[FEA] Support `contains( strCol, expr )`

[mythrocks]

It would be good if spark-rapids would support the STRING contains() with a column or expression as the second argument.
See https://spark.apache.org/docs/latest/api/sql/index.html#contains.

Currently, when running queries using contains() such as the following:

SELECT COUNT(1) FROM fact WHERE contains( val, CAST( id AS STRING ) )

We see the operator fall off the GPU as follows:

           !Expression <Contains> Contains(cast(val#1L as string), cast(id#0L as string)) cannot run on GPU because search only supports StringType if it is a literal value

This then necessitates row-column transposition to get the data to/from the CPU operator.

Putting this operator on the GPU should enable some speedup.