Exposed canonical alg

README.md

@@ -57,10 +57,12 @@ where $M$ is the size in bits, $N$ is the number of compounds, and $\epsilon$ is
 ### Build with Python
 
 ```py
-from molbloom import CustomFilter
+from molbloom import CustomFilter, canon
 bf = CustomFilter(100, 1000, 'myfilter')
 bf.add('CCCO')
-
+# canonicalize one
+s = canon("CCCOC")
+bf.add(s)
 # save it
 bf.save('test.bloom')
 ```

molbloom/__init__.py

@@ -51,16 +51,20 @@ def _load_filter(name):
             _load_big_filter(name)
 
 
+def canon(smiles):
+    try:
+        from rdkit import Chem
+    except ImportError:
+        raise ImportError("To canonicalize SMILES, rdkit is required.")
+    return Chem.MolToSmiles(Chem.MolFromSmiles(smiles), canonical=True)
+
+
 def buy(smiles, instock=True, canonicalize=False):
     if instock:
         name = "instock"
     else:
         name = "zinc20"
     _load_filter(name)
     if canonicalize:
-        try:
-            from rdkit import Chem
-        except ImportError:
-            raise ImportError("To canonicalize SMILES, rdkit is required.")
-        smiles = Chem.MolToSmiles(Chem.MolFromSmiles(smiles), canonical=True)
+        smiles = canon(smiles)
     return smiles in filters[name]

tests/test_molbloom.py

@@ -71,6 +71,9 @@ def test_build_custom():
     """Build a custom filter"""
     bf = molbloom.CustomFilter(100, 1000, "test")
     bf.add("CCCO")
+    # canonicalize
+    s = molbloom.canon("CCCOC")
+    bf.add(s)
     assert "CCCO" in bf
     assert "CCCOO" not in bf
     bf.save("test.bloom")