consolidating the checking of the RMD sequence dictionary against the reference into a single function, and adding an integration test to test that empty VCFs pass (both the indexing and the seq dictionary validation).

git-svn-id: file:///humgen/gsa-scr1/gsa-engineering/svn_contents/trunk@4750 348d0f76-0448-11de-a6fe-93d51630548a

java/src/org/broadinstitute/sting/gatk/AbstractGenomeAnalysisEngine.java

@@ -334,7 +334,7 @@ public abstract class AbstractGenomeAnalysisEngine {
         validateSuppliedReferenceOrderedData(tracks);
 
         // validate all the sequence dictionaries against the reference
-        validateSourcesAgainstReference(readsDataSource, referenceDataSource.getReference(), tracks);
+        validateSourcesAgainstReference(readsDataSource, referenceDataSource.getReference(), tracks, manager);
 
         rodDataSources = getReferenceOrderedDataSources(tracks);
     }
@@ -505,7 +505,7 @@ public abstract class AbstractGenomeAnalysisEngine {
      * @param reference Reference data source.
      * @param tracks    a collection of the reference ordered data tracks
      */
-    private void validateSourcesAgainstReference(SAMDataSource reads, ReferenceSequenceFile reference, Collection<RMDTrack> tracks) {
+    private void validateSourcesAgainstReference(SAMDataSource reads, ReferenceSequenceFile reference, Collection<RMDTrack> tracks, RMDTrackBuilder manager) {
         if ((reads.isEmpty() && (tracks == null || tracks.isEmpty())) || reference == null )
             return;
 
@@ -531,21 +531,8 @@ public abstract class AbstractGenomeAnalysisEngine {
         }
 
         // compare the tracks to the reference, if they have a sequence dictionary
-        for (RMDTrack track : tracks) {
+        for (RMDTrack track : tracks)
-            SAMSequenceDictionary trackDict = track.getSequenceDictionary();
+            manager.validateTrackSequenceDictionary(track.getName(),track.getSequenceDictionary(),referenceDictionary);
-
-            // hack: if the sequence dictionary is empty (as well as null which means it doesn't have a dictionary), skip validation
-            if (trackDict == null || trackDict.size() == 0) {
-                logger.info("Track " + track.getName() + " doesn't have a sequence dictionary built in, skipping dictionary validation");
-                continue;
-            }
-
-            Set<String> trackSequences = new TreeSet<String>();
-            for (SAMSequenceRecord dictionaryEntry : trackDict.getSequences())
-                trackSequences.add(dictionaryEntry.getSequenceName());
-            SequenceDictionaryUtils.validateDictionaries(logger, getArguments().unsafe, track.getName(), trackDict, "reference", referenceDictionary);
-        }
-
     }
 
     /**

java/src/org/broadinstitute/sting/gatk/refdata/tracks/builders/RMDTrackBuilder.java

@@ -503,7 +503,7 @@ public class RMDTrackBuilder extends PluginManager<FeatureCodec> {
         if (dict == null) return;
 
         SAMSequenceDictionary currentDict = createSequenceDictionaryFromContigList(index, new SAMSequenceDictionary());
-        SequenceDictionaryUtils.validateDictionaries(logger,validationExclusionType,"GATK",dict,inputFile.getAbsolutePath(),currentDict);
+        validateTrackSequenceDictionary(inputFile.getAbsolutePath(),currentDict,dict);
 
         // check that every contig in the RMD contig list is at least in the sequence dictionary we're being asked to set
         for (SAMSequenceRecord seq : currentDict.getSequences()) {
@@ -518,4 +518,17 @@ public class RMDTrackBuilder extends PluginManager<FeatureCodec> {
             logger.warn("Unable to update index with the sequence dictionary for file " + indexFile + "; this will not effect your run of the GATK");
         }
     }
+
+
+    public void validateTrackSequenceDictionary(String trackName, SAMSequenceDictionary trackDict, SAMSequenceDictionary referenceDict) {
+        // if the sequence dictionary is empty (as well as null which means it doesn't have a dictionary), skip validation
+        if (trackDict == null || trackDict.size() == 0)
+            logger.info("Track " + trackName + " doesn't have a sequence dictionary built in, skipping dictionary validation");
+        else {
+            Set<String> trackSequences = new TreeSet<String>();
+            for (SAMSequenceRecord dictionaryEntry : trackDict.getSequences())
+                trackSequences.add(dictionaryEntry.getSequenceName());
+            SequenceDictionaryUtils.validateDictionaries(logger, validationExclusionType, trackName, trackDict, "reference", referenceDict);
+        }
+    }
 }

java/test/org/broadinstitute/sting/playground/gatk/walkers/validation/RodSystemValidationIntegrationTest.java

@@ -3,10 +3,13 @@ package org.broadinstitute.sting.playground.gatk.walkers.validation;
 import org.broadinstitute.sting.WalkerTest;
 import org.testng.annotations.Test;
 
+import java.io.File;
+import java.util.ArrayList;
 import java.util.Arrays;
+import java.util.List;
 
 /**
- * The pile-up tests, that test any changes to the underlying ROD system
+ * tests for the ROD system in general; from rod system validation to empty VCF files
  */
 public class RodSystemValidationIntegrationTest extends WalkerTest {
 
@@ -31,6 +34,18 @@ public class RodSystemValidationIntegrationTest extends WalkerTest {
         executeTest("testSimpleVCFPileup", spec);
     }
 
+    @Test
+    public void testEmptyVCF() {
+        File vcf = new File(validationDataLocation + "justHeader.vcf.idx");
+        if (vcf.exists()) vcf.delete();
+
+        WalkerTestSpec spec = new WalkerTestSpec(
+                baseTestString1KG() + " -B:eval,VCF " + validationDataLocation + "justHeader.vcf", 1,
+                Arrays.asList("b3e80d77f8c05edf2a385b0a87708670"));
+        executeTest("testEmptyVCF", spec);
+    }
+
+
     @Test
     public void testComplexVCFPileup() {
         WalkerTestSpec spec = new WalkerTestSpec(