Add support for generic record builders

[yruslan]

Background

Currently, in order to parse mainframe records, you need an instance of Spark for both batch processing and streaming.

The COBOL parser can be utilized in more generic EBCDIC records processing workflow, such as parsing records coming from MQ.

Feature

In order to allow for such a workflow, Cobrix should provide

• Generic data structures to hold records (similar to Row in Spark)
• Ways to traverse these data structures
• Example builders for converting records to JSON, XML.
• Encoders to be able to convert from these generic data structures to an EBCDIC record