Merge pull request #503 from datamol-io/ordered_true

MAJOR-BUG for node/edge predictions: Added ordered=True to `to_mol`

Author: DomInvivo

graphium/data/smiles_transform.py

@@ -27,7 +27,9 @@ def smiles_to_unique_mol_id(smiles: str) -> Optional[str]:
         mol_id: a string unique ID
     """
     try:
-        mol = dm.to_mol(mol=smiles)
+        mol = dm.to_mol(
+            mol=smiles
+        )  # Doesn't need `ordered=True` because the unique_id doesn't depend on the atom order
         mol_id = dm.unique_id(mol)
     except:
         mol_id = ""

graphium/features/featurizer.py

@@ -743,7 +743,7 @@ def mol_to_adj_and_features(
     """
 
     if isinstance(mol, str):
-        mol = dm.to_mol(mol)
+        mol = dm.to_mol(mol, ordered=True)
 
     # Add or remove explicit hydrogens
     if explicit_H:
@@ -1071,7 +1071,7 @@ def mol_to_graph_dict(
     input_mol = mol
     try:
         if isinstance(mol, str):
-            mol = dm.to_mol(mol)
+            mol = dm.to_mol(mol, ordered=True)
         if explicit_H:
             mol = Chem.AddHs(mol)
         else:

graphium/features/properties.py

@@ -18,6 +18,7 @@
 import datamol as dm
 
 from rdkit.Chem import rdMolDescriptors as rdMD
+from loguru import logger
 
 
 def get_prop_or_none(
@@ -33,6 +34,7 @@ def get_prop_or_none(
     Returns:
         The property or a list of `None` with lenght `n`.
     """
+    logger.warning("get_prop_or_none is deprecated. Use `datamol.to_fp` instead.")
     try:
         return prop(*args, **kwargs)
     except RuntimeError:
@@ -75,8 +77,12 @@ def get_props_from_mol(
 
     """
 
+    logger.warning("get_props_from_mol is deprecated. Use `datamol.to_fp` instead.")
+
     if isinstance(mol, str):
-        mol = dm.to_mol(mol)
+        mol = dm.to_mol(
+            mol
+        )  # Doesn't need `ordered=True` because the fingerprints don't depend on the atom order
 
     if isinstance(properties, str):
         properties = [properties]

profiling/profile_mol_to_graph.py

@@ -67,7 +67,9 @@ def main():
 
     graphs = []
     for s in tqdm(smiles):
-        mol = dm.to_mol(s)
+        mol = dm.to_mol(
+            s
+        )  # Doesn't need `ordered=True` because this is just to test the speed of the featurizer
         graphs.append(mol_to_graph_dict(mol, **featurizer))
 
     print(graphs[0])