Merge remote-tracking branch 'origin/main' into 313-bug-regression-tests-fail

harryswift01 · harryswift01 · commit c71cb40b6cba · 2026-05-07T11:25:59.000+01:00
diff --git a/.github/workflows/update_regression_baselines.yaml b/.github/workflows/update_regression_baselines.yaml
@@ -0,0 +1,44 @@
+name: Update Regression Baselines
+
+on:
+  workflow_dispatch:
+
+concurrency:
+  group: update-regression-baselines-${{ github.ref }}
+  cancel-in-progress: true
+
+permissions:
+  contents: write
+
+jobs:
+  update-baselines:
+    name: Update regression baselines
+    runs-on: ubuntu-24.04
+    timeout-minutes: 30
+
+    steps:
+      - name: Checkout
+        uses: actions/checkout@8e8c483db84b4bee98b60c0593521ed34d9990e8 # v6
+        with:
+          ref: ${{ github.ref }}
+
+      - name: Set up Python
+        uses: actions/setup-python@a309ff8b426b58ec0e2a45f0f869d46889d02405 # v6.2.0
+        with:
+          python-version: "3.14"
+          cache: pip
+
+      - name: Install testing dependencies
+        run: |
+          python -m pip install --upgrade pip
+          python -m pip install -e .[testing]
+
+      - name: Update regression baselines
+        run: pytest tests/regression --update-baselines
+
+      - name: Commit updated baselines
+        uses: stefanzweifel/git-auto-commit-action@778341af668090896ca464160c2def5d1d1a3eb0 # v6
+        with:
+          commit_message: "test(regression): update baselines"
+          commit_user_name: github-actions[bot]
+          commit_user_email: 41898282+github-actions[bot]@users.noreply.github.com
diff --git a/CITATION.cff b/CITATION.cff
@@ -76,5 +76,5 @@ keywords:
   - biomolecular simulations
   - protein flexibility
 license: MIT
-version: 2.1.1
-date-released: '2026-04-01'
+version: 2.2.2
+date-released: '2026-05-01'
diff --git a/CodeEntropy/__init__.py b/CodeEntropy/__init__.py
@@ -8,4 +8,4 @@
 and statistical mechanics.
 """
 
-__version__ = "2.1.1"
+__version__ = "2.2.2"
diff --git a/CodeEntropy/config/runtime.py b/CodeEntropy/config/runtime.py
@@ -335,8 +335,14 @@ def _build_universe(
         kcal_units = args.kcal_force_units
 
         if forcefile is None:
-            logger.debug(f"Loading Universe with {tprfile} and {trrfile}")
-            return mda.Universe(tprfile, trrfile, format=fileformat)
+            if fileformat == "LAMMPSDUMP":
+                logger.debug(
+                    f"Loading Universe with {tprfile} and {trrfile} (LAMMPSDUMP)"
+                )
+                return universe_operations.convert_lammps(tprfile, trrfile, fileformat)
+            else:
+                logger.debug(f"Loading Universe with {tprfile} and {trrfile}")
+                return mda.Universe(tprfile, trrfile, format=fileformat)
 
         return universe_operations.merge_forces(
             tprfile, trrfile, forcefile, fileformat, kcal_units
diff --git a/CodeEntropy/entropy/nodes/vibrational.py b/CodeEntropy/entropy/nodes/vibrational.py
@@ -323,7 +323,7 @@ def _compute_force_torque_entropy(
             EntropyPair containing translational entropy (from force covariance) and
             rotational entropy (from torque covariance).
         """
-        if fmat is None or tmat is None:
+        if fmat is None and tmat is None:
             return EntropyPair(trans=0.0, rot=0.0)
 
         f = self._mat_ops.filter_zero_rows_columns(
@@ -333,7 +333,7 @@ def _compute_force_torque_entropy(
             np.asarray(tmat), atol=self._zero_atol
         )
 
-        if f.size == 0 or t.size == 0:
+        if f.size == 0 and t.size == 0:
             return EntropyPair(trans=0.0, rot=0.0)
 
         s_trans = ve.vibrational_entropy_calculation(
diff --git a/CodeEntropy/levels/mda.py b/CodeEntropy/levels/mda.py
@@ -121,6 +121,97 @@ def extract_fragment(
         selection_string = f"index {frag.indices[0]}:{frag.indices[-1]}"
         return self.select_atoms(universe, selection_string)
 
+    def convert_lammps(
+        self,
+        tprfile: str,
+        trrfile,
+        fileformat: str | None = None,
+    ) -> mda.Universe:
+        """Update the units produced from the universe produced from LAMMPS
+        format topology and trajectory files. MDA currently has a bug that
+        results in forces not being converted to the correct units
+        (see issue for more details:
+        https://github.com/MDAnalysis/mdanalysis/issues/5115
+        )
+        The method currently expects the following additional columns in the
+        lammps dump file: fx fy fz c_5 c_7
+        where c_5 and c_7 are the atom potential and kinetic energies respectively.
+
+        This method loads:
+
+        - Coordinates and dimensions from the coordinate trajectory
+          (``tprfile`` + ``trrfile``).
+
+        Args:
+            tprfile: Topology input file.
+            trrfile: Coordinate trajectory file(s). This can be a single path or a
+                list, as accepted by MDAnalysis.
+            fileformat: Optional file format for the coordinate trajectory, as
+                recognised by MDAnalysis.
+
+        Returns:
+            MDAnalysis.Universe: A new Universe containing coordinates, forces and
+            dimensions loaded into memory.
+
+        Raises:
+            ValueError: If fileformat is not one of the supported values.
+        """
+
+        def _convert_lammps_forces_energies(ts):
+            """
+            Convert lammps forces from kcal/mol/Ang to kJ/mol/Ang.
+            Assumes columns for per-atom potential (c_5) and kinetic energies (c_7)
+            are provided and converts these too.
+
+            Args:
+                ts: MDAnalysis timeseries from the trajectory.
+
+            Returns:
+                A converted time series.
+            """
+            ts.forces *= 4.184
+            ts.data["c_5"] *= 4.184
+            ts.data["c_7"] *= 4.184
+            return ts
+
+        def _convert_lammps_forces(ts):
+            """
+            Convert lammps forces from kcal/mol/Ang to kJ/mol/Ang.
+
+            Args:
+                ts: MDAnalysis timeseries from the trajectory.
+
+            Returns:
+                A converted time series.
+            """
+            ts.forces *= 4.184
+            return ts
+
+        if fileformat == "LAMMPSDUMP":
+            try:
+                return mda.Universe(
+                    tprfile,
+                    trrfile,
+                    format=fileformat,
+                    additional_columns=["fx", "fy", "fz", "c_5", "c_7"],
+                    transformations=[_convert_lammps_forces_energies],
+                )
+            except KeyError:
+                logger.debug(
+                    f"Warning: Energy columns not found in LAMMPSDUMP: {trrfile}"
+                )
+                return mda.Universe(
+                    tprfile,
+                    trrfile,
+                    format=fileformat,
+                    additional_columns=["fx", "fy", "fz"],
+                    transformations=[_convert_lammps_forces],
+                )
+        else:
+            raise ValueError(
+                f"Incorrect file format: {fileformat}, LAMMPSDUMP expected"
+            )
+
     def merge_forces(
         self,
         tprfile: str,
diff --git a/CodeEntropy/levels/neighbors.py b/CodeEntropy/levels/neighbors.py
@@ -167,7 +167,11 @@ def _get_rdkit_mol(self, universe, mol_id):
             rdkit_mol = frag.convert_to("RDKIT", force=True, inferrer=None)
             logger.debug("Warning: Dummy atoms found")
         else:
-            rdkit_mol = molecule.convert_to("RDKIT", force=True)
+            try:
+                rdkit_mol = molecule.convert_to("RDKIT", force=True)
+            except Exception:
+                logger.debug("Warning: Constraint bonds to H atoms found")
+                rdkit_mol = molecule.convert_to("RDKIT", force=True, inferrer=None)
 
         number_heavy = rdkit_mol.GetNumHeavyAtoms()
         number_hydrogen = rdkit_mol.GetNumAtoms() - number_heavy
diff --git a/conda-recipe/meta.yaml b/conda-recipe/meta.yaml
@@ -39,7 +39,7 @@ requirements:
     - distributed >=2026.1.2,<2026.2.0
     - dask-jobqueue >=0.9,<0.10
     - pytest-xdist >=3.8, <3.9
-    - numba >=0.65.0, <0.7
+    - numba >=0.65.1, <0.70
 
 test:
   imports:
@@ -48,7 +48,13 @@ test:
     - CodeEntropy --help
 
 about:
-  home: https://ccpbiosim.github.io/CodeEntropy/
+  home: https://github.com/CCPBioSim/CodeEntropy
   license: MIT
+  license_family: MIT
   license_file: LICENSE
-  summary: "Entropy analysis tools for macromolecular systems and MD simulation"
+  license_url: https://github.com/CCPBioSim/CodeEntropy/blob/main/LICENSE
+  summary: Python package for calculating configurational entropy from molecular dynamics simulations using the multiscale cell correlation method.
+  description: "CodeEntropy is a Python package for computing the configurational entropy of macromolecular systems using forces sampled from molecular dynamics (MD) simulations. It implements the multiscale cell correlation method to provide accurate and efficient entropy estimates, supporting a wide range of applications in molecular simulation and statistical mechanics."
+  dev_url: https://github.com/CCPBioSim/CodeEntropy
+  doc_url: https://codeentropy.readthedocs.io/en/latest/
+  doc_source_url: https://github.com/CCPBioSim/CodeEntropy/blob/main/README.md
diff --git a/docs/developer_guide.rst b/docs/developer_guide.rst
@@ -60,13 +60,46 @@ Run tests with coverage::
 
     pytest --cov CodeEntropy --cov-report=term-missing
 
-Update regression baselines::
+Run a specific test::
+
+    pytest tests/unit/.../test_file.py::test_function
+
+Updating Regression Baselines
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+Regression baselines should only be updated when a change intentionally alters
+the expected numerical output of CodeEntropy. Do not update baselines simply
+because a regression test fails.
+
+Before updating baselines, confirm that:
+
+- The code change is intentional and understood.
+- The new output has been reviewed.
+- The regression difference is expected.
+- The pull request explains why the baselines changed.
+
+For local updates, run::
 
     pytest tests/regression --update-baselines
 
-Run a specific test::
+For pull requests, baselines can also be updated using the GitHub Actions
+workflow **Update Regression Baselines**.
 
-    pytest tests/unit/.../test_file.py::test_function
+To use it:
+
+1. Push your changes to your PR branch.
+2. Open the **Actions** tab on GitHub.
+3. Select **Update Regression Baselines**.
+4. Click **Run workflow**.
+5. Select your PR branch.
+6. Run the workflow.
+
+The workflow runs the regression tests with ``--update-baselines`` and commits
+any changed baseline files back to the selected branch.
+
+Only use this workflow when the baseline changes are intentional. If a regression
+test fails unexpectedly, investigate the failure instead of updating the
+baseline.
 
 Regression Test Data
 --------------------
diff --git a/tests/unit/CodeEntropy/entropy/nodes/test_vibrational_node.py b/tests/unit/CodeEntropy/entropy/nodes/test_vibrational_node.py
@@ -321,7 +321,7 @@ def test_compute_force_torque_entropy_returns_zero_when_missing_matrix(shared_gr
     node = VibrationalEntropyNode()
     ve = MagicMock()
     pair = node._compute_force_torque_entropy(
-        ve=ve, temp=298.0, fmat=None, tmat=np.eye(3), flexible=0, highest=True
+        ve=ve, temp=298.0, fmat=None, tmat=None, flexible=0, highest=True
     )
     assert pair == EntropyPair(trans=0.0, rot=0.0)
 
diff --git a/tests/unit/CodeEntropy/levels/test_mda_universe_operations.py b/tests/unit/CodeEntropy/levels/test_mda_universe_operations.py
@@ -118,6 +118,59 @@ def test_merge_forces_scales_kcal(monkeypatch):
     assert np.allclose(forces, np.ones((2, 2, 3)) * 4.184)
 
 
+def test_convert_lammps_transforms_forces_and_energies(monkeypatch):
+    ops = UniverseOperations()
+
+    mock_universe = MagicMock()
+    transformations_captured = []
+
+    def capture_universe(*args, **kwargs):
+        if "transformations" in kwargs:
+            transformations_captured.extend(kwargs["transformations"])
+        return mock_universe
+
+    monkeypatch.setattr("CodeEntropy.levels.mda.mda.Universe", capture_universe)
+
+    ops.convert_lammps("tpr", "trr", "LAMMPSDUMP")
+
+    ts = MagicMock()
+    ts.forces = np.array([[1.0, 2.0, 3.0]], dtype=float)
+    ts.data = {"c_5": np.array([1.0]), "c_7": np.array([2.0])}
+
+    transformations_captured[0](ts)
+
+    assert np.allclose(ts.forces, np.array([[1.0, 2.0, 3.0]], dtype=float) * 4.184)
+    assert np.allclose(ts.data["c_5"], np.array([1.0], dtype=float) * 4.184)
+    assert np.allclose(ts.data["c_7"], np.array([2.0], dtype=float) * 4.184)
+
+
+def test_convert_lammps_fallback_on_keyerror(monkeypatch):
+    ops = UniverseOperations()
+
+    transformations_captured = []
+    call_count = [0]
+
+    def mock_universe(*args, **kwargs):
+        call_count[0] += 1
+        if call_count[0] == 1:
+            raise KeyError("c_5")
+        if "transformations" in kwargs:
+            transformations_captured.extend(kwargs["transformations"])
+        return MagicMock()
+
+    monkeypatch.setattr("CodeEntropy.levels.mda.mda.Universe", mock_universe)
+
+    ops.convert_lammps("tpr", "trr", "LAMMPSDUMP")
+
+    ts = MagicMock()
+    ts.forces = np.array([[1.0, 2.0, 3.0]], dtype=float)
+
+    transformations_captured[0](ts)
+
+    assert np.allclose(ts.forces, np.array([[1.0, 2.0, 3.0]], dtype=float) * 4.184)
+    assert call_count[0] == 2
+
+
 def test_select_atoms_builds_merged_universe_and_loads_timeseries(monkeypatch):
     ops = UniverseOperations()
 

Original file line number	Diff line number	Diff line change
`@@ -321,7 +321,7 @@ def test_compute_force_torque_entropy_returns_zero_when_missing_matrix(shared_gr`
`321`	`321`	`node = VibrationalEntropyNode()`
`322`	`322`	`ve = MagicMock()`
`323`	`323`	`pair = node._compute_force_torque_entropy(`
`324`		`- ve=ve, temp=298.0, fmat=None, tmat=np.eye(3), flexible=0, highest=True`
	`324`	`+ ve=ve, temp=298.0, fmat=None, tmat=None, flexible=0, highest=True`
`325`	`325`	`)`
`326`	`326`	`assert pair == EntropyPair(trans=0.0, rot=0.0)`
`327`	`327`