Ignore prefix underscore on R labels.

cdk · Dec 3, 2018 · 23baae9 · 23baae9
1 parent 9e7cb8f
commit 23baae9
Show file tree

Hide file tree

Showing 2 changed files with 18 additions and 1 deletion.
diff --git a/storage/smiles/src/main/java/org/openscience/cdk/smiles/CxSmilesParser.java b/storage/smiles/src/main/java/org/openscience/cdk/smiles/CxSmilesParser.java
@@ -78,9 +78,15 @@ private static boolean processAtomLabels(final CharIter iter, final Map<Integer,
                 return true;
             } else {
                 iter.pos--; // push back
-                final int beg = iter.pos;
+                int beg = iter.pos;
                 int rollback = beg;
                 while (iter.hasNext()) {
+
+                    if (iter.pos == beg && iter.curr() == '_' &&
+                        iter.peek() == 'R') {
+                        ++beg;
+                    }
+
                     // correct step over of escaped label
                     if (iter.curr() == '&') {
                         rollback = iter.pos;
@@ -677,6 +683,10 @@ char next() {
             return str.charAt(pos++);
         }
 
+        public char peek() {
+            return pos < str.length() ? str.charAt(pos+1) : '\0';
+        }
+
 
         /**
          * Access a substring from the iterator.

diff --git a/storage/smiles/src/test/java/org/openscience/cdk/smiles/CxSmilesParserTest.java b/storage/smiles/src/test/java/org/openscience/cdk/smiles/CxSmilesParserTest.java
@@ -88,6 +88,13 @@ public void atomLabelsTruncated3() {
         assertThat(CxSmilesParser.processCx("|$;;;Het;$", state), is(-1));
     }
 
+    @Test
+    public void removeUnderscore() {
+        CxSmilesState state = new CxSmilesState();
+        CxSmilesParser.processCx("|$;;;_R1;$|", state);
+        assertThat(state.atomLabels.get(3), is("R1"));
+    }
+
     @Test
     public void skipCis() {
         CxSmilesState state = new CxSmilesState();