FindInPDFCollection.scala

import sys.process._

def read(f: File): String = {
  val fin = new java.io.FileInputStream(f)
  try {
    val sz = fin.available()
    val arr = new Array[Byte](sz)
    fin.read(arr)
    new String(arr, "UTF-8")
  } finally {
    fin.close()
  }
}

def run(dir: File, word: String = "swarm"): Unit = {
  val f1 = dir.children(_.extL == "pdf").sorted(File.NameOrdering)
  f1.foreach { f0 =>
    val tf = file("/data/temp/test.txt")
    tf.delete()
    val res = List("pdftotext", f0.path, tf.path).!
    if (res == 0) {
      val t = read(tf)
      val i = t.toLowerCase.indexOf(word)
      // val ok = if (i < 0) "  " else "!!"
      // println(s"$ok ${f0.name}")
      if (i >= 0) println(f0.name)
    }
  }
}

val d0 = file("/data/texts/Papers")
run(d0 / "B")

Variant:

val base = file("/data/texts/Papers")
val all = base.children(_.isDirectory).flatMap(_.children(_.extL == "pdf"))

def run(word: String): Unit =
  new Thread {
    override def run(): Unit = {
      var last = "?"
      all.sorted(File.NameOrdering)foreach { f =>
        import scala.sys.process._
        import scala.util.Try
        val h = f.name.substring(0, 1).toLowerCase
        if (h != last) {
          println(s"---- LETTER: $h ----")
          last = h
        }
        val contents: String = Try(Seq("pdftotext", "-q", f.path, "-").!!).getOrElse("")
        if (contents.toLowerCase.contains(word)) println(f.name)
      }
    }
      
    start()
  }

run("gradient")