fixing the dbSNP test. Also removing unnessasary comments from the GenomeLocParser, added some tests, and commented out the performance test

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@1676 348d0f76-0448-11de-a6fe-93d51630548a

java/src/org/broadinstitute/sting/utils/GenomeLocParser.java

@@ -52,7 +52,7 @@ import java.util.regex.Pattern;
 public class GenomeLocParser {
     private static Logger logger = Logger.getLogger(GenomeLocParser.class);
 
-    private static final Pattern mPattern = Pattern.compile("([\\w&&[^:]]+):([\\d,]+)?\\+?(-)?([\\d,]+)?$");  // matches case 3
+    private static final Pattern mPattern = Pattern.compile("([\\w&&[^:]]+):([\\d,]+)?(\\+)?(-)?([\\d,]+)?$");  // matches case 3
 
 
     // --------------------------------------------------------------------------------------------------------------
@@ -61,7 +61,7 @@ public class GenomeLocParser {
     //
     // --------------------------------------------------------------------------------------------------------------
     //public static Map<String, Integer> refContigOrdering = null;
-    private static SAMSequenceDictionary contigInfo = null;
+    protected static SAMSequenceDictionary contigInfo = null;
 
     /**
      * do we have a contig ordering setup?
@@ -140,11 +140,6 @@ public class GenomeLocParser {
     public static GenomeLoc parseGenomeLoc(final String str) {
         // 'chr2', 'chr2:1000000' or 'chr2:1,000,000-2,000,000'
         //System.out.printf("Parsing location '%s'%n", str);
-        /*try {
-            Thread.sleep(2000);
-        } catch (InterruptedException e) {
-            e.printStackTrace();  //To change body of catch statement use File | Settings | File Templates.
-        } */
         
         String contig = null;
         long start = 1;
@@ -156,29 +151,21 @@ public class GenomeLocParser {
         try {
             if (match.matches()) {
                 contig = match.group(1);
-                if (match.groupCount() > 1) {
+                if (match.groupCount() == 5) {
-                switch (match.groupCount()) {
-                    case 2:
-                        start = stop = parsePosition(match.group(2));
-                        break;
-                    case 3:
                     start = parsePosition(match.group(2));
-                        if (!match.group(3).equals("+")) bad = true;
+                    if (match.group(3) != null && match.group(3).equals("+") && match.group(5) == null) {
-                        break;
+                        // do nothing
-                    case 4:
+                    } else if (match.group(5) != null)
-                        start = parsePosition(match.group(2));
+                        stop = parsePosition(match.group(5));
-                        stop = parsePosition(match.group(4));
+                    else if (match.group(5) == null && match.group(4) == null && match.group(3) == null)
-                        break;
+                        stop = start;
-                    default:
+                    else
                         bad = true;
-                        break;
-
-                }
-                }
                 } else {
                     bad = true;
                 }
             }
+        }
 
         catch (Exception e) {
             bad = true;
@@ -194,6 +181,7 @@ public class GenomeLocParser {
 
         if (!isContigValid(contig))
             throw new MalformedGenomeLocException("Contig " + contig + " does not match any contig in the GATK sequence dictionary derived from the reference.");
+
         GenomeLoc loc = parseGenomeLoc(contig, start, stop);
         return loc;
     }

java/test/org/broadinstitute/sting/gatk/refdata/rodDbSNPTest.java

@@ -75,7 +75,7 @@ public class rodDbSNPTest extends BaseTest {
                 stop = (var.getLocation().getContig().equals("1") && var.getLocation().getStart() > 11000000);
                 line = stream.readLine();
             }
-            Assert.assertEquals(3717,snpCount);
+            Assert.assertEquals(3615,snpCount);
             Assert.assertEquals(9902,indelCount);
 
         } catch (IOException e) {

java/test/org/broadinstitute/sting/utils/GenomeLocParserTest.java

@@ -3,16 +3,11 @@ package org.broadinstitute.sting.utils;
 import static junit.framework.Assert.assertTrue;
 import net.sf.samtools.SAMFileHeader;
 import org.broadinstitute.sting.BaseTest;
-import org.broadinstitute.sting.gatk.GenomeAnalysisEngine;
 import org.broadinstitute.sting.utils.sam.ArtificialSAMUtils;
 import static org.junit.Assert.assertEquals;
+import org.junit.BeforeClass;
 import org.junit.Test;
 
-import java.util.Arrays;
-import java.util.Collections;
-import java.util.LinkedList;
-import java.util.List;
-
 
 /**
  * @author aaron
@@ -22,15 +17,22 @@ import java.util.List;
  *         Test out the functionality of the new genome loc parser
  */
 public class GenomeLocParserTest extends BaseTest {
-
     @Test(expected = StingException.class)
     public void testUnsetupException() {
+        GenomeLocParser.contigInfo = null;
         GenomeLocParser.createGenomeLoc(0, 0, 0);
     }
 
+    @BeforeClass
+    public static void init() {
+        SAMFileHeader header = ArtificialSAMUtils.createArtificialSamHeader(1, 1, 10);
+        GenomeLocParser.setupRefContigOrdering(header.getSequenceDictionary());
+    }
+
     @Test
     public void testKnownContigOrder() {
         SAMFileHeader header = ArtificialSAMUtils.createArtificialSamHeader(1, 1, 10);
+        GenomeLocParser.contigInfo = null;
         // assert that it's false when the contig ordering is not setup
         assertTrue(!GenomeLocParser.hasKnownContigOrdering());
         GenomeLocParser.setupRefContigOrdering(header.getSequenceDictionary());
@@ -97,6 +99,14 @@ public class GenomeLocParserTest extends BaseTest {
         assertEquals(1, loc.getStart());
     }
 
+    @Test
+    public void testCreateGenomeLoc1point5() { // in honor of VAAL!
+        GenomeLoc loc = GenomeLocParser.parseGenomeLoc("chr1:1");
+        assertEquals(loc.getContigIndex(), 0);
+        assertEquals(1, loc.getStop());
+        assertEquals(1, loc.getStart());
+    }
+
     @Test
     public void testCreateGenomeLoc2() {
         GenomeLoc loc = GenomeLocParser.createGenomeLoc(0, 1, 100);
@@ -130,7 +140,7 @@ public class GenomeLocParserTest extends BaseTest {
         assertEquals(1, copy.getStart());
     }
 
-    @Test
+    /*@Test // - uncomment if you want to test speed
     public void testGenomeLocParserList() {
         long start = System.currentTimeMillis();
         List<GenomeLoc> parsedIntervals = GenomeAnalysisEngine.parseIntervalRegion(Arrays.asList(new String[]{"/humgen/gsa-scr1/GATK_Data/Validation_Data/bigChr1IntervalList.list"}));
@@ -138,5 +148,21 @@ public class GenomeLocParserTest extends BaseTest {
         LinkedList<GenomeLoc> loc = new LinkedList<GenomeLoc>(GenomeLocParser.mergeOverlappingLocations(parsedIntervals));
         long stop = System.currentTimeMillis();
         logger.warn("Elapsed time = " + (stop - start));
+    }*/
+
+    @Test
+    public void testGenomeLocPlusSign() {
+        GenomeLoc loc = GenomeLocParser.parseGenomeLoc("chr1:1+");
+        assertEquals(0, loc.getContigIndex());
+        assertEquals(10, loc.getStop()); // the size
+        assertEquals(1, loc.getStart());
+    }
+
+    @Test(expected = RuntimeException.class)
+    public void testGenomeLocBad() {
+        GenomeLoc loc = GenomeLocParser.parseGenomeLoc("chr1:1-");
+        assertEquals(0, loc.getContigIndex());
+        assertEquals(10, loc.getStop()); // the size
+        assertEquals(1, loc.getStart());
     }
 }