Allow KuduContext to create table with column encodings

Currently createTable method in KuduContext has no way to specify the column encodings. It uses org.apache.spark.sql.types.StructType, and there seems to be the way to specify optional Metadata in StructField.

new
StructField(name: String, dataType: DataType, nullable: Boolean = true, metadata: Metadata = Metadata.empty)

Can we update the createTable method to parse out optional metadata and add the encodings if specify.