Triplet S2 perception (#778)

Fixes an issue with perceiving S2 as a triplet from xyz, building upon #756
ReactionMechanismGenerator · Dec 8, 2024 · 6fda63e · 6fda63e
2 parents 84898ee + 92bbc31
commit 6fda63e
Show file tree

Hide file tree

Showing 4 changed files with 32 additions and 13 deletions.
diff --git a/arc/species/converter.py b/arc/species/converter.py
@@ -1380,12 +1380,14 @@ def molecules_from_xyz(xyz: Optional[Union[dict, str]],
         return None, None
     xyz = check_xyz_dict(xyz)
 
-    if xyz['symbols'] == ('O', 'O') and multiplicity != 1:
-        coords = np.asarray(xyz['coords'], dtype=np.float32)
-        vector = coords[0] - coords[1]
-        if float(np.dot(vector, vector) ** 0.5) < 1.4:
-            # Special case for O2 triplet
-            return Molecule(smiles='[O][O]'), Molecule(smiles='[O][O]')
+    if len(xyz['symbols']) == 2:
+        for element, bond_length in zip(['O', 'S'], [1.4, 2.1]):
+            if xyz['symbols'] == (element, element) and multiplicity != 1:
+                coords = np.asarray(xyz['coords'], dtype=np.float32)
+                vector = coords[0] - coords[1]
+                if float(np.dot(vector, vector) ** 0.5) < bond_length:
+                    # Special case for O2 and S2 triplet
+                    return Molecule(smiles=f'[{element}][{element}]'), Molecule(smiles=f'[{element}][{element}]')
 
     # 1. Generate a molecule with no bond order information with atoms ordered as in xyz.
     mol_graph = MolGraph(symbols=xyz['symbols'], coords=xyz['coords'])

diff --git a/arc/species/converter_test.py b/arc/species/converter_test.py
@@ -3588,8 +3588,10 @@ def test_xyz_to_smiles(self):
         mol18 = converter.molecules_from_xyz(converter.str_to_xyz(xyz18))[1]
         mol19 = converter.molecules_from_xyz(converter.str_to_xyz(xyz19))[1]
         mol20 = converter.molecules_from_xyz(converter.str_to_xyz(xyz20))[1]
-        mol21 = converter.molecules_from_xyz("""O	0.0000000	0.0000000	0.6076340
-                                                O	0.0000000	0.0000000	-0.6076340""")[1]
+        mol21 = converter.molecules_from_xyz("""O    0.0000000    0.0000000    0.6076340
+                                                O    0.0000000    0.0000000   -0.6076340""")[1]
+        mol22 = converter.molecules_from_xyz("""S    0.0000000    0.0000000    0.9547220
+                                                S    0.0000000    0.0000000   -0.9547220""")[1]
 
         self.assertEqual(mol1.to_smiles(), '[NH-][S+](=O)(O)C')
         self.assertIn(mol2.to_smiles(), ['COC1=C(CO)C=C([C](C)C)C=C1', 'COC1C=CC(=CC=1CO)[C](C)C'])
@@ -3622,6 +3624,8 @@ def test_xyz_to_smiles(self):
         self.assertEqual(mol20.to_smiles(), 'C=C[CH]C=CC')
         self.assertEqual(mol21.to_smiles(), '[O][O]')
         self.assertEqual(mol21.multiplicity, 3)
+        self.assertEqual(mol22.to_smiles(), '[S][S]')
+        self.assertEqual(mol22.multiplicity, 3)
 
     def test_to_rdkit_mol(self):
         """Test converting an RMG Molecule object to an RDKit Molecule object"""

diff --git a/arc/species/species.py b/arc/species/species.py
@@ -1583,11 +1583,6 @@ def mol_from_xyz(self,
         if xyz is None:
             return None
 
-        if len(xyz['symbols']) == 2 and xyz['symbols'][0] == xyz['symbols'][1] \
-                and xyz['symbols'][0] in ['O', 'S'] and self.multiplicity == 3:
-            # Hard-coded for triplet O2 and S2: Don't perceive mol.
-            return None
-
         if self.mol is not None:
             if len(self.mol.atoms) != len(xyz['symbols']):
                 raise SpeciesError(f'The number of atoms in the molecule and in the coordinates of {self.label} is different.'

diff --git a/arc/species/species_test.py b/arc/species/species_test.py
@@ -1686,6 +1686,24 @@ def test_mol_from_xyz(self):
         so2.mol_from_xyz(xyz=so2_t_xyz, get_cheap=False)
         self.assertEqual([atom.element.symbol for atom in so2.mol.atoms], ['S', 'O', 'O'])
 
+        # S2
+        s2_xyz = """S    0.0000000    0.0000000    0.9547220
+                    S    0.0000000    0.0000000   -0.9547220"""
+        s2 = ARCSpecies(label='S2', smiles='[S][S]')
+        self.assertEqual(s2.multiplicity, 3)
+        self.assertEqual(s2.charge, 0)
+        s2.mol_from_xyz(xyz=str_to_xyz(s2_xyz), get_cheap=False)
+        self.assertEqual([atom.element.symbol for atom in s2.mol.atoms], ['S', 'S'])
+        self.assertEqual(s2.mol.to_smiles(), '[S][S]')
+        self.assertEqual(s2.mol.multiplicity, 3)
+
+        # O2
+        o2_xyz = {'symbols': ('O', 'O'), 'isotopes': (16, 16), 'coords': ((0.0, 0.0, 0.6029), (0.0, 0.0, -0.6029))}
+        o2 = ARCSpecies(label='O2aa', xyz=o2_xyz, multiplicity=3)
+        self.assertEqual(o2.multiplicity, 3)
+        self.assertEqual(o2.mol.multiplicity, 3)
+        self.assertEqual(o2.mol.to_smiles(), '[O][O]')
+
     def test_consistent_atom_order(self):
         """Test that the atom order is preserved whether starting from SMILES or from xyz"""
         xyz9 = """O      -1.17310019   -0.30822930    0.16269772