[ADAM-1939] Allow validation stringency to waive off FLAG arrays.

Resolves #1939.
bigdatagenomics · Mar 7, 2018 · bc8c6f4 · bc8c6f4
1 parent 7d07c71
commit bc8c6f4
Show file tree

Hide file tree

Showing 3 changed files with 29 additions and 3 deletions.
diff --git a/adam-core/src/main/scala/org/bdgenomics/adam/converters/VariantContextConverter.scala b/adam-core/src/main/scala/org/bdgenomics/adam/converters/VariantContextConverter.scala
@@ -296,7 +296,7 @@ class VariantContextConverter(
   import VariantContextConverter._
 
   // format fns gatk --> bdg, extract fns bdg --> gatk
-  private val variantFormatFn = makeVariantFormatFn(headerLines)
+  private val variantFormatFn = makeVariantFormatFn(headerLines, stringency)
   private val variantExtractFn = makeVariantExtractFn(headerLines)
   private val genotypeFormatFn = makeGenotypeFormatFn(headerLines)
   private val genotypeExtractFn = makeGenotypeExtractFn(headerLines)
@@ -1636,7 +1636,8 @@ class VariantContextConverter(
   }
 
   private def makeVariantFormatFn(
-    headerLines: Seq[VCFHeaderLine]): (HtsjdkVariantContext, Option[String], Int, Boolean) => (Variant, Variant) = {
+    headerLines: Seq[VCFHeaderLine],
+    stringency: ValidationStringency = ValidationStringency.STRICT): (HtsjdkVariantContext, Option[String], Int, Boolean) => (Variant, Variant) = {
 
     val attributeFns: Iterable[(HtsjdkVariantContext, Int, Array[Int]) => Option[(String, String)]] = headerLines
       .flatMap(hl => hl match {
@@ -1648,7 +1649,20 @@ class VariantContextConverter(
           if (DefaultHeaderLines.infoHeaderLines
             .find(_.getID == key)
             .isEmpty) {
-            lineToVariantContextExtractor(il)
+            try {
+              lineToVariantContextExtractor(il)
+            } catch {
+              case t: Throwable => {
+                if (stringency == ValidationStringency.STRICT) {
+                  throw t
+                } else {
+                  if (stringency == ValidationStringency.LENIENT) {
+                    log.warn("Saw invalid info field %s. Ignoring...".format(t))
+                  }
+                  None
+                }
+              }
+            }
           } else {
             None
           }

diff --git a/adam-core/src/test/resources/invalid/small.INFO_flag.vcf b/adam-core/src/test/resources/invalid/small.INFO_flag.vcf
@@ -0,0 +1,5 @@
+##fileformat=VCFv4.1
+##INFO=<ID=ABADFLAG,Number=.,Type=Flag,Description="A no good, very bad flag.">
+##contig=<ID=1,length=249250621>
+#CHROM	POS	ID	REF	ALT	QUAL	FILTER	INFO	FORMAT	NA12878	NA12891	NA12892
+1	14397	.	CTGT	C	139.12	IndelQD	AC=2;AF=0.333;AN=6;BaseQRankSum=1.800;ClippingRankSum=0.138;DP=69;FS=7.786;MLEAC=2;MLEAF=0.333;MQ=26.84;MQ0=0;MQRankSum=-1.906;QD=1.55;ReadPosRankSum=0.384	GT:AD:DP:FT:GQ:PL	0/1:16,4:20:rd:99:120,0,827	0/1:8,2:10:dp;rd:60:60,0,414	0/0:39,0:39:PASS:99:0,116,2114
diff --git a/adam-core/src/test/scala/org/bdgenomics/adam/rdd/ADAMContextSuite.scala b/adam-core/src/test/scala/org/bdgenomics/adam/rdd/ADAMContextSuite.scala
@@ -179,6 +179,13 @@ class ADAMContextSuite extends ADAMFunSuite {
     assert(last.getName === "gn|BPY2C;ccds|CCDS44030;ens|ENSG00000185894;vega|OTTHUMG00000045199")
   }
 
+  sparkTest("can read a small .vcf file with a validation issue") {
+    val path = testFile("invalid/small.INFO_flag.vcf")
+
+    val vcs = sc.loadVcf(path, stringency = ValidationStringency.LENIENT)
+    assert(vcs.rdd.count === 1)
+  }
+
   sparkTest("can read a small .vcf file") {
     val path = testFile("small.vcf")