[ADAM-1165] Support merging shards across multiple file systems.

Resolves bigdatagenomics#1165.

adam-cli/src/main/scala/org/bdgenomics/adam/cli/MergeShards.scala

@@ -67,10 +67,11 @@ class MergeShards(val args: MergeShardsArgs) extends BDGSparkCommand[MergeShards
     val optHeadPath = Option(args.headerPath).map(p => new Path(p))
     val tailPath = new Path(args.inputPath)
     val outputPath = new Path(args.outputPath)
-    val fs = tailPath.getFileSystem(conf)
+    val fsIn = tailPath.getFileSystem(conf)
+    val fsOut = outputPath.getFileSystem(conf)
 
     // merge the files
-    FileMerger.mergeFiles(fs,
+    FileMerger.mergeFilesAcrossFilesystems(fsIn, fsOut,
       outputPath, tailPath, optHeadPath,
       writeEmptyGzipBlock = args.gzipAtEof,
       bufferSize = args.bufferSize)

adam-core/src/main/scala/org/bdgenomics/adam/rdd/FileMerger.scala

@@ -38,16 +38,45 @@ private[adam] object FileMerger extends Logging {
    *   been written.
    * @param writeEmptyGzipBlock If true, we write an empty GZIP block at the
    *   end of the merged file.
+   *
+   * @see mergeFilesAcrossFilesystems
    */
   def mergeFiles(fs: FileSystem,
                  outputPath: Path,
                  tailPath: Path,
                  optHeaderPath: Option[Path] = None,
                  writeEmptyGzipBlock: Boolean = false,
                  bufferSize: Int = 4 * 1024 * 1024) {
+    mergeFilesAcrossFilesystems(fs, fs,
+      outputPath, tailPath, optHeaderPath = optHeaderPath,
+      writeEmptyGzipBlock = writeEmptyGzipBlock,
+      bufferSize = bufferSize)
+  }
+
+  /**
+   * Merges together sharded files, while preserving partition ordering.
+   *
+   * Can read files from a different filesystem then they are written to.
+   *
+   * @param fsIn The file system implementation to use for the tail/head paths.
+   * @param fsOut The file system implementation to use for the output path.
+   * @param outputPath The location to write the merged file at.
+   * @param tailPath The location where the sharded files have been written.
+   * @param optHeaderPath Optionally, the location where a header file has
+   *   been written.
+   * @param writeEmptyGzipBlock If true, we write an empty GZIP block at the
+   *   end of the merged file.
+   */
+  def mergeFilesAcrossFilesystems(fsIn: FileSystem,
+                                  fsOut: FileSystem,
+                                  outputPath: Path,
+                                  tailPath: Path,
+                                  optHeaderPath: Option[Path] = None,
+                                  writeEmptyGzipBlock: Boolean = false,
+                                  bufferSize: Int = 4 * 1024 * 1024) {
 
     // get a list of all of the files in the tail file
-    val tailFiles = fs.globStatus(new Path("%s/part-*".format(tailPath)))
+    val tailFiles = fsIn.globStatus(new Path("%s/part-*".format(tailPath)))
       .toSeq
       .map(_.getPath)
       .sortBy(_.getName)
@@ -66,7 +95,7 @@ private[adam] object FileMerger extends Logging {
     // but! it is correct.
 
     // open our output file
-    val os = fs.create(outputPath)
+    val os = fsOut.create(outputPath)
 
     // here is a byte array for copying
     val ba = new Array[Byte](bufferSize)
@@ -91,7 +120,7 @@ private[adam] object FileMerger extends Logging {
       log.info("Copying header file (%s)".format(p))
 
       // open our input file
-      val is = fs.open(p)
+      val is = fsIn.open(p)
 
       // until we are out of bytes, copy
       copy(is, os)
@@ -112,7 +141,7 @@ private[adam] object FileMerger extends Logging {
         numFiles))
 
       // open our input file
-      val is = fs.open(p)
+      val is = fsIn.open(p)
 
       // until we are out of bytes, copy
       copy(is, os)
@@ -134,7 +163,7 @@ private[adam] object FileMerger extends Logging {
     os.close()
 
     // delete temp files
-    optHeaderPath.foreach(headPath => fs.delete(headPath, true))
-    fs.delete(tailPath, true)
+    optHeaderPath.foreach(headPath => fsIn.delete(headPath, true))
+    fsIn.delete(tailPath, true)
   }
 }