tests and account for bond order

cgsmiles/pysmiles_utils.py

@@ -67,12 +67,12 @@ def rebuild_h_atoms(mol_graph, keep_bonding=False):
         pysmiles.smiles_helper.correct_aromatic_rings(mol_graph, strict=True)
     except SyntaxError as pysmiles_err:
         print(pysmiles_err)
-        msg = ("Likely you are writing an aromatic molecule that does not "
-               "show delocalization-induced molecular equivalency and thus "
-               "is not considered aromatic. For example, 4-methyl imidazole "
-               "is often written as [nH]1cc(nc1)C, but should be written as "
-               "[NH]1C=C(N=C1)C. A corresponding CGSmiles string would be "
-               "{[#A]1[#B][#C]1}.{#A=[>][<]N,#B=[$]N=C[>],#C=[$]C(C)=C[<]}")
+        msg = (r"Likely you are writing an aromatic molecule that does not "
+               r"show delocalization-induced molecular equivalency and thus "
+               r"is not considered aromatic. For example, 4-methyl imidazole "
+               r"is often written as [nH]1cc(nc1)C, but should be written as "
+               r"[NH]1C=C(N=C1)C. A corresponding CGSmiles string would be "
+               r"{[#A]1[#B][#C]1}.{#A=[>][<]N,#B=[$]N=C[>],#C=[$]C(C)=C[<]}")
         raise SyntaxError(msg)
     nx.set_node_attributes(mol_graph, 0, 'hcount')
 
@@ -84,7 +84,9 @@ def rebuild_h_atoms(mol_graph, keep_bonding=False):
     if keep_bonding:
         bonding_nodes = nx.get_node_attributes(mol_graph, 'bonding')
         for node, bond_ops in bonding_nodes.items():
-            mol_graph.nodes[node]['hcount'] -= len(bond_ops)
+            print(bond_ops)
+            print(sum([int(bond[-1]) for bond in bond_ops]))
+            mol_graph.nodes[node]['hcount'] -= sum([int(bond[-1]) for bond in bond_ops])
 
     # now we add the hydrogen atoms
     pysmiles.smiles_helper.add_explicit_hydrogens(mol_graph)

cgsmiles/tests/test_utils.py

@@ -0,0 +1,32 @@
+import re
+import pytest
+import cgsmiles
+
+err_msg_rebuild_h = ("Likely you are writing an aromatic molecule that does not "
+                     "show delocalization-induced molecular equivalency and thus "
+                     "is not considered aromatic. For example, 4-methyl imidazole "
+                     "is often written as [nH]1cc(nc1)C, but should be written as "
+                     "[NH]1C=C(N=C1)C. A corresponding CGSmiles string would be "
+                     "{[#A]1[#B][#C]1}.{#A=[>][<]N,#B=[$]N=C[>],#C=[$]C(C)=C[<]}")
+
+@pytest.mark.parametrize('frag_str, hatoms_ref, error_type, err_msg', (
+                        ('{#A=[$]CCC[$]}', 6, None, None),
+                        ('{#A=CCC}', 8, None, None),
+                        ('{#A=C[!]CC}', 7, None, None),
+                        ('{#A=[$]=CCC=[$]}', 4, None, None),
+                        ('{#A=[$]cccc}',5, None, None),
+                        ('{#A=[$]ccc}', 0, SyntaxError, err_msg_rebuild_h),
+))
+def test_rebuild_hatoms(frag_str, hatoms_ref, error_type, err_msg):
+    frag_dict = cgsmiles.read_fragments(frag_str)
+    frag_graph = frag_dict['A']
+    if error_type:
+        with pytest.raises(error_type, match=re.escape(err_msg)):
+            cgsmiles.pysmiles_utils.rebuild_h_atoms(frag_graph, keep_bonding=True)
+    else:
+        cgsmiles.pysmiles_utils.rebuild_h_atoms(frag_graph, keep_bonding=True)
+        hatoms = 0
+        for node, ele in frag_graph.nodes(data='element'):
+            if ele == 'H':
+                hatoms += 1
+        assert hatoms == hatoms_ref