Make lowercasing by Normalizer optional

[lambdaofgod]

Description

I've noted that Normalizer lowercases all tokens that are fed into it. I added lowercase parameter that, when set to false, doesn't lowercase tokens. By default it's true so it behaves like previous version.

Motivation and Context

It seems it could be advantageous to not lowercase some inputs (for example if someone wanted to feed output of SparkNLP pipeline to some other algorithms that are case-sensitive).

How Has This Been Tested?

I've checked old tests for Normalizer and they still work, and I added tests for lowercase = true.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• Code improvements with no or little impact
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)

Checklist:

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have read the CONTRIBUTING page.
• I have added tests to cover my changes.
• All new and existing tests passed.

[lambdaofgod]

Should I make this PR to master or tokenizer-normalizer-params? I thought about the latter, but since it wasn't updated for a bit it looked like I'd change over 100 files

[saif-ellafi]

Looks good. I will test it shortly. Thank you for your contribution. Looking for more!

[saif-ellafi]

I am thinking of how else the Normalizer could be improved... maybe standardizing dates? (perhaps using date matcher), clearing up some abbreviations (Sentence detector has some regex work with abbreviations) or something else..