Screen ellipsoid ode45 cases against refined full-duration source

.github/workflows/ellipsoid-ode45-convergence.yml

@@ -0,0 +1,73 @@
+name: Ellipsoid ode45 full-duration source convergence
+
+on:
+  pull_request:
+    paths:
+      - '.github/workflows/ellipsoid-ode45-convergence.yml'
+      - 'tests/matlab/ellipsoid_ode45_convergence.m'
+      - 'tests/test_ellipsoid_ode45_convergence.py'
+      - 'tests/test_ellipsoid_ode45_convergence_portable.py'
+      - 'tests/fixtures/matlab/ellipsoid_ode45_convergence/**'
+      - 'wecsim/hydrodynamics/nonlinearHydro.py'
+      - 'wecsim/dynamics/caseDynamics.py'
+  workflow_dispatch:
+
+jobs:
+  compare:
+    strategy:
+      fail-fast: false
+      matrix:
+        wave: [Regular, RegularCIC]
+    runs-on: ubuntu-latest
+    timeout-minutes: 120
+    name: ${{ matrix.wave }}
+    steps:
+      - uses: actions/checkout@v7
+      - uses: actions/checkout@v7
+        with:
+          repository: WEC-Sim/WEC-Sim
+          ref: 0753b2e47f2457c078751dcfe5d251d1767b80ab
+          path: matlab-ref
+          filter: blob:none
+          sparse-checkout: source
+      - uses: actions/checkout@v7
+        with:
+          repository: WEC-Sim/WEC-Sim_Applications
+          ref: d53d4d4c9eda2581f04204f5d394a6ef84bb099e
+          path: applications
+          filter: blob:none
+          lfs: true
+          sparse-checkout: Nonlinear_Hydro
+      - uses: matlab-actions/setup-matlab@v3
+        with:
+          release: R2025b
+          products: Simulink Simscape Simscape_Multibody
+      - uses: matlab-actions/run-command@v3
+        with:
+          command: addpath('tests/matlab'); ellipsoid_ode45_convergence('${{ matrix.wave }}', 'step005')
+      - uses: matlab-actions/run-command@v3
+        with:
+          command: addpath('tests/matlab'); ellipsoid_ode45_convergence('${{ matrix.wave }}', 'step0025')
+      - uses: matlab-actions/run-command@v3
+        with:
+          command: addpath('tests/matlab'); ellipsoid_ode45_convergence('${{ matrix.wave }}', 'step00125')
+      - uses: matlab-actions/run-command@v3
+        with:
+          command: addpath('tests/matlab'); ellipsoid_ode45_convergence('${{ matrix.wave }}', 'step000625', 0.005)
+      - uses: actions/setup-python@v7
+        with:
+          python-version: '3.12'
+      - run: python -m pip install -e . pytest
+      - run: python -m pytest -q -s tests/test_ellipsoid_ode45_convergence.py
+        env:
+          WEC_SIM_APPLICATIONS_DIR: applications
+          WEC_SIM_ELLIPSOID_CONVERGENCE_DIR: matlab-ellipsoid-ode45-convergence
+          WEC_SIM_ELLIPSOID_CASE: ${{ matrix.wave }}
+      - uses: actions/upload-artifact@v4
+        if: always()
+        with:
+          name: ellipsoid-ode45-convergence-${{ matrix.wave }}
+          path: |
+            matlab-ellipsoid-ode45-convergence/${{ matrix.wave }}/
+            applications/Nonlinear_Hydro/hydroData/ellipsoid.h5
+            applications/Nonlinear_Hydro/geometry/elipsoid.stl

.github/workflows/ellipsoid-ode45-output-grid-ab.yml

@@ -0,0 +1,59 @@
+name: Ellipsoid ode45 source output-grid A/B
+
+on:
+  pull_request:
+    paths:
+      - '.github/workflows/ellipsoid-ode45-output-grid-ab.yml'
+      - 'tests/matlab/ellipsoid_ode45_output_grid_ab.m'
+      - 'tests/test_ellipsoid_ode45_output_grid_ab.py'
+  workflow_dispatch:
+
+jobs:
+  compare:
+    strategy:
+      fail-fast: false
+      matrix:
+        wave: [Regular, RegularCIC]
+    runs-on: ubuntu-latest
+    timeout-minutes: 90
+    name: ${{ matrix.wave }}
+    steps:
+      - uses: actions/checkout@v7
+      - uses: actions/checkout@v7
+        with:
+          repository: WEC-Sim/WEC-Sim
+          ref: 0753b2e47f2457c078751dcfe5d251d1767b80ab
+          path: matlab-ref
+          filter: blob:none
+          sparse-checkout: source
+      - uses: actions/checkout@v7
+        with:
+          repository: WEC-Sim/WEC-Sim_Applications
+          ref: d53d4d4c9eda2581f04204f5d394a6ef84bb099e
+          path: applications
+          filter: blob:none
+          lfs: true
+          sparse-checkout: Nonlinear_Hydro
+      - uses: matlab-actions/setup-matlab@v3
+        with:
+          release: R2025b
+          products: Simulink Simscape Simscape_Multibody
+      - uses: matlab-actions/run-command@v3
+        with:
+          command: addpath('tests/matlab'); ellipsoid_ode45_output_grid_ab('${{ matrix.wave }}')
+      - uses: actions/setup-python@v7
+        with:
+          python-version: '3.12'
+      - run: python -m pip install -e . pytest
+      - run: python -m pytest -q -s tests/test_ellipsoid_ode45_output_grid_ab.py
+        env:
+          WEC_SIM_ELLIPSOID_AB_DIR: matlab-ellipsoid-ode45-output-grid-ab
+          WEC_SIM_ELLIPSOID_AB_CASE: ${{ matrix.wave }}
+      - uses: actions/upload-artifact@v4
+        if: always()
+        with:
+          name: ellipsoid-ode45-output-grid-ab-${{ matrix.wave }}
+          path: |
+            matlab-ellipsoid-ode45-output-grid-ab/${{ matrix.wave }}/
+            applications/Nonlinear_Hydro/hydroData/ellipsoid.h5
+            applications/Nonlinear_Hydro/geometry/elipsoid.stl

tests/fixtures/matlab/ellipsoid_ode45_convergence/README.md

@@ -0,0 +1,104 @@
+# Ellipsoid ode45 derived source-step screen
+
+This fixture retains full 150 s R2025b source histories for the
+`Nonlinear_Hydro/ode45/Regular` and `RegularCIC` applications. The source
+workflow checks out WEC-Sim core
+`0753b2e47f2457c078751dcfe5d251d1767b80ab` and WEC-Sim_Applications
+`d53d4d4c9eda2581f04204f5d394a6ef84bb099e`. It creates a copy of each
+application input for every run, replacing the published `simu.dt = 0.05`
+with 0.005, 0.0025, 0.00125, or 0.000625 s. At runtime it verifies that
+`simu.dt`, `nonlinearDt`, `cicDt`, and Simulink `MaxStep` all equal the chosen
+physical step. The first three arms also use that step for `dtOut`. The finest
+arm explicitly sets `dtOut=0.005` s and retains 0.000625 s for all four
+physical clocks. Each arm runs in a fresh R2025b process over the full 150 s.
+The workflow generates the archived BEM HDF5 from pinned WAMIT output in the
+application checkout and copies its STL; neither upstream source repository
+is changed.
+
+`source.npz` stores all eleven output channels and five checked source settings
+at every source step. The finest history uses the common 0.005 s output grid.
+`hydroData/ellipsoid.h5` and `geometry/elipsoid.stl` let the portable Python
+test advance an independent physical model from the same input. The archive
+tool checks all source settings, validates the 0.005 s history against the
+earlier [full-step source archive](../ellipsoid_ode45_maxstep_only/README.md),
+checks wave identity, and checks contraction of heave, speed, PTO force, PTO
+power, and signed PTO work across the refinement sequence. Pointwise source
+changes are screened on the same 0.005 s grid, with their original numerical
+limits unchanged. Signed work is integrated on the full available grid for
+the first three arms and on the 0.005 s grid for the finest; the test checks
+the full/common integration difference in each earlier arm.
+For every arm, the archive and portable gate also check PTO vertical speed
+against body heave speed and verify `Fz = -1,200,000 vz` and `P = -Fz vz`
+within 1e-9 m/s, 1e-6 N, and 1e-6 W, respectively. These identities reject
+the previously encountered inactive local-axis export.
+
+The [controlled output-grid A/B](https://github.com/cmudrc/wec-sim-python/actions/runs/40104326372)
+found exact common-grid trajectory and PTO equality when `dtOut` changed from
+0.00125 to 0.005 s with the physical clocks fixed at 0.00125 s. This supports
+an output-sampling explanation but establishes neutrality only at that tested
+physical step. Four all-clock 0.000625 s attempts, including an isolated fresh
+process, lost hosted runners during simulation without a MATLAB or physical
+gate result. The mixed-output finest source arm is therefore labeled as a
+derived source input in its own right, with unchanged full physical gates.
+
+The source model retains its 1e-7 s added-mass feedback Transport Delay. Its
+logged restoring force samples the preceding state at the 0.00125 s all-clock
+arm. The 0.000625 s arm does not log every solver step, so its immediately
+preceding source state is unavailable. The independent Python model evaluates
+force from its current state; the test exposes this timing law explicitly at
+0.00125 s. Passing this step-screened, derived reference does not change the
+separate, pinned 0.05 s published-input
+diagnostic or its `qualified_motion` status.
+
+The portable gate rechecks all source convergence conditions, advances Python
+at 0.0025 and 0.00125 s over the full duration, and compares its finest run
+directly against the 0.000625 s source on the 0.005 s output grid. Its original
+pointwise motion/PTO limits remain unchanged. The Python self-step limits are
+0.1 mm, 0.1 mm/s, 100 N, and 100 W. The signed source-work final-step change
+must be below 0.1% and contract by a factor below 0.7 at each refinement;
+the direct Python/source signed-work difference must also be below 0.1%.
+
+Original all-clock source run: [R2025b workflow 40100441636](https://github.com/cmudrc/wec-sim-python/actions/runs/40100441636).
+Isolated all-clock diagnostic: [R2025b workflow 40105605489](https://github.com/cmudrc/wec-sim-python/actions/runs/40105605489).
+
+The archived corrected-axis source histories came from these Python-repository
+workflow runs and exact commits:
+
+| Source arms | R2025b run | Python-repository commit |
+| --- | --- | --- |
+| Both cases, 0.005/0.0025/0.00125 s | [40107175117](https://github.com/cmudrc/wec-sim-python/actions/runs/40107175117) | `daa282c9c3ccb6041b7ccccb89304944ea351d36` |
+| Regular, physical 0.000625 s with `dtOut=0.005` s | [40106411780](https://github.com/cmudrc/wec-sim-python/actions/runs/40106411780) | `44b0809b295e2163d433be98b246be2ebf58113d` |
+| RegularCIC, physical 0.000625 s with `dtOut=0.005` s | [40107064277](https://github.com/cmudrc/wec-sim-python/actions/runs/40107064277) | `1e13e500f815dcb91168e985c2af060bc95b077d` |
+
+The measured source changes below compare adjacent arms on their common
+0.005 s grid. Each channel contracted through both intermediate steps, with
+successive change ratios below the predeclared 0.7 bound.
+
+| Case | Final heave change | Final speed change | Final PTO force change | Final PTO power change | Final signed-work change |
+| --- | ---: | ---: | ---: | ---: | ---: |
+| Regular | 0.000231452 m | 0.000341197 m/s | 409.436 N | 333.393 W | 0.069817% |
+| RegularCIC | 0.000241166 m | 0.000362903 m/s | 435.484 N | 360.546 W | 0.073934% |
+
+The four signed-work measurements on each arm's available output grid are
+20.154279, 20.098017, 20.069978, and 20.055976 MJ for Regular; and 20.186951, 20.127222,
+20.097460, and 20.082612 MJ for RegularCIC. The final source-work change is
+below the declared 0.1% limit for both cases. Common-grid work also contracts
+and passes the same limit. Source settings, wave identity, and archived 0.005 s
+continuity pass for both cases. All 38 BEM HDF5 datasets agree in value across
+the source arms, and the STL files have identical SHA-256 hashes.
+
+The independent Python 0.0025→0.00125 s self-step gates pass. Against the
+finest source arm, direct Python heave/speed/PTO-force/PTO-power errors are
+0.000231346 m, 0.000338943 m/s, 406.732 N, and 328.493 W for Regular;
+0.000241018 m, 0.000362768 m/s, 435.321 N, and 359.533 W for RegularCIC.
+Their signed-work differences are 0.069727% and 0.073878%, respectively,
+below the unchanged 0.1% direct gate. These measurements apply only to this
+derived source-step screen.
+
+Archived file SHA-256 hashes:
+
+| File | SHA-256 |
+| --- | --- |
+| `source.npz` | `8dd1d8d28feab13617eab78afad02e128e1fee017364db906af7d29f4e808f23` |
+| `hydroData/ellipsoid.h5` | `cd437b3fcdb2837b1534abadc3591d21cf3f001812407973b1cdd12517c5f489` |
+| `geometry/elipsoid.stl` | `a225213457fb9273a327a141273b9d68266e650cf973a98f84e84e8607cbb7ec` |