Some suggestions for code improvement

[philshem]

I'm a big fan the code you've done to make this great dataset usable. I've had to request a major international company install "sexmachine", so that's how useful it has been *. In using the code, I've come up with some suggestions for how to improve. I'm happy to implement some if you think they are useful.

• Country naming support: Although defined in detector.py,

  COUNTRIES = u"""great_britain ireland usa italy ... other_countries""".split()

Names can be additionally mapped by ISO 3166-1-alpha-2 codes. For example,

d.get_gender(u"Andrea", 'IT')

would return the same as

d.get_gender(u"Andrea", 'italy')

• Distinguish "not found" and "androgynous": Names that are not found should be distinguished from names that are androgynous. For example,

  print d.get_gender(u"Pauley")
  print d.get_gender(u"dssad12jkasdl")

Should print "andy", "unknown" instead of twice printing "andy".

• Include some data: Return country probability for a name. For example:

  print d.get_gender(u"Andrea")

would return something like:

italy, male, 99%
usa, mostly_female, 85%
global, female, 95%

* The Ruby package was renamed gender_detector based on the blog post here.

[alisonjo315]

I think those are really thoughtful and good suggestions, so I'm just expressing my interest and support (oh and maybe a little bit bumping this thread). I'd be happy to help, too, if that makes a difference.

[dukebody]

The mantainer doesn't look very responsive. Perhaps we can continue this on https://github.com/eddies/sexmachine ? @eddies

[dukebody]

@philshem @alisonjo2786 I've implemented the 'Distinguish "not found" and "androgynous"' feature in the dev version of https://github.com/lead-ratings/gender-guesser. Please take a look and see if we can continue the development there.

[alisonjo315]

Thank you @dukebody ! I can't review this week, but I will make a note to try to review it this weekend or next week -- sorry, and thanks!