Updated HIN reader to fix bug 2984581

src/main/org/openscience/cdk/io/HINReader.java

@@ -20,18 +20,6 @@
  */
 package org.openscience.cdk.io;
 
-import java.io.BufferedReader;
-import java.io.IOException;
-import java.io.InputStream;
-import java.io.InputStreamReader;
-import java.io.Reader;
-import java.io.StringReader;
-import java.util.ArrayList;
-import java.util.List;
-import java.util.StringTokenizer;
-
-import javax.vecmath.Point3d;
-
 import org.openscience.cdk.CDKConstants;
 import org.openscience.cdk.annotations.TestClass;
 import org.openscience.cdk.annotations.TestMethod;
@@ -48,6 +36,17 @@
 import org.openscience.cdk.io.formats.HINFormat;
 import org.openscience.cdk.io.formats.IResourceFormat;
 
+import javax.vecmath.Point3d;
+import java.io.BufferedReader;
+import java.io.IOException;
+import java.io.InputStream;
+import java.io.InputStreamReader;
+import java.io.Reader;
+import java.io.StringReader;
+import java.util.ArrayList;
+import java.util.List;
+import java.util.StringTokenizer;
+
 /**
  * Reads an object from HIN formated input.
  *
@@ -197,7 +196,7 @@ private IChemFile readChemFile(IChemFile file) {
                 // read data for current molecule
                 int atomSerial = 0;
                 while (true) {
-                    if (line.indexOf("endmol ") >= 0) {
+                    if (line == null || line.indexOf("endmol ") >= 0) {
                         break;
                     }
                     if (line.indexOf(';') == 0) continue; // comment line
@@ -257,7 +256,15 @@ private IChemFile readChemFile(IChemFile file) {
                         m.addBond(file.getBuilder().newBond(s, e, bo));
                 }
                 setOfMolecules.addMolecule(m);
-                line = input.readLine(); // read in the 'mol N'
+
+                // we may not get a 'mol N' immediately since
+                // the aromaticring keyword might be present
+                // and doesn't seem to be located within the molecule
+                // block
+                while (true) {
+                    line = input.readLine();
+                    if (line == null || line.indexOf("mol ") == 0) break;
+                }
             }
 
             // got all the molecule in the HIN file (hopefully!)

src/test/data/hin/bug2984581.hin

@@ -0,0 +1,31 @@
+forcefield mm+
+sys 0 0 1
+view 40 0.1471 55 15 0.5586844 -0.7573253 0.3381274 0.8112321 0.583803 -0.03280994 -0.172552 0.2926302 0.9405283 -1.5498 -0.69331 -55.277
+seed 0
+mol 1
+atom 1 - C CA - -0.1921654 0.5444375 -0.480369 0.1121711 3 3 a 2 a 14 s
+atom 2 - C CA - -0.07865381 1.48883 -1.507176 0.1121711 3 5 a 1 a 15 s
+atom 3 - C CA - 0.1260283 0.9869497 0.8709679 0.1121711 3 1 a 4 a 7 s
+atom 4 - C CA - -0.1945596 2.357387 1.141245 0.1121711 3 3 a 6 a 20 s
+atom 5 - C CA - -0.1604924 2.854465 -1.227169 0.1121711 3 2 a 6 a 21 s
+atom 6 - C CA - -0.08668089 3.281385 0.09564244 0.1121711 3 5 a 4 a 16 s
+atom 7 - O O2 - -0.1584311 -0.06161446 1.767233 0.1121711 2 3 s 8 s
+atom 8 - C CA - 0.09127665 -0.009773561 3.152164 0.1121711 3 13 a 7 s 9 a
+atom 9 - C CA - -0.1865864 1.073688 4.026796 0.1121711 3 10 a 22 s 8 a
+atom 10 - C CA - -0.1031656 0.8394372 5.407645 0.1121711 3 9 a 11 a 17 s
+atom 11 - C CA - -0.153739 -0.452265 5.91707 0.1121711 3 10 a 12 a 18 s
+atom 12 - C CA - -0.1025772 -1.539669 5.037201 0.1121711 3 11 a 13 a 19 s
+atom 13 - C CA - -0.1454592 -1.331929 3.667263 0.1121711 3 12 a 8 a 23 s
+atom 14 - Br BR - 0.07038784 -1.265702 -0.9827618 0.1121711 1 1 s
+atom 15 - H H - 0.149111 1.15467 -2.557919 0.1121711 1 2 s
+atom 16 - H H - 0.1373875 4.357067 0.3297046 0.1121711 1 6 s
+atom 17 - H H - 0.1337775 1.701587 6.091895 0.1121711 1 10 s
+atom 18 - H H - 0.1358719 -0.6226702 7.002957 0.1121711 1 11 s
+atom 19 - H H - 0.1372966 -2.566783 5.432249 0.1121711 1 12 s
+atom 20 - H H - 0.145665 2.75555 2.159567 0.1121711 1 4 s
+atom 21 - H H - 0.1408581 3.583701 -2.04989 0.1121711 1 5 s
+atom 22 - H H - 0.1395921 2.117928 3.702678 0.1121711 1 9 s
+atom 23 - H H - 0.1552575 -2.176271 2.961693 0.1121711 1 13 s
+endmol 1
+aromaticring 6 1 1 1 3 1 4 1 6 1 5 1 2
+aromaticring 6 1 8 1 13 1 12 1 11 1 10 1 9

src/test/org/openscience/cdk/io/HINReaderTest.java

@@ -27,20 +27,19 @@
  *  */
 package org.openscience.cdk.io;
 
-import java.io.InputStream;
-import java.util.List;
-
 import org.junit.Assert;
 import org.junit.BeforeClass;
 import org.junit.Test;
 import org.openscience.cdk.ChemFile;
 import org.openscience.cdk.ChemObject;
-import org.openscience.cdk.exception.CDKException;
 import org.openscience.cdk.interfaces.IAtomContainer;
 import org.openscience.cdk.interfaces.IChemFile;
 import org.openscience.cdk.tools.LoggingTool;
 import org.openscience.cdk.tools.manipulator.ChemFileManipulator;
 
+import java.io.InputStream;
+import java.util.List;
+
 /**
  * TestCase for the reading HIN mol files using one test file.
  *
@@ -139,4 +138,18 @@ public class HINReaderTest extends SimpleChemObjectReaderTest {
         Assert.assertEquals(57, ac.getAtomCount());
         Assert.assertEquals(59, ac.getBondCount());
     }
+
+    /**
+     * @cdk.bug 2984581
+     * @throws Exception
+     */
+    @Test public void testAromaticRingsLine() throws Exception {
+        String filename = "data/hin/bug2984581.hin";
+        InputStream ins = this.getClass().getClassLoader().getResourceAsStream(filename);
+        ISimpleChemObjectReader reader = new HINReader(ins);
+        IChemFile content = (IChemFile) reader.read(new ChemFile());
+        List<IAtomContainer> cList = ChemFileManipulator.getAllAtomContainers(content);
+        Assert.assertEquals(1, cList.size());
+        IAtomContainer ac = cList.get(0);        
+    }
 }