Fix error in charge assignment and spin

cell2mol/charge_assignment.py

@@ -35,7 +35,7 @@ def get_possible_charge_state(spec: object, debug: int=0):
     if spec.protonation_states is None: 
         return None
 
-    if spec.formula in ["O4-Cl", "N3", "I3"]:
+    if spec.formula in ["O4-Cl", "N3", "I3", "N2"]:
         ch_state = get_charge_manual(spec, debug=debug)
         possible_cs = [ch_state]
         return possible_cs
@@ -437,7 +437,8 @@ def get_protonation_states_specie(specie: object, debug: int=0) -> list:
                             addedlist[idx] = 1
                     else:
                         # nitrogen with at least 3 adjacencies doesnt need H
-                        if a.connec >= 3:
+                        # if a.connec >= 3:
+                        if (a.connec - a.mconnec) >= 3 :
                             block[idx] = 1
                             # needs_nonlocal = True
                             # non_local_groups += 1
@@ -700,7 +701,11 @@ def get_charge_manual(spec, debug: int=0):
                 break  # Found the target atom, no need to check further
 
         if debug >= 2: print(f"I3: new_order={new_order}")
-
+    elif spec.formula == "N2":
+        smiles = "N#N"
+        charge = 0
+        order = [0, 1]
+        new_order = order
 
     temp_mol = Chem.MolFromSmiles(smiles, sanitize=False)
     mol = Chem.RenumberAtoms(temp_mol, new_order)
@@ -943,9 +948,9 @@ def get_metal_poscharges(metal: object, debug: int=0) -> list:
     at_charge[23] = [1, 2, 3, 4, 5]  # V
     at_charge[24] = [0, 2, 3] # Cr ; including 5 leads to worse results
     at_charge[25] = [1, 2, 3]  # Mn
-    at_charge[26] = [2, 3]  # Fe
+    at_charge[26] = [0, 2, 3]  # Fe
     at_charge[27] = [1, 2, 3]  # Co
-    at_charge[28] = [2, 3]  # Ni
+    at_charge[28] = [0, 2, 3]  # Ni
     at_charge[29] = [1, 2]  # Cu
     at_charge[30] = [2]  # Zn
     # 2nd-row transition metals.
@@ -954,7 +959,7 @@ def get_metal_poscharges(metal: object, debug: int=0) -> list:
     at_charge[41] = [1, 3, 4, 5]  # Nb
     at_charge[42] = [0, 2, 4, 5, 6]  # Mo
     at_charge[43] = [1, 2, 3, 4, 5]  # Tc
-    at_charge[44] = [2, 3, 4]  # Ru
+    at_charge[44] = [0, 2, 3, 4]  # Ru
     at_charge[45] = [1, 2, 3]  # Rh
     at_charge[46] = [0, 2]  # Pd
     at_charge[47] = [1]  # Ag
@@ -971,14 +976,14 @@ def get_metal_poscharges(metal: object, debug: int=0) -> list:
     at_charge[80] = [2]  # Hg
 
     # post-transition metals
-    at_charge[13] = [0, 3]  # Al
-    at_charge[31] = [0, 3]  # Ga
-    at_charge[32] = [0, 2, 4]  # Ge
-    at_charge[49] = [0, 3]  # In
-    at_charge[50] = [0, 2, 4]  # Sn
-    at_charge[81] = [0, 1, 3]  # Tl
-    at_charge[82] = [0, 2, 4]  # Pb
-    at_charge[83] = [0, 3]  # Bi
+    at_charge[13] = [3]  # Al
+    at_charge[31] = [3]  # Ga
+    at_charge[32] = [2, 4]  # Ge
+    at_charge[49] = [3]  # In
+    at_charge[50] = [2, 4]  # Sn
+    at_charge[81] = [1, 3]  # Tl
+    at_charge[82] = [2, 4]  # Pb
+    at_charge[83] = [3]  # Bi
 
     # Lanthanides (atomic numbers 57 to 71)
     at_charge[57] = [3]  # La
@@ -1227,6 +1232,8 @@ def correct_smiles_ligand(ligand: object, debug: int=0) -> Tuple[str, object]:
             ismetal_2 = elemdatabase.elementblock[b.atom2.label] == "d" or elemdatabase.elementblock[b.atom2.label] == "f"
             if ismetal_1 or ismetal_2:
                 pass
+            elif len(get_non_transition_metal_idxs([b.atom1.label, b.atom2.label])) > 0:
+                pass
             else:
                 begin_idx = b.atom1.get_parent_index("ligand")
                 end_idx = b.atom2.get_parent_index("ligand")

cell2mol/spin.py

@@ -29,14 +29,18 @@ def assign_spin_metal (metal:object, debug: int=0) -> None:
         elif valence_elec in [1, 9]:                                       return 2
         elif valence_elec in [2, 3] and metal.get_parent("molecule").is_haptic == False :         return (valence_elec + 1)
         elif valence_elec in [4, 5, 6, 7, 8] or (valence_elec in [2, 3] and metal.get_parent("molecule").is_haptic == True) :
-            # Predict spin multiplicity of metal based on Random forest model
-            feature = generate_feature_vector (metal, target_prop="spin", debug=debug)
-            path_rf = os.path.join( os.path.abspath(os.path.dirname(__file__)), "total_spin_3131.pkl")
-            ramdom_forest = pickle.load(open(path_rf, 'rb'))
-            predictions = ramdom_forest.predict(feature)
-            spin_rf = predictions[0]
-            print(f"ASSIGN_SPIN_METAL: Spin multiplicity of the metal {metal.label} is predicted as {spin_rf} using Random Forest model")
-            return spin_rf
+            if hasattr(metal, "coord_geometry") and metal.coord_geometry != "Undefined":  
+                # Predict spin multiplicity of metal based on Random forest model
+                feature = generate_feature_vector (metal, target_prop="spin", debug=debug)
+                path_rf = os.path.join( os.path.abspath(os.path.dirname(__file__)), "total_spin_3131.pkl")
+                ramdom_forest = pickle.load(open(path_rf, 'rb'))
+                predictions = ramdom_forest.predict(feature)
+                spin_rf = predictions[0]
+                print(f"ASSIGN_SPIN_METAL: Spin multiplicity of the metal {metal.label} is predicted as {spin_rf} using Random Forest model")
+                return spin_rf
+            else:
+                print("ASSIGN_SPIN_METAL: Error! Coordination geometry of the metal is not defined.")
+                return None
         else :
             print("ASSIGN_SPIN_METAL: Error! Spin multiplicity could not be assigned to the metal with valence electrons: ", valence_elec)
             return None

cell2mol/xyz2mol.py

@@ -604,7 +604,7 @@ def AC2BO(AC, atoms, charge, allow_charged_fragments=True, use_graph=True):
                 and valences_not_too_large(BO, valences)
                 and charge_OK
             ):
-                print(f"\tAC2BO: status", status, "BO.sum()", BO.sum(), "best_BO.sum()", best_BO.sum())
+                #print(f"\tAC2BO: status", status, "BO.sum()", BO.sum(), "best_BO.sum()", best_BO.sum())
                 best_BO = BO.copy()
 
             count += 1