From f21eb611044169a07a7d4c6e2f69dd772fa7c872 Mon Sep 17 00:00:00 2001
From: Jim Procter <jprocter@dundee.ac.uk>
Date: Thu, 12 Feb 2015 16:49:32 +0000
Subject: [PATCH] JAL-1350 cloned method and tested equivalence for using
 column selection to subdivide groups

---
 src/jalview/analysis/Grouping.java      |   82 ++++++++++++++++++++++++++++
 test/jalview/analysis/GroupingTest.java |   89 +++++++++++++++++++++++++++++++
 2 files changed, 171 insertions(+)
 create mode 100644 test/jalview/analysis/GroupingTest.java

diff --git a/src/jalview/analysis/Grouping.java b/src/jalview/analysis/Grouping.java
index 7b6dda1..ea10c9a 100644
--- a/src/jalview/analysis/Grouping.java
+++ b/src/jalview/analysis/Grouping.java
@@ -102,6 +102,88 @@ public class Grouping
   }
 
   /**
+   * Divide the given sequences based on the equivalence of characters at
+   * selected columns If exgroups is provided, existing groups will be
+   * subdivided.
+   * 
+   * @param sequences
+   * @param columnSelection
+   * @param list
+   * @return
+   */
+  public static SequenceGroup[] makeGroupsFromCols(SequenceI[] sequences,
+          ColumnSelection cs, List<SequenceGroup> list)
+  {
+    // TODO: determine how to get/recover input data for group generation
+    Map<String, List<SequenceI>> gps = new HashMap<String, List<SequenceI>>();
+    Map<String, SequenceGroup> pgroup = new HashMap<String, SequenceGroup>();
+    if (list != null)
+    {
+      for (SequenceGroup sg : list)
+      {
+        for (SequenceI sq : sg.getSequences(null))
+        {
+          pgroup.put(sq.toString(), sg);
+        }
+      }
+    }
+    int[] spos = new int[cs.getSelected().size()];
+    int width = -1;
+    int i = 0;
+    for (Integer pos : cs.getSelected())
+    {
+      spos[i++] = pos.intValue();
+    }
+    ;
+    for (i = 0; i < sequences.length; i++)
+    {
+      int slen = sequences[i].getLength();
+      if (width < slen)
+      {
+        width = slen;
+      }
+
+      SequenceGroup pgp = pgroup.get(((Object) sequences[i]).toString());
+      StringBuilder schar = new StringBuilder();
+      if (pgp != null)
+      {
+        schar.append(pgp.getName() + ":");
+      }
+      for (int p : spos)
+      {
+        if (p >= slen)
+        {
+          schar.append("~");
+        }
+        else
+        {
+          schar.append(sequences[i].getCharAt(p));
+        }
+      }
+      List<SequenceI> svec = gps.get(schar.toString());
+      if (svec == null)
+      {
+        svec = new ArrayList<SequenceI>();
+        gps.put(schar.toString(), svec);
+      }
+      svec.add(sequences[i]);
+    }
+    // make some groups
+    SequenceGroup[] groups = new SequenceGroup[gps.size()];
+    i = 0;
+    for (String key : gps.keySet())
+    {
+      SequenceGroup group = new SequenceGroup(gps.get(key), "Subseq: "
+              + key, null, true, true, false, 0, width - 1);
+
+      groups[i++] = group;
+    }
+    gps.clear();
+    pgroup.clear();
+    return groups;
+  }
+
+  /**
    * subdivide the given sequences based on the distribution of features
    * 
    * @param featureLabels
diff --git a/test/jalview/analysis/GroupingTest.java b/test/jalview/analysis/GroupingTest.java
new file mode 100644
index 0000000..61bd480
--- /dev/null
+++ b/test/jalview/analysis/GroupingTest.java
@@ -0,0 +1,89 @@
+package jalview.analysis;
+
+import static org.junit.Assert.fail;
+import jalview.datamodel.Alignment;
+import jalview.datamodel.AlignmentI;
+import jalview.datamodel.ColumnSelection;
+import jalview.datamodel.Sequence;
+import jalview.datamodel.SequenceGroup;
+import jalview.datamodel.SequenceI;
+
+import java.util.ArrayList;
+import java.util.Arrays;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+public class GroupingTest
+{
+  Sequence s1 = new Sequence("s1", "AAAADDDDEEEE");
+
+  Sequence s2 = new Sequence("s2", "AAAADDDDEEEE");
+
+  Sequence s3 = new Sequence("s3", "ACAADDEDEEEE");
+
+  Sequence s4 = new Sequence("s4", "AAAADDEDEEEE");
+
+  Sequence s5 = new Sequence("s5", "AAAADDEDTTEE");
+
+  SequenceGroup sg1 = new SequenceGroup(Arrays.asList(new SequenceI[]
+  { s1, s2 }), "Group1", null, false, false, false, 0, 5);
+
+  SequenceGroup sg2 = new SequenceGroup(Arrays.asList(new SequenceI[]
+  { s3, s4, s5 }), "Group2", null, false, false, false, 0, 5);
+
+  AlignmentI alignment = new Alignment(new SequenceI[]
+  { s1, s2, s3, s4, s5 });
+
+  int[] positions = new int[]
+  { 1, 7, 9 };
+
+  @Test
+  public void testMakeGroupsWithBoth()
+  {
+    ArrayList<String> str = new ArrayList<String>();
+    for (SequenceI s : alignment.getSequences())
+    {
+      StringBuilder sb = new StringBuilder();
+      for (int p : positions)
+      {
+        sb.append(s.getCharAt(p));
+      }
+      str.add(sb.toString());
+    }
+    SequenceGroup[] seqgroupsString = Grouping.makeGroupsFrom(
+            alignment.getSequencesArray(),
+            str.toArray(new String[str.size()]),
+            Arrays.asList(new SequenceGroup[]
+            { sg1, sg2 }));
+    ColumnSelection cs = new ColumnSelection();
+    for (int p : positions)
+    {
+      cs.addElement(p);
+    }
+    SequenceGroup[] seqgroupsColSel = Grouping.makeGroupsFromCols(
+            alignment.getSequencesArray(), cs,
+            Arrays.asList(new SequenceGroup[]
+            { sg1, sg2 }));
+    Assert.assertEquals(seqgroupsString.length, seqgroupsColSel.length);
+    for (int p = 0; p < seqgroupsString.length; p++)
+    {
+      Assert.assertEquals(seqgroupsString[p].getName(),
+              seqgroupsColSel[p].getName());
+      Assert.assertArrayEquals(
+              seqgroupsString[p].getSequencesInOrder(alignment),
+              seqgroupsColSel[p].getSequencesInOrder(alignment));
+      if (seqgroupsString[p].getSequences().contains(s2))
+      {
+        Assert.assertTrue(seqgroupsString[p].getSize() == 2);
+      }
+    }
+  }
+
+  @Test
+  public void testMakeGroupsFromCols()
+  {
+    fail("Not yet implemented"); // TODO
+  }
+
+}
-- 
1.7.10.2