fix(ENGKNOW-3981): refbases below position 1 returns N instead of throwing

gortools/src/test/java/gorsat/parser/UTestGenomicFunctions.java

@@ -89,6 +89,19 @@ public void testRefBases() {
         Assert.assertEquals("cct", TestUtils.getCalculatedWithArgs("refbases('chr1', 10101, 10103)", new String[]{"-config", "../tests/data/ref_mini/gor_config.txt"}));
     }
 
+    @Test
+    public void testRefBasesBelowPositionOne() {
+        String[] args = new String[]{"gorrow chr17,0 | calc r refbases(chrom,pos,pos+3)", "-config", "../tests/data/ref_mini/gor_config.txt"};
+        String lines = TestUtils.runGorPipe(args);
+        Assert.assertEquals("chrom\tpos\tr\nchr17\t0\tNAAG\n", lines);
+    }
+
+    @Test
+    public void testRefBasesWithBuildBelowPositionOne() {
+        String lines = TestUtils.runGorPipe("gorrow chr17,0 | calc r refbases_with_build(chrom,pos,pos+3,'../tests/data/ref_mini/chromSeq')");
+        Assert.assertEquals("chrom\tpos\tr\nchr17\t0\tNAAG\n", lines);
+    }
+
     @Test
     public void testRefBasesWithBuildUnQuoted() {
         TestUtils.assertCalculated("refbases_with_build('chr1', 10101, 10103, ../tests/data/ref_mini/chromSeq)", "cct");

model/src/main/scala/org/gorpipe/model/gor/iterators/RefSeqFromChromSeq.scala

@@ -114,6 +114,8 @@ class RefSeqFromChromSeq(ipath : String, fileReader : FileReader) extends RefSeq
 
   def getBase(chr: String, pos: Int): Char = {
     if (noReferenceBuildFound) return 'N'
+    // Positions below 1 are outside the chromosome, same as positions past its end.
+    if (pos < 1) return 'N'
     try {
       val (buffKey, offset) = getKeyAndOffset(chr, pos)
 
@@ -165,6 +167,12 @@ class RefSeqFromChromSeq(ipath : String, fileReader : FileReader) extends RefSeq
   }
 
   def getBases(chr: String, pos1: Int, pos2: Int): String = {
+    // Positions below 1 are outside the chromosome and read as 'N', same as positions past its end.
+    // Handle them before the buffer lookup: (pos - 1) / buffLength truncates toward zero, so pos 0 maps to buffer 0.
+    if (pos1 < 1) {
+      val leading = "N" * (math.min(pos2, 0) - pos1 + 1)
+      return if (pos2 < 1) leading else leading + getBases(chr, 1, pos2)
+    }
     if (pos1 == pos2) return getBase(chr, pos1).toString
     if ((pos1 - 1) / buffLength == (pos2 - 1) / buffLength) {
       val (buffKey, offset) = getKeyAndOffset(chr, pos1)

model/src/test/java/org/gorpipe/model/gor/iterators/UTestRefSeqFromChromSeq.java

@@ -79,6 +79,46 @@ public void testGetRefbases() {
 
     }
 
+    // Positions below 1 (e.g. pos 0 from an unmapped liftover row) read as 'N', like positions past the chromosome end.
+    @Test
+    public void testGetRefbasesBelowPositionOne() {
+        String path = "../tests/data/ref_mini/chromSeq";
+
+        // chr17 in ref_mini starts with real bases (AAGC...).
+        RefSeqFromChromSeq refseq = new RefSeqFromChromSeq(path, new DriverBackedFileReader(""));
+        Assert.assertEquals("NAAGC", refseq.getBases("chr17", 0, 4));
+
+        // Same, with the buffer already loaded.
+        Assert.assertEquals("NAAGC", refseq.getBases("chr17", 0, 4));
+
+        // Fresh refseq, no buffer loaded yet.
+        refseq = new RefSeqFromChromSeq(path, new DriverBackedFileReader(""));
+        Assert.assertEquals("NNNNN", refseq.getBases("chr17", -5, -1));
+
+        Assert.assertEquals("N", refseq.getBases("chr17", 0, 0));
+        Assert.assertEquals('N', refseq.getBase("chr17", 0));
+        Assert.assertEquals('N', refseq.getBase("chr17", -1));
+        Assert.assertEquals('N', refseq.getBase("chr17", -10001));
+
+        // Starting below 1 and crossing the buffer boundary (10000).
+        String expectedTail = new RefSeqFromChromSeq(path, new DriverBackedFileReader("")).getBases("chr17", 1, 10002);
+        refseq = new RefSeqFromChromSeq(path, new DriverBackedFileReader(""));
+        String bases = refseq.getBases("chr17", -2, 10002);
+        Assert.assertEquals(10005, bases.length());
+        Assert.assertEquals("NNN" + expectedTail, bases);
+        Assert.assertTrue(bases.startsWith("NNNAAGC"));
+        Assert.assertTrue(bases.endsWith("aactctt"));
+
+        // Starting more than a buffer below 1.
+        refseq = new RefSeqFromChromSeq(path, new DriverBackedFileReader(""));
+        Assert.assertEquals("N".repeat(10006) + "AA", refseq.getBases("chr17", -10005, 2));
+
+        // Positive positions are unchanged.
+        Assert.assertEquals("AAGC", refseq.getBases("chr17", 1, 4));
+        Assert.assertEquals('A', refseq.getBase("chr17", 1));
+        Assert.assertEquals("aactcttgac", refseq.getBases("chr17", 9996, 10005));
+    }
+
     @Ignore("Run manually to test from same buffer optimization")
     @Test
     public void testGetRefbasesPerformance() {