JAL-3140 NCList now via IntervalStore library

[jalview.git] / src / jalview / datamodel / Sequence.java

diff --git a/src/jalview/datamodel/Sequence.java b/src/jalview/datamodel/Sequence.java
index 441d8d0..674fabe 100755 (executable)
--- a/src/jalview/datamodel/Sequence.java
+++ b/src/jalview/datamodel/Sequence.java
@@ -34,11 +34,14 @@ import java.util.Arrays;
 import java.util.BitSet;
 import java.util.Collections;
 import java.util.Enumeration;
+import java.util.Iterator;
 import java.util.List;
 import java.util.ListIterator;
 import java.util.Vector;
 
 import fr.orsay.lri.varna.models.rna.RNA;
+import intervalstore.api.IntervalI;
+import intervalstore.impl.Range;
 
 /**
  * 
@@ -408,7 +411,7 @@ public class Sequence extends ASequence implements SequenceI
   {
     if (pdbIds == null)
     {
-      pdbIds = new Vector<PDBEntry>();
+      pdbIds = new Vector<>();
       pdbIds.add(entry);
       return true;
     }
@@ -444,7 +447,7 @@ public class Sequence extends ASequence implements SequenceI
   @Override
   public Vector<PDBEntry> getAllPDBEntries()
   {
-    return pdbIds == null ? new Vector<PDBEntry>() : pdbIds;
+    return pdbIds == null ? new Vector<>() : pdbIds;
   }
 
   /**
@@ -814,7 +817,7 @@ public class Sequence extends ASequence implements SequenceI
    * @param curs
    * @return
    */
-  protected int findIndex(int pos, SequenceCursor curs)
+  protected int findIndex(final int pos, SequenceCursor curs)
   {
     if (!isValidCursor(curs))
     {
@@ -840,10 +843,15 @@ public class Sequence extends ASequence implements SequenceI
     while (newPos != pos)
     {
       col += delta; // shift one column left or right
-      if (col < 0 || col == sequence.length)
+      if (col < 0)
       {
         break;
       }
+      if (col == sequence.length)
+      {
+        col--; // return last column if we failed to reach pos
+        break;
+      }
       if (!Comparison.isGap(sequence[col]))
       {
         newPos += delta;
@@ -851,7 +859,14 @@ public class Sequence extends ASequence implements SequenceI
     }
 
     col++; // convert back to base 1
-    updateCursor(pos, col, curs.firstColumnPosition);
+
+    /*
+     * only update cursor if we found the target position
+     */
+    if (newPos == pos)
+    {
+      updateCursor(pos, col, curs.firstColumnPosition);
+    }
 
     return col;
   }
@@ -1056,7 +1071,7 @@ public class Sequence extends ASequence implements SequenceI
    * {@inheritDoc}
    */
   @Override
-  public Range findPositions(int fromColumn, int toColumn)
+  public IntervalI findPositions(int fromColumn, int toColumn)
   {
     if (toColumn < fromColumn || fromColumn < 1)
     {
@@ -1128,6 +1143,27 @@ public class Sequence extends ASequence implements SequenceI
     return map;
   }
 
+  /**
+   * Build a bitset corresponding to sequence gaps
+   * 
+   * @return a BitSet where set values correspond to gaps in the sequence
+   */
+  @Override
+  public BitSet gapBitset()
+  {
+    BitSet gaps = new BitSet(sequence.length);
+    int j = 0;
+    while (j < sequence.length)
+    {
+      if (jalview.util.Comparison.isGap(sequence[j]))
+      {
+        gaps.set(j);
+      }
+      j++;
+    }
+    return gaps;
+  }
+
   @Override
   public int[] findPositionMap()
   {
@@ -1151,7 +1187,7 @@ public class Sequence extends ASequence implements SequenceI
   @Override
   public List<int[]> getInsertions()
   {
-    ArrayList<int[]> map = new ArrayList<int[]>();
+    ArrayList<int[]> map = new ArrayList<>();
     int lastj = -1, j = 0;
     int pos = start;
     int seqlen = sequence.length;
@@ -1462,7 +1498,7 @@ public class Sequence extends ASequence implements SequenceI
   {
     if (this.annotation == null)
     {
-      this.annotation = new Vector<AlignmentAnnotation>();
+      this.annotation = new Vector<>();
     }
     if (!this.annotation.contains(annotation))
     {
@@ -1629,7 +1665,7 @@ public class Sequence extends ASequence implements SequenceI
       return null;
     }
 
-    Vector<AlignmentAnnotation> subset = new Vector<AlignmentAnnotation>();
+    Vector<AlignmentAnnotation> subset = new Vector<>();
     Enumeration<AlignmentAnnotation> e = annotation.elements();
     while (e.hasMoreElements())
     {
@@ -1763,7 +1799,7 @@ public class Sequence extends ASequence implements SequenceI
   public List<AlignmentAnnotation> getAlignmentAnnotations(String calcId,
           String label)
   {
-    List<AlignmentAnnotation> result = new ArrayList<AlignmentAnnotation>();
+    List<AlignmentAnnotation> result = new ArrayList<>();
     if (this.annotation != null)
     {
       for (AlignmentAnnotation ann : annotation)
@@ -1819,7 +1855,7 @@ public class Sequence extends ASequence implements SequenceI
     }
     synchronized (dbrefs)
     {
-      List<DBRefEntry> primaries = new ArrayList<DBRefEntry>();
+      List<DBRefEntry> primaries = new ArrayList<>();
       DBRefEntry[] tmp = new DBRefEntry[1];
       for (DBRefEntry ref : dbrefs)
       {
@@ -1961,4 +1997,73 @@ public class Sequence extends ASequence implements SequenceI
 
     return count;
   }
+
+  @Override
+  public String getSequenceStringFromIterator(Iterator<int[]> it)
+  {
+    StringBuilder newSequence = new StringBuilder();
+    while (it.hasNext())
+    {
+      int[] block = it.next();
+      if (it.hasNext())
+      {
+        newSequence.append(getSequence(block[0], block[1] + 1));
+      }
+      else
+      {
+        newSequence.append(getSequence(block[0], block[1]));
+      }
+    }
+
+    return newSequence.toString();
+  }
+
+  @Override
+  public int firstResidueOutsideIterator(Iterator<int[]> regions)
+  {
+    int start = 0;
+
+    if (!regions.hasNext())
+    {
+      return findIndex(getStart()) - 1;
+    }
+
+    // Simply walk along the sequence whilst watching for region
+    // boundaries
+    int hideStart = getLength();
+    int hideEnd = -1;
+    boolean foundStart = false;
+
+    // step through the non-gapped positions of the sequence
+    for (int i = getStart(); i <= getEnd() && (!foundStart); i++)
+    {
+      // get alignment position of this residue in the sequence
+      int p = findIndex(i) - 1;
+
+      // update region start/end
+      while (hideEnd < p && regions.hasNext())
+      {
+        int[] region = regions.next();
+        hideStart = region[0];
+        hideEnd = region[1];
+      }
+      if (hideEnd < p)
+      {
+        hideStart = getLength();
+      }
+      // update boundary for sequence
+      if (p < hideStart)
+      {
+        start = p;
+        foundStart = true;
+      }
+    }
+
+    if (foundStart)
+    {
+      return start;
+    }
+    // otherwise, sequence was completely hidden
+    return 0;
+  }
 }