fix: correct pubchem crawling from smiles with multiple entries

pytoda/preprocessing/crawlers.py

@@ -128,9 +128,7 @@ def get_smiles_from_pubchem(
             path = '{}/{}/{}/{}/{}/{}'.format(
                 PUBCHEM_START, query_type, drug, PUBCHEM_MID, option, PUBCHEM_END
             )
-            smiles = (
-                urllib_request.urlopen(path).read().decode('UTF-8').replace('\n', '')
-            )
+            smiles = urllib_request.urlopen(path).read().decode('UTF-8').split()[0]
             if not kekulize:
                 smiles = Chem.MolToSmiles(Chem.MolFromSmiles(smiles, sanitize=sanitize))
             return smiles

pytoda/preprocessing/tests/test_crawlers.py

@@ -1,4 +1,5 @@
 """Testing Crawlers."""
+
 import unittest
 
 from pytoda.preprocessing.crawlers import (  # query_pubchem,; remove_pubchem_smiles,
@@ -31,7 +32,7 @@ def test_get_smiles_from_zinc(self) -> None:
         self.assertEqual(smiles, ground_truth)
 
     def test_get_smiles_from_pubchem(self) -> None:
-        """Test get_smiles_from_zinc"""
+        """Test get_smiles_from_pubchem"""
 
         for sanitize in [True, False]:
 
@@ -83,6 +84,14 @@ def test_get_smiles_from_pubchem(self) -> None:
                 )
                 self.assertEqual(smiles, ground_truth)
 
+        # Test molecule where landing page has several entries
+        gt_smiles = (
+            'CC12C(C(CC(O1)N3C4=CC=CC=C4C5=C6C(=C7C8=CC=CC=C8N2C7=C53)CNC6=O)NC)OC'
+        )
+        drug = 'Staurosporine'
+        smiles = get_smiles_from_pubchem(drug, use_isomeric=False, kekulize=True)
+        self.assertEqual(smiles, gt_smiles)
+
     def test_query_pubchem(self) -> None:
         """Test query_pubchem"""
         # pass