Unary estimator for detecting names and transforming to gender

[MWYang]

The PR implements the following algorithm for detecting names in text fields and inferring gender as a standalone unary estimator:

• First, a majority (or some custom threshold) of entries must be found in a large name dictionary
• If the threshold is reached, the transformer uses several heuristics (the first or last token; RegEx patterns; presence of honorifics like Mr. or Ms.) to find the likely first name and infer the gender.

Additional context
I completed this work as part of my ongoing Salesforce internship. This PR cleans up and replaces many previous ones: #440, #437, #428. Per internal review with @leahmcguire, I'm including my name identification and name-to-gender code in two PRs. This first one implements the standalone unary estimator where most of the actual logic is in a utils file, NameDetectUtils.scala. The code also makes use of Algebird monoids in order to run in a single pass over the data. Pending approval on this PR, I will then make a separate PR with the proper (and hopefully fewer) changes made in SmartTextVectorizer.

[tovbinm]

let's please follow the convention of camelCasing the parameters, e.g. replace guard_maxNumberOfTokens with guardMaxNumberOfTokens @MWYang

[tovbinm]

BTW, where are the GenderDictionary_USandUK.csv and Names_JRC_Combined.txt coming from? What's the license on these datasets? @MWYang

[MWYang]

BTW, where are the GenderDictionary_USandUK.csv and Names_JRC_Combined.txt coming from? What's the license on these datasets? @MWYang

They are both from governmental organizations. For the first, US and UK birth registry data, and for the second, the EU science commission. I also documented the pre-processing of those files in a separate repo: https://github.com/MWYang/InternationalNames

[tovbinm]

looks good! some minor comments.

[tovbinm]

@MWYang kudos on your 1st contribution!!! 🔥

[salesforce-cla]

Thanks for the contribution! It looks like @MWYang is an internal user so signing the CLA is not required. However, we need to confirm this.