Allow user to specify whether the field names in a Dataset query should be lowercased or not

[akudiyar]

Currently, when a schema is generated from the Dataset, the field names are implicitly lowercased: https://github.com/tarantool/cartridge-spark/blob/master/src/main/scala/org/apache/spark/sql/tarantool/TarantoolSchema.scala#L32. That was done so because usually the queries have the field names uppercased and therefore not compatible with the schema in Tarantool. But if the schema in Tarantool has uppercase letters in the field names, there is the same problem.

Acceptance criteria:

• Users are able to specify the field transformation
• The field names are not transformed by default