Fixed hydrogen bug and disabled evaluation test;

JonasSchaub · Feb 1, 2024 · 9f4f57c · 9f4f57c
1 parent 6214bac
commit 9f4f57c
Show file tree

Hide file tree

Showing 3 changed files with 24 additions and 5 deletions.
diff --git a/src/main/java/org/openscience/cdk/tools/ErtlFunctionalGroupsFinder.java b/src/main/java/org/openscience/cdk/tools/ErtlFunctionalGroupsFinder.java
@@ -761,12 +761,12 @@ private void markAtoms(IAtomContainer aMolecule) {
                 // if H...
                 // convert to implicit H
                 IAtom tmpConnectedAtom;
-                try {
+                if (this.adjListCache[idx].length > 0) {
                     tmpConnectedAtom = aMolecule.getAtom(this.adjListCache[idx][0]);
-                } catch(ArrayIndexOutOfBoundsException anException) {
-                    //TODO: this happens too often to ignore, investigate!
-                    ErtlFunctionalGroupsFinder.LOGGING_TOOL.warn("Explicit H was included in atom count but not correctly in adjacency list");
-                    break;
+                } else {
+                    //unconnected, explicit hydrogen atoms (like e.g. in CHEBI:365445) have an array of bond partners of size 0
+                    // nothing to do about them, but they also do not concern us
+                    continue;
                 }
                 if (Objects.isNull(tmpConnectedAtom.getImplicitHydrogenCount())) {
                     tmpConnectedAtom.setImplicitHydrogenCount(1);

diff --git a/src/test/java/org/openscience/cdk/tools/ErtlFunctionalGroupsFinderEvaluationTest.java b/src/test/java/org/openscience/cdk/tools/ErtlFunctionalGroupsFinderEvaluationTest.java
@@ -22,6 +22,7 @@
 
 import org.junit.jupiter.api.Assertions;
 import org.junit.jupiter.api.Assumptions;
+import org.junit.jupiter.api.Disabled;
 import org.junit.jupiter.api.Test;
 import org.openscience.cdk.Atom;
 import org.openscience.cdk.CDKConstants;
@@ -93,6 +94,7 @@
  * @author Jonas Schaub
  * @version 1.3
  */
+@Disabled("Legacy code and lengthy analyses")
 public class ErtlFunctionalGroupsFinderEvaluationTest {
 
     //<editor-fold defaultstate="collapsed" desc="Private static final constants">

diff --git a/src/test/java/org/openscience/cdk/tools/ErtlFunctionalGroupsFinderTest.java b/src/test/java/org/openscience/cdk/tools/ErtlFunctionalGroupsFinderTest.java
@@ -537,6 +537,23 @@ public void testRAtoms1() throws Exception {
         this.testFind(tmpMoleculeSmiles, tmpExpectedFGs);
     }
     //
+    /**
+     * Tests that a former bug concerning unconnected, explicit hydrogen atoms does not occur anymore.
+     *
+     * @throws Exception if anything goes wrong
+     * @author Jonas Schaub
+     */
+    @Test
+    public void testHydrogenBug() throws Exception {
+        String tmpMoleculeSmiles = "[H+].[H+].[O-]C(=O)\\C=C/C([O-])=O.[H][C@@]12Cc3c[nH]c4cccc(C1=C[C@@H](COC(=O)C1CCCCC1)CN2C)c34"; //CHEBI:365445
+        String[] tmpExpectedFGs = new String[]{"O=C([O-])[C]=[C]C(=O)[O-]", "[C]=[C]", "*OC(*)=O", "[R]N([R])[R]", "NarR3"};
+        this.testFind(tmpMoleculeSmiles, tmpExpectedFGs);
+
+        tmpMoleculeSmiles = "[HH].O=C1N([C@H](C)C(C1=C(O)[C@]2([C@]3([C@H](C=C([C@H]2[C@@H](C(=O)O)CC)C)C[C@H](C)CC3)C)C)=O)C"; //CHEBI:223373
+        tmpExpectedFGs = new String[]{"*C(=O)C(=[C]O[H])C(=O)N(*)*", "[C]=[C]", "*C(=O)O[H]"};
+        this.testFind(tmpMoleculeSmiles, tmpExpectedFGs);
+    }
+    //
     /**
      * Applies EFGF to detect functional groups in the given molecule and compares the identified FG to the given
      * expected FG, using i.a. an identity search. Note that the order of the given FG must match the order of the detected