/* Extract PDF bytes and PDF metadata. */
    def extractPDFDetailsDF(): DataFrame = {
      val records = rdd
        .map(r =>
            (r, (DetectMimeTypeTika(r.getBinaryBytes)))
            )
        .filter(r => r._2 == "application/pdf")
        .map(r => {
          val bytes = r._1.getBinaryBytes
          val md5Hash = new String(Hex.encodeHex(MessageDigest.getInstance("MD5").digest(bytes)))
          val sha1Hash = new String(Hex.encodeHex(MessageDigest.getInstance("MD5").digest(bytes)))
          val encodedBytes = Base64.getEncoder.encodeToString(bytes)
          val url = new URL(r._1.getUrl)
          val filename = FilenameUtils.getName(url.getPath())
          val extension = GetExtensionMime(url.getPath(), r._2)
          (r._1.getUrl, filename, extension, r._1.getMimeType,
            DetectMimeTypeTika(r._1.getBinaryBytes), md5Hash, sha1Hash, encodedBytes)
        })
        .map(t => Row(t._1, t._2, t._3, t._4, t._5, t._6, t._7, t._8))

      val schema = new StructType()
        .add(StructField("url", StringType, true))
        .add(StructField("filename", StringType, true))
        .add(StructField("extension", StringType, true))
        .add(StructField("mime_type_web_server", StringType, true))
        .add(StructField("mime_type_tika", StringType, true))
        .add(StructField("md5", StringType, true))
        .add(StructField("sha1", StringType, true))
        .add(StructField("bytes", StringType, true))

      val sqlContext = SparkSession.builder();
      sqlContext.getOrCreate().createDataFrame(records, schema)
    }