I believe you can use a visitor and manipulate the context for each of the rules you want to extra.
So when you visit the EvaluateStatement, you can access the context and call:
context.statement() to get a list of all the statement contexts that are children of the evaluateStatement.
If you are looking to get all the text associated with the paragraph 200-IDENTITY-TXNID, implement the part for the visitor of the paragraph rule and then you can access:
context.paragraphName() and context.sentence(). Then for each sentence you can call the sentence.statement() contexts.
Hope that helps a bit