Provide equals and hashCode for ParserModel and TokenizerModel

The tests opennlp.tools.tokenize.TokenizerModelTest, opennlp.tools.parser.chunking.ParserTest, and opennlp.tools.parser.treeinsert.ParserTest signal by TODOs that no actual assertions are made to check whether (de-)serialized Model instances are correctly read in again.
In other words: Those (Tokenizer/Parser) models lacks a valid equals/hashCode implementation by which one could verify a valid state (=equality) after (de-)serialization has occurred.

Aim:

• Provide an improved implementation of the related "Model" classes.
• Remove existing TODOs in TokenizerModelTest and ParserTest
• Improve the three test classes with further test assertions.