 1.0)
+def result(stdout, *, code=0, stderr=""):
+    return CommandResult(("ros2",), code, stdout, stderr, 1.0)
 
 
 class RosTrajectoryV09Tests(unittest.TestCase):
+    def test_joint_state_echo_accepts_ros2_trailing_document_separator(self):
+        stdout = """header:
+  stamp:
+    sec: 10
+    nanosec: 20
+  frame_id: ''
+name:
+- shoulder_pan_joint
+- shoulder_lift_joint
+position:
+- 0.125
+- -1.25
+velocity: []
+effort: []
+---
+"""
+        runner = FakeRunner(result(stdout))
+
+        snapshot = read_joint_state(runner)
+
+        self.assertEqual(
+            snapshot.names,
+            ("shoulder_pan_joint", "shoulder_lift_joint"),
+        )
+        self.assertEqual(snapshot.positions_rad, (0.125, -1.25))
+        self.assertEqual(
+            runner.commands[0][0],
+            ("ros2", "topic", "echo", "/joint_states", "--once"),
+        )
+
+    def test_joint_state_echo_rejects_multiple_nonempty_documents(self):
+        stdout = """name: [j1]
+position: [0.0]
+---
+name: [j1]
+position: [0.1]
+---
+"""
+        runner = FakeRunner(result(stdout))
+        with self.assertRaisesRegex(
+            RosTrajectoryError,
+            "joint_state_snapshot_invalid:joint_state_document_count_invalid",
+        ):
+            read_joint_state(runner)
+
+    def test_joint_state_echo_preserves_command_failure_detail(self):
+        runner = FakeRunner(result("", code=1, stderr="topic unavailable"))
+        with self.assertRaisesRegex(
+            RosTrajectoryError,
+            "joint_state_snapshot_failed:topic unavailable",
+        ):
+            read_joint_state(runner)
+
     def test_succeeded_action_is_accepted(self):
         runner = FakeRunner(result("Goal accepted\nResult: error_code=0\nGoal finished with status: SUCCEEDED"))
         completed = execute_joint_trajectory(runner, motion())