devagent-physical-engine 2.0.0__tar.gz → 2.1.0__tar.gz

{devagent_physical_engine-2.0.0/src/devagent_physical_engine.egg-info → devagent_physical_engine-2.1.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: devagent-physical-engine
-Version: 2.0.0
+Version: 2.1.0
 Summary: Verification-first robotics commissioning control plane with immutable engineering findings, deterministic pre-site readiness, evidence authority, exact-Twin lineage, selective simulation, regression, replay, and FAT reporting.
 Author: Tom Ha
 License: All Rights Reserved
@@ -81,7 +81,7 @@ DevAgent turns customer engineering inputs into a bounded, evidence-driven pre-s
 
 <p align="center"><sub><b>Real engine screenshots — not product renders.</b> ROS 2 Jazzy · Gazebo Harmonic · MoveIt 2 · Pilz · UR5e exact-Twin verification.</sub></p>
 
-## v2.0 commissioning verification flow
+## v2.1 commissioning verification flow
 
 ```text
 Customer engineering files
@@ -125,7 +125,7 @@ That decision means the declared pre-site engineering scope has complete, curren
 Python 3.11+ is required.
 
 ```bash
-python -m pip install --upgrade devagent-physical-engine==2.0.0
+python -m pip install --upgrade devagent-physical-engine==2.1.0
 ```
 
 Verify the installed package:
@@ -137,7 +137,7 @@ python -c "import devagent_physical_engine as d; print(d.__version__)"
 Expected:
 
 ```text
-2.0.0
+2.1.0
 ```
 
 Run the normal customer workflow:
@@ -235,7 +235,7 @@ A generated plan is not physical proof. A replay is not a new execution. An AI r
 
 ## Engineering findings and risk authority
 
-v2.0 records first-class immutable artifacts for:
+The v2 commissioning line records first-class immutable artifacts for:
 
 ```text
 engineering_finding
@@ -377,14 +377,15 @@ The target-proven UR5e path retains controller-manager/action readiness binding,
 
 Unsupported customer Twins, unqualified robot profiles, boundary/fault cases outside the qualified scope, ambiguous case mappings, or incomplete lineage fail closed.
 
-ABB IRB1200 remains a separate experimental visual simulation path. FANUC/KUKA and other robot/vendor paths retain their truthful support tiers until separately qualified.
+Exact ABB IRB1200, FANUC CRX-10iA, and KUKA KR6 R900-2 simulation qualification infrastructure is included in v2.1.0. Their packaged simulation tiers remain `EXPERIMENTAL` until accepted target-workstation reports are reviewed and a separate authority-promotion change is merged. Generic ABB/FANUC/KUKA family profiles remain `MODEL_ONLY`.
 
 Inspect robot support:
 
 ```bash
 devagent robots
 devagent robot doctor abb_irb1200
-devagent robot visualize abb_irb1200
+devagent robot doctor fanuc_crx10ia
+devagent robot doctor kuka_kr6_r900_2
 ```
 
 ## Customer input
@@ -520,9 +521,9 @@ devagent-physical-ai optional provider-backed engineering front end
 
 # Project status
 
-**v2.0.0 — Production/Stable software workflow; commissioning verification control plane with fail-closed pre-site readiness**
+**v2.1.0 — Production/Stable software workflow; commissioning verification control plane with fail-closed multivendor simulation qualification**
 
-v2.0.0 adds immutable engineering findings/dispositions, deterministic verification-scope authority, stronger internal/external evidence authority, commissioning-aware Evidence Graph lineage, whole-project staleness detection, and an immutable pre-site commissioning readiness decision.
+v2.1.0 preserves the v2.0 commissioning-verification authority model and adds exact-model ABB/FANUC/KUKA qualification infrastructure with direct runtime readiness probes, verified simulation-clock and robot-description contracts, authoritative joint-state measurement, controller-feedback tracking, immutable qualification reports, software-provenance hashing, and a three-report multivendor acceptance gate. Exact ABB/FANUC/KUKA simulation tiers are promoted only after accepted target-workstation evidence and a code-reviewed authority change.
 
 Production-oriented software capabilities include:
 
@@ -553,7 +554,7 @@ immutable READY / NOT_READY pre-site commissioning decision
 provider-neutral optional AI front end
 ```
 
-Production/Stable describes the bounded software workflow. The target-proven UR5e claim remains limited to the qualified packaged scope and acceptance environment. v2.0.0 does not claim arbitrary customer cells are physically qualified, site-qualified, functionally safe, commissioned, or authorized for autonomous real execution.
+Production/Stable describes the bounded software workflow. The target-proven UR5e claim remains limited to the qualified packaged scope and acceptance environment. v2.1.0 does not claim arbitrary customer cells are physically qualified, site-qualified, functionally safe, commissioned, or authorized for autonomous real execution.
 
 Current hard authority boundaries remain:
 
@@ -568,6 +569,9 @@ unless separately proven by an appropriate qualified process.
 
 ## Documentation
 
+v2.1.0 release notes:
+https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V2_1_0.md
+
 v2.0.0 release notes:
 https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V2_0_0.md
 

{devagent_physical_engine-2.0.0 → devagent_physical_engine-2.1.0}/README.md

@@ -39,7 +39,7 @@ DevAgent turns customer engineering inputs into a bounded, evidence-driven pre-s
 
 <p align="center"><sub><b>Real engine screenshots — not product renders.</b> ROS 2 Jazzy · Gazebo Harmonic · MoveIt 2 · Pilz · UR5e exact-Twin verification.</sub></p>
 
-## v2.0 commissioning verification flow
+## v2.1 commissioning verification flow
 
 ```text
 Customer engineering files
@@ -83,7 +83,7 @@ That decision means the declared pre-site engineering scope has complete, curren
 Python 3.11+ is required.
 
 ```bash
-python -m pip install --upgrade devagent-physical-engine==2.0.0
+python -m pip install --upgrade devagent-physical-engine==2.1.0
 ```
 
 Verify the installed package:
@@ -95,7 +95,7 @@ python -c "import devagent_physical_engine as d; print(d.__version__)"
 Expected:
 
 ```text
-2.0.0
+2.1.0
 ```
 
 Run the normal customer workflow:
@@ -193,7 +193,7 @@ A generated plan is not physical proof. A replay is not a new execution. An AI r
 
 ## Engineering findings and risk authority
 
-v2.0 records first-class immutable artifacts for:
+The v2 commissioning line records first-class immutable artifacts for:
 
 ```text
 engineering_finding
@@ -335,14 +335,15 @@ The target-proven UR5e path retains controller-manager/action readiness binding,
 
 Unsupported customer Twins, unqualified robot profiles, boundary/fault cases outside the qualified scope, ambiguous case mappings, or incomplete lineage fail closed.
 
-ABB IRB1200 remains a separate experimental visual simulation path. FANUC/KUKA and other robot/vendor paths retain their truthful support tiers until separately qualified.
+Exact ABB IRB1200, FANUC CRX-10iA, and KUKA KR6 R900-2 simulation qualification infrastructure is included in v2.1.0. Their packaged simulation tiers remain `EXPERIMENTAL` until accepted target-workstation reports are reviewed and a separate authority-promotion change is merged. Generic ABB/FANUC/KUKA family profiles remain `MODEL_ONLY`.
 
 Inspect robot support:
 
 ```bash
 devagent robots
 devagent robot doctor abb_irb1200
-devagent robot visualize abb_irb1200
+devagent robot doctor fanuc_crx10ia
+devagent robot doctor kuka_kr6_r900_2
 ```
 
 ## Customer input
@@ -478,9 +479,9 @@ devagent-physical-ai optional provider-backed engineering front end
 
 # Project status
 
-**v2.0.0 — Production/Stable software workflow; commissioning verification control plane with fail-closed pre-site readiness**
+**v2.1.0 — Production/Stable software workflow; commissioning verification control plane with fail-closed multivendor simulation qualification**
 
-v2.0.0 adds immutable engineering findings/dispositions, deterministic verification-scope authority, stronger internal/external evidence authority, commissioning-aware Evidence Graph lineage, whole-project staleness detection, and an immutable pre-site commissioning readiness decision.
+v2.1.0 preserves the v2.0 commissioning-verification authority model and adds exact-model ABB/FANUC/KUKA qualification infrastructure with direct runtime readiness probes, verified simulation-clock and robot-description contracts, authoritative joint-state measurement, controller-feedback tracking, immutable qualification reports, software-provenance hashing, and a three-report multivendor acceptance gate. Exact ABB/FANUC/KUKA simulation tiers are promoted only after accepted target-workstation evidence and a code-reviewed authority change.
 
 Production-oriented software capabilities include:
 
@@ -511,7 +512,7 @@ immutable READY / NOT_READY pre-site commissioning decision
 provider-neutral optional AI front end
 ```
 
-Production/Stable describes the bounded software workflow. The target-proven UR5e claim remains limited to the qualified packaged scope and acceptance environment. v2.0.0 does not claim arbitrary customer cells are physically qualified, site-qualified, functionally safe, commissioned, or authorized for autonomous real execution.
+Production/Stable describes the bounded software workflow. The target-proven UR5e claim remains limited to the qualified packaged scope and acceptance environment. v2.1.0 does not claim arbitrary customer cells are physically qualified, site-qualified, functionally safe, commissioned, or authorized for autonomous real execution.
 
 Current hard authority boundaries remain:
 
@@ -526,6 +527,9 @@ unless separately proven by an appropriate qualified process.
 
 ## Documentation
 
+v2.1.0 release notes:
+https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V2_1_0.md
+
 v2.0.0 release notes:
 https://github.com/tomha85/devagent-physical-engine/blob/main/docs/RELEASE_NOTES_V2_0_0.md
 

{devagent_physical_engine-2.0.0 → devagent_physical_engine-2.1.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "devagent-physical-engine"
-version = "2.0.0"
+version = "2.1.0"
 description = "Verification-first robotics commissioning control plane with immutable engineering findings, deterministic pre-site readiness, evidence authority, exact-Twin lineage, selective simulation, regression, replay, and FAT reporting."
 readme = { file = "README.md", content-type = "text/markdown" }
 requires-python = ">=3.11"
@@ -89,9 +89,14 @@ omit = [
   "src/devagent_physical_engine/ros2/workpiece_gazebo_follower.py",
   "src/devagent_physical_engine/ros2/abb_irb1200.py",
   "src/devagent_physical_engine/ros2/abb_irb1200_preview.py",
+  "src/devagent_physical_engine/ros2/fanuc_crx10ia.py",
+  "src/devagent_physical_engine/ros2/fanuc_crx10ia_physics.py",
+  "src/devagent_physical_engine/ros2/kuka_kr6_r900_2.py",
+  "src/devagent_physical_engine/ros2/industrial_arm_task_runtime.py",
+  "src/devagent_physical_engine/ros2/industrial_arm_task_adapters.py",
 ]
 
 [tool.coverage.report]
 fail_under = 75
 show_missing = true
-skip_covered = false
+skip_covered = false

{devagent_physical_engine-2.0.0 → devagent_physical_engine-2.1.0}/src/devagent_physical_engine/__init__.py

@@ -249,4 +249,4 @@ install_industrial_motion_v170()
 install_controller_runtime_hotfix_v181()
 install_one_command_runtime()
 
-__version__ = "2.0.0"
+__version__ = "2.1.0"

{devagent_physical_engine-2.0.0 → devagent_physical_engine-2.1.0}/src/devagent_physical_engine/controller_runtime_hotfix_v181.py

@@ -6,11 +6,13 @@ from typing import Mapping
 
 from .ros2 import trajectory as _trajectory
 from .ros2.commands import CommandExecutionError, RosCommandRunner
+from .ros2.direct_controller_discovery_runner import DirectControllerDiscoveryRunner
 from .ros2.trajectory import RosTrajectoryError
 from .ros2.ur5e_adapter import UR5E_JOINTS, UR5eGazeboAdapter
 
 
 _DEFAULT_TRAJECTORY_ACTION = "/joint_trajectory_controller/follow_joint_trajectory"
+_ORIGINAL_RESOLVE_ACTIVE_TRAJECTORY_ACTION = _trajectory.resolve_active_trajectory_action
 
 
 def _last_json_mapping(stdout: str, *, code: str) -> Mapping[str, object]:
@@ -36,6 +38,23 @@ def resolve_active_trajectory_action_v181(
     ):
         raise RosTrajectoryError("trajectory_action_must_be_absolute")
 
+    # Exact multivendor qualification adapters wrap their runner with
+    # DirectControllerDiscoveryRunner. Keep execution-time controller selection
+    # on the same direct-rclpy authority path already used by readiness instead
+    # of bypassing it through the legacy one-shot v1.8.1 helper.
+    #
+    # The original resolver asks the runner for ros2 control list_controllers.
+    # DirectControllerDiscoveryRunner intercepts exactly that read-only command
+    # and synthesizes its result from controller_manager/ListControllers through
+    # direct rclpy discovery. No ros2cli daemon is involved.
+    if isinstance(runner, DirectControllerDiscoveryRunner):
+        return _ORIGINAL_RESOLVE_ACTIVE_TRAJECTORY_ACTION(
+            runner,
+            preferred_action=preferred_action,
+        )
+
+    # Preserve the existing direct machine probe for mature runtimes such as
+    # UR5e that are not wrapped by the multivendor controller-discovery runner.
     service_timeout = 3.0
     action_timeout = 3.0
     helper = Path(_trajectory.__file__).with_name("controller_action_probe.py").resolve()

devagent_physical_engine-2.1.0/src/devagent_physical_engine/robot_exercise_cli.py

@@ -0,0 +1,167 @@
+from __future__ import annotations
+
+import argparse
+import json
+from pathlib import Path
+import sys
+
+from .robots import PROFILE_REGISTRY
+from .ros2.commands import CommandExecutionError, SubprocessRosRunner
+from .ros2.industrial_arm_task_adapters import TASK_ADAPTERS, TASK_SPECS
+from .ros2.industrial_arm_task_runtime import run_commissioning_exercise
+from .simulation_platform import SimulationPlatformError
+
+
+def _parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(
+        prog="devagent robot exercise",
+        description=(
+            "Run the bounded MoveIt/controller commissioning exercise for an exact "
+            "experimental industrial-robot runtime. This never promotes FAT evidence."
+        ),
+    )
+    parser.add_argument("robot")
+    parser.add_argument("--ros-setup", default="/opt/ros/jazzy/setup.bash")
+    parser.add_argument("--workspace-setup")
+    parser.add_argument("--log-dir", default="~/.devagent/robot-exercise")
+    parser.add_argument("--startup-timeout", type=float, default=120.0)
+    parser.add_argument(
+        "--no-replay",
+        action="store_true",
+        help="skip the deterministic replay leg (not recommended for qualification)",
+    )
+    parser.add_argument("--json", action="store_true")
+    parser.add_argument(
+        "--dry-run",
+        action="store_true",
+        help="show the exact bounded exercise contract without starting ROS",
+    )
+    return parser
+
+
+def _print_dry_run(profile_key: str, *, json_output: bool) -> int:
+    spec = TASK_SPECS[profile_key]
+    payload = {
+        "profile_key": spec.profile_key,
+        "joint_names": list(spec.joint_names),
+        "planning_group": spec.planning_group,
+        "planning_frame": spec.planning_frame,
+        "fk_link_name": spec.fk_link_name,
+        "home_positions_rad": list(spec.home_positions_rad),
+        "exercise_positions_rad": list(spec.exercise_positions_rad),
+        "runtime_kind": spec.runtime_kind,
+        "physics_simulation": spec.physics_simulation,
+        "moveit_preplan_required": True,
+        "follow_joint_trajectory_required": True,
+        "joint_state_measurement_required": True,
+        "controller_feedback_verification_required": True,
+        "fk_metric_collection_required": True,
+        "replay_supported": True,
+        "fat_evidence_qualified": False,
+        "real_execution_allowed": False,
+    }
+    if json_output:
+        print(json.dumps(payload, indent=2, sort_keys=True))
+    else:
+        print(f"Robot exercise dry-run: {spec.profile_key}")
+        print(f"Runtime: {spec.runtime_kind}")
+        print(f"Physics simulation: {spec.physics_simulation}")
+        print("MoveIt preplan: REQUIRED")
+        print("FollowJointTrajectory execution: REQUIRED")
+        print("Controller feedback verification: REQUIRED")
+        print("Persistent /joint_states measurement: REQUIRED")
+        print("MoveIt FK endpoint/path metrics: REQUIRED")
+        print("Replay: SUPPORTED")
+        print("FAT evidence qualified: False")
+        print("Real execution allowed: False")
+        print("Home: " + ", ".join(f"{value:.4f}" for value in spec.home_positions_rad))
+        print(
+            "Exercise: "
+            + ", ".join(f"{value:.4f}" for value in spec.exercise_positions_rad)
+        )
+    return 0
+
+
+def main(argv: list[str] | None = None) -> int:
+    args = _parser().parse_args(argv)
+    try:
+        profile = PROFILE_REGISTRY.resolve(args.robot)
+    except Exception as exc:
+        print(f"Robot exercise blocked: {exc}", file=sys.stderr)
+        return 41
+
+    if profile.key not in TASK_ADAPTERS:
+        print(
+            f"Robot exercise blocked: no experimental task adapter registered for {profile.key}",
+            file=sys.stderr,
+        )
+        return 41
+    if args.startup_timeout <= 0:
+        print("Robot exercise blocked: startup timeout must be positive", file=sys.stderr)
+        return 2
+    if args.dry_run:
+        return _print_dry_run(profile.key, json_output=args.json)
+
+    workspace = Path(args.workspace_setup).expanduser() if args.workspace_setup else None
+    runner = SubprocessRosRunner(
+        ros_setup=Path(args.ros_setup).expanduser(),
+        workspace_setup=workspace,
+    )
+    adapter_type = TASK_ADAPTERS[profile.key]
+    log_root = Path(args.log_dir).expanduser() / profile.key
+    adapter = adapter_type(
+        runner,
+        ros_setup=args.ros_setup,
+        log_dir=log_root,
+        startup_timeout_s=args.startup_timeout,
+    )
+    if not args.json:
+        print(
+            f"[DevAgent] START robot={profile.key} runtime={TASK_SPECS[profile.key].runtime_kind}",
+            flush=True,
+        )
+        print(f"[DevAgent] LOGS  {log_root.resolve()}", flush=True)
+        print(
+            "[DevAgent] PREPARE launching exact simulator/controller/MoveIt stack; "
+            "no FAT evidence can be promoted by this command",
+            flush=True,
+        )
+    try:
+        report = run_commissioning_exercise(
+            adapter,
+            work_dir=log_root / "exercise",
+            replay=not args.no_replay,
+        )
+    except (
+        CommandExecutionError,
+        SimulationPlatformError,
+        OSError,
+        ValueError,
+    ) as exc:
+        print(f"Robot exercise blocked: {exc}", file=sys.stderr)
+        print(f"Diagnostic logs: {log_root.resolve()}", file=sys.stderr)
+        print("No FAT evidence was promoted and real robot execution remains locked.")
+        return 41
+
+    if args.json:
+        print(json.dumps(report.to_dict(), indent=2, sort_keys=True))
+    else:
+        print(f"Robot: {report.profile_key}")
+        print(f"Runtime: {report.runtime_kind}")
+        print(f"Physics simulation: {report.physics_simulation}")
+        print("Phases: " + ", ".join(report.phases))
+        print(f"Replay verified: {report.replay_verified}")
+        print(f"Result: {'PASS' if report.passed else 'FAIL'}")
+        print("FAT evidence qualified: False")
+        print("Real execution allowed: False")
+        for index, metrics in enumerate(report.motion_metrics, start=1):
+            print(
+                f"Motion {index}: duration={metrics.get('observed_duration_s')} "
+                f"tracking={metrics.get('max_tracking_error_rad')} "
+                f"tcp_error={metrics.get('final_tcp_error_m')}"
+            )
+    return 0 if report.passed else 41
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

{devagent_physical_engine-2.0.0 → devagent_physical_engine-2.1.0}/src/devagent_physical_engine/robot_platform.py

@@ -129,6 +129,10 @@ class RobotProfile:
     controller_family: str | None = None
     legacy_qualification: str | None = None
     notes: tuple[str, ...] = ()
+    # Independent authority bit for the packaged DevAgent runtime. A robot may be
+    # fully simulation-qualified while still being forbidden from producing measured
+    # customer FAT evidence. Never infer this authority from simulation tier badges.
+    measured_fat_authority: bool = False
 
     def __post_init__(self) -> None:
         key = self.key.strip().lower()
@@ -148,6 +152,19 @@ class RobotProfile:
             raise RobotPlatformError("empty_legacy_qualification")
         if any(not note.strip() for note in self.notes):
             raise RobotPlatformError("empty_robot_note")
+        if self.measured_fat_authority:
+            required = (
+                SimulationTier.KINEMATIC,
+                SimulationTier.PHYSICS,
+                SimulationTier.VISUAL,
+            )
+            if any(
+                self.qualification.state_for(tier) is not QualificationState.QUALIFIED
+                for tier in required
+            ):
+                raise RobotPlatformError(
+                    "measured_fat_authority_requires_qualified_kinematic_physics_visual"
+                )
         object.__setattr__(self, "key", key)
         object.__setattr__(self, "vendor", vendor)
         object.__setattr__(self, "model", model)
@@ -205,6 +222,7 @@ class RobotProfile:
             "site_qualification": self.qualification.site_qualification.value,
             "lifecycle_level": int(self.lifecycle_level),
             "lifecycle_name": self.lifecycle_level.name.lower(),
+            "measured_fat_authority": self.measured_fat_authority,
             "real_execution_locked": True,
         }
         if self.controller_family:
@@ -294,6 +312,7 @@ class RobotProfileRegistry:
                     "hardware_shadow": profile.qualification.hardware_shadow.value,
                     "sim_real": profile.qualification.sim_real_correlation.value,
                     "site": profile.qualification.site_qualification.value,
+                    "measured_fat_authority": profile.measured_fat_authority,
                 }
             )
         return tuple(rows)
@@ -355,6 +374,7 @@ BUILTIN_ROBOT_PROFILES = RobotProfileRegistry(
             notes=(
                 "DevAgent visual Gazebo/MoveIt workstation qualification exists; real execution remains locked.",
             ),
+            measured_fat_authority=True,
         ),
         RobotProfile(
             key="fanuc_crx",