From 9246472f702c72ef5134faa079c219718af4bbcb Mon Sep 17 00:00:00 2001
From: Ryan Williams <ryan.blake.williams@gmail.com>
Date: Wed, 23 Mar 2016 16:28:56 +0000
Subject: [PATCH] add validation stringency to bam parsing, flagstat

---
 .../org/bdgenomics/adam/cli/FlagStat.scala    | 19 ++++++++++++++-----
 .../org/bdgenomics/adam/rdd/ADAMContext.scala |  6 ++++--
 2 files changed, 18 insertions(+), 7 deletions(-)

diff --git a/adam-cli/src/main/scala/org/bdgenomics/adam/cli/FlagStat.scala b/adam-cli/src/main/scala/org/bdgenomics/adam/cli/FlagStat.scala
index 5c4f485631..7ca51bfb6c 100644
--- a/adam-cli/src/main/scala/org/bdgenomics/adam/cli/FlagStat.scala
+++ b/adam-cli/src/main/scala/org/bdgenomics/adam/cli/FlagStat.scala
@@ -17,15 +17,15 @@
  */
 package org.bdgenomics.adam.cli
 
-import org.apache.hadoop.fs.{ Path, FileSystem }
-import org.apache.hadoop.mapreduce.Job
+import htsjdk.samtools.ValidationStringency
+import org.apache.hadoop.fs.{ FileSystem, Path }
 import org.apache.spark.SparkContext
 import org.apache.spark.rdd.RDD
-import org.bdgenomics.adam.projections.{ Projection, AlignmentRecordField }
+import org.bdgenomics.adam.projections.{ AlignmentRecordField, Projection }
 import org.bdgenomics.adam.rdd.ADAMContext._
 import org.bdgenomics.formats.avro.AlignmentRecord
 import org.bdgenomics.utils.cli._
-import org.kohsuke.args4j.Argument
+import org.kohsuke.args4j.{ Argument, Option ⇒ Args4jOption }
 
 object FlagStat extends BDGCommandCompanion {
   val commandName: String = "flagstat"
@@ -41,6 +41,8 @@ class FlagStatArgs extends Args4jBase with ParquetArgs {
   val inputPath: String = null
   @Argument(required = false, metaVar = "OUTPUT", usage = "Optionally write the stats to this file.", index = 1)
   val outputPath: String = null
+  @Args4jOption(required = false, name = "-stringency", usage = "Set the parsing stringency: SILENT, LENIENT, STRICT.")
+  val stringency: String = "SILENT"
 }
 
 class FlagStat(protected val args: FlagStatArgs) extends BDGSparkCommand[FlagStatArgs] {
@@ -65,7 +67,14 @@ class FlagStat(protected val args: FlagStatArgs) extends BDGSparkCommand[FlagSta
       AlignmentRecordField.supplementaryAlignment
     )
 
-    val adamFile: RDD[AlignmentRecord] = sc.loadAlignments(args.inputPath, projection = Some(projection))
+    val stringency = ValidationStringency.valueOf(args.stringency)
+
+    val adamFile: RDD[AlignmentRecord] =
+      sc.loadAlignments(
+        args.inputPath,
+        projection = Some(projection),
+        stringency = stringency
+      )
 
     val (failedVendorQuality, passedVendorQuality) = adamFile.adamFlagStat()
 
diff --git a/adam-core/src/main/scala/org/bdgenomics/adam/rdd/ADAMContext.scala b/adam-core/src/main/scala/org/bdgenomics/adam/rdd/ADAMContext.scala
index 31e97ceb39..70c4a577c2 100644
--- a/adam-core/src/main/scala/org/bdgenomics/adam/rdd/ADAMContext.scala
+++ b/adam-core/src/main/scala/org/bdgenomics/adam/rdd/ADAMContext.scala
@@ -261,7 +261,8 @@ class ADAMContext(@transient val sc: SparkContext) extends Serializable with Log
    *
    * @see loadAlignments
    */
-  def loadBam(filePath: String): AlignmentRecordRDD = {
+  def loadBam(filePath: String,
+              validationStringency: ValidationStringency = ValidationStringency.STRICT): AlignmentRecordRDD = {
     val path = new Path(filePath)
     val fs =
       Option(
@@ -289,6 +290,7 @@ class ADAMContext(@transient val sc: SparkContext) extends Serializable with Log
             // We need to separately read the header, so that we can inject the sequence dictionary
             // data into each individual Read (see the argument to samRecordConverter.convert,
             // below).
+            sc.hadoopConfiguration.set(SAMHeaderReader.VALIDATION_STRINGENCY_PROPERTY, validationStringency.toString)
             val samHeader = SAMHeaderReader.readSAMHeaderFrom(fp, sc.hadoopConfiguration)
             log.info("Loaded header from " + fp)
             val sd = adamBamDictionaryLoad(samHeader)
@@ -858,7 +860,7 @@ class ADAMContext(@transient val sc: SparkContext) extends Serializable with Log
     if (filePath.endsWith(".sam") ||
       filePath.endsWith(".bam")) {
       log.info("Loading " + filePath + " as SAM/BAM and converting to AlignmentRecords. Projection is ignored.")
-      loadBam(filePath)
+      loadBam(filePath, stringency)
     } else if (filePath.endsWith(".ifq")) {
       log.info("Loading " + filePath + " as interleaved FASTQ and converting to AlignmentRecords. Projection is ignored.")
       loadInterleavedFastq(filePath)