Assign numbers to unlabelled atoms (i.e. hydrogens). Beam will handle…

… the correct hydrogen ordering (CDK API makes this difficult).

Signed-off-by: Stephan Beisken <sbeisken@gmail.com>
Signed-off-by: Egon Willighagen <egonw@users.sourceforge.net>

src/main/org/openscience/cdk/graph/invariant/InChINumbersTools.java

@@ -92,6 +92,11 @@ public static long[] getUSmilesNumbers(IAtomContainer container) throws CDKExcep
      * charged oxygen atom, start instead at any carbonyl oxygen attached to the
      * same neighbour." <p/>
      * 
+     * All unlabelled atoms (e.g. hydrogens) are assigned the same label which
+     * is different but larger then all other labels. The hydrogen
+     * labelling then needs to be adjusted externally as universal SMILES 
+     * suggests hydrogens should be visited first.
+     * 
      * @param aux       inchi AuxInfo
      * @param container the structure to obtain the numbering of           
      * @return the numbers string to use
@@ -102,8 +107,8 @@ static long[] parseUSmilesNumbers(String aux, IAtomContainer container) {
         int index;
         long[] numbers = new long[container.getAtomCount()];
         int[]  first   = null;
-        int label = 1;
-
+        int    label   = 1;
+                
         if ((index = aux.indexOf("/R:")) >= 0) { // reconnected metal numbers
             String[] baseNumbers = aux.substring(index + 8, aux.indexOf('/', index + 8)).split(";");
             first = new int[baseNumbers.length];
@@ -176,6 +181,11 @@ static long[] parseUSmilesNumbers(String aux, IAtomContainer container) {
                 }
             }
         }
+
+        // assign unlabelled atoms
+        for (int i = 0; i < numbers.length; i++)
+            if (numbers[i] == 0)
+                numbers[i] = label;
 
         return numbers;
     }

src/test/org/openscience/cdk/graph/invariant/InChINumbersToolsTest.java

@@ -157,6 +157,12 @@ public void fixedH() throws Exception {
                    is(new long[]{3, 2, 1}));
     }
 
+    @Test public void unlabelledHydrogens() throws Exception {
+        IAtomContainer container = new SmilesParser(SilentChemObjectBuilder.getInstance()).parseSmiles("[H]C([H])([H])[H]");
+        assertThat(InChINumbersTools.getUSmilesNumbers(container),
+                   is(new long[]{2, 1, 2, 2, 2}));   
+    }
+
     static IAtomContainer mock(int nAtoms) {
         IAtomContainer container = Mockito.mock(IAtomContainer.class);
         when(container.getAtomCount()).thenReturn(nAtoms);