[2] Aromatic fragments #5
Conversation
| @@ -49,6 +49,7 @@ def merge_graphs(source_graph, target_graph, max_node=None): | |||
| for node1, node2 in target_graph.edges: | |||
| if correspondence[node1] != correspondence[node2]: | |||
| attrs = target_graph.edges[(node1, node2)] | |||
| print(attrs) | |||
There was a problem hiding this comment.
| print(attrs) |
| organic_subset = 'B C N O P S F Cl Br I * b c n o s p'.split() | ||
| batom = False | ||
| for idx, node in enumerate(smiles_str): | ||
| if node == '[': | ||
| batom = True | ||
| start = idx | ||
|
|
||
| if node == ']' and batom: | ||
| stop = idx+1 | ||
| batom = False | ||
| yield start, stop | ||
|
|
||
| if node in organic_subset and not batom: | ||
| yield idx, idx + 1 |
There was a problem hiding this comment.
This won't work for multi-letter organic atoms, such as Br!
There was a problem hiding this comment.
indeed; hence PR #4 🤣
|
|
||
| def strip_aromatic_nodes(smiles_str): | ||
| """ | ||
| Find all aromatic nodes and change them to lower |
There was a problem hiding this comment.
| Find all aromatic nodes and change them to lower | |
| Find all aromatic nodes and change them to upper |
| if node in organic_subset and not batom: | ||
| yield idx, idx + 1 | ||
|
|
||
| def strip_aromatic_nodes(smiles_str): |
There was a problem hiding this comment.
I don't like the name of the function very much, since it deals with the smiles string, and not with graph nodes
| @@ -102,7 +110,18 @@ def fragment_iter(fragment_str, all_atom=True): | |||
| mol_graph.add_node(0, element="H", bonding=bonding_descrpt[0]) | |||
| nx.set_node_attributes(mol_graph, bonding_descrpt, 'bonding') | |||
| elif all_atom: | |||
| mol_graph = pysmiles.read_smiles(smile) | |||
| try: | |||
| mol_graph = pysmiles.read_smiles(smile) | |||
There was a problem hiding this comment.
| mol_graph = pysmiles.read_smiles(smile) | |
| mol_graph = pysmiles.read_smiles(smile, reinterpret_aromatic=False) |
There was a problem hiding this comment.
Unfortunately, not. The error is raised already at the parsing stage of the cgsmiles string. You categorically do not allow non-cyclic aromatics. We could modify pysmiles though to allow this if 'reinterpret_aromatic' is set to False
| @@ -165,7 +165,8 @@ def edges_from_bonding_descrpt(self): | |||
| bonding descriptors that formed the edge. Later unconsumed | |||
| bonding descriptors are replaced by hydrogen atoms. | |||
| """ | |||
| for prev_node, node in nx.dfs_edges(self.meta_graph): | |||
| for prev_node, node in self.meta_graph.edges: | |||
| print(prev_node, node) | |||
There was a problem hiding this comment.
| print(prev_node, node) |
| order = re.findall("[-+]?[.]?[\d]+(?:,\d\d\d)*[\.]?\d*(?:[eE][-+]?\d+)?", bonding[0]) | ||
| print(order) |
There was a problem hiding this comment.
| order = re.findall("[-+]?[.]?[\d]+(?:,\d\d\d)*[\.]?\d*(?:[eE][-+]?\d+)?", bonding[0]) | |
| print(order) | |
| order = re.findall("[-+]?[.]?[\d]+(?:,\d\d\d)*[\.]?\d*(?:[eE][-+]?\d+)?", bonding[0]) |
Cool, this got more complicated. At least needs a comment, but we should probably finish the discussion on #1 about this first.
|
no need for this anymore; the recent pysmiles changes fix everything |
Pysmiles does not allow partial aromatic fragments (for good reason) yet we still want to do them. Thus here we have a workaround. If two atoms are marked as aromatic we simply convert them to regular elements and later change it back. It's stupid but works.
Example Case Martini3 p-cresol:
The aromatic ring is fragmented into four fragments.