[FEA] Add regular expression support to GPU implementation of StringSplit

[andygrove]

Is your feature request related to a problem? Please describe.
We currently implement StringSplit but only support non-regexp patterns.

Describe the solution you'd like
We should support regular expressions in the split pattern, compatible with Spark.

Describe alternatives you've considered
None

Additional context
None

[andygrove]

Depends on rapidsai/cudf#3584

[ttnghia]

Depends on rapidsai/cudf#10128 and rapidsai/cudf#10139.

[sameerz]

Example from #4658 (comment)

I wish we support split with regular expressions.

Mini repro:

val address = Seq((1,"abc.com"),
(2,"...abc"),
(3,".a.b.c"))

import spark.implicits._
val df = address.toDF("id","txt")
df.write.mode("overwrite").format("parquet").save("/tmp/testparquet")
val df2=spark.read.parquet("/tmp/testparquet")
df2.createOrReplaceTempView("df2")
spark.sql("select id,txt, split(txt, '\\.')[0] AS new_txt from df2").show()

Not-supported-messages:

!Expression split(txt#209, ., -1) cannot run on GPU because regular expressions are not supported yet