From 04134a52ba4355211d96a1a883663508008a584b Mon Sep 17 00:00:00 2001
From: Mateusz Warowny <mmzwarowny@dundee.ac.uk>
Date: Wed, 13 Sep 2023 16:08:54 +0200
Subject: [PATCH] JAL-4241 add SeqsetUtils.filterSequence function

---
 src/jalview/analysis/SeqsetUtils.java      |   29 +++++++++++++++++++++++++++
 test/jalview/analysis/SeqsetUtilsTest.java |   30 ++++++++++++++++++++++++++++
 2 files changed, 59 insertions(+)

diff --git a/src/jalview/analysis/SeqsetUtils.java b/src/jalview/analysis/SeqsetUtils.java
index adb70e3..5420aff 100755
--- a/src/jalview/analysis/SeqsetUtils.java
+++ b/src/jalview/analysis/SeqsetUtils.java
@@ -30,6 +30,7 @@ import jalview.datamodel.SequenceFeature;
 import jalview.datamodel.SequenceI;
 
 import java.util.ArrayList;
+import java.util.BitSet;
 import java.util.Enumeration;
 import java.util.HashMap;
 import java.util.Hashtable;
@@ -40,6 +41,8 @@ import java.util.Optional;
 import java.util.Vector;
 import static java.lang.String.format;
 
+import java.nio.CharBuffer;
+
 public class SeqsetUtils
 {
   public static class SequenceInfo {
@@ -90,6 +93,32 @@ public class SeqsetUtils
   }
 
   /**
+   * Filter the sequence through the mask leaving only characters at positions
+   * where the mask value was true. The length of the resulting array is
+   * the cardinality of the mask from 0 to sequence length.
+   *
+   * @param sequence
+   *          input sequence
+   * @param mask
+   *          mask used to filter the sequence characters
+   * @return input array filtered through the mask
+   */
+  public static char[] filterSequence(char[] sequence, BitSet mask)
+  {
+    mask = mask.get(0, sequence.length);
+    char[] result = new char[mask.cardinality()];
+    for (int i = mask.nextSetBit(0), j = 0; i >= 0;)
+    {
+      result[j++] = sequence[i];
+      if (i == Integer.MAX_VALUE)
+        // prevents integer overflow of (i + 1)
+        break;
+      i = mask.nextSetBit(i + 1);
+    }
+    return result;
+  }
+
+  /**
    * Recover essential properties of a sequence from a hashtable TODO: replace
    * these methods with something more elegant.
    * 
diff --git a/test/jalview/analysis/SeqsetUtilsTest.java b/test/jalview/analysis/SeqsetUtilsTest.java
index 71e5bfd..fbd5812 100644
--- a/test/jalview/analysis/SeqsetUtilsTest.java
+++ b/test/jalview/analysis/SeqsetUtilsTest.java
@@ -27,11 +27,16 @@ import jalview.datamodel.SequenceFeature;
 import jalview.datamodel.SequenceI;
 import jalview.gui.JvOptionPane;
 
+import static org.hamcrest.MatcherAssert.assertThat;
+import static org.hamcrest.Matchers.equalTo;
+
+import java.util.BitSet;
 import java.util.Hashtable;
 import java.util.Map;
 
 import org.testng.Assert;
 import org.testng.annotations.BeforeClass;
+import org.testng.annotations.DataProvider;
 import org.testng.annotations.Test;
 
 /**
@@ -82,4 +87,29 @@ public class SeqsetUtilsTest
     Assert.assertSame(sqset[0].getSequenceFeatures().get(1),
             sqset2[0].getSequenceFeatures().get(1));
   }
+  
+  @DataProvider
+  public Object[][] sequenceAndMask()
+  {
+    return new Object[][] {
+      { "AAAABBBBCCCCDDDD", 0xFFFFL, "AAAABBBBCCCCDDDD" },
+      { "AAAABBBBCCCCDDDD", 0x000FL, "AAAA" },
+      { "---A---B---C---D", 0x8888L, "ABCD" },
+      { "---A---B---C---D", 0x9999L, "-A-B-C-D" },
+      { "ABCDABCDABCDABCD", 0xC5A3L, "ABBDACCD" },
+      { "", 0xFFFFL, "" },
+      { "AAAABBBBCCCCDDDD", 0x0000L, "" },
+      { "AAABBBCCC", 0xFFFF, "AAABBBCCC" },
+      { "AAAABBBB", 0xD000L, "" },
+      { "AAAABBBB", 0xAA0AL, "AA" },
+    };
+  }
+  
+  @Test(groups = {"Functional"}, dataProvider = "sequenceAndMask")
+  public void testFilterSequence(String sequence, long mask, String expected)
+  {
+    BitSet bitMask = BitSet.valueOf(new long[] {mask});
+    var result = SeqsetUtils.filterSequence(sequence.toCharArray(), bitMask);
+    assertThat(result, equalTo(expected.toCharArray()));
+  }
 }
-- 
1.7.10.2