This is a simple library to parse Chinese names into surname + given name. It can also guess the gender from the given name. Chinese names are less strictly separated by gender than in many other cultures, but in this way it is possible to get an educated guess. The library was orginally created for a project that involved identifying the genders of authors when the only information available was their names.
The gender estimates are based on character-level statistics from names of people born 1930-2008 in mainland China.
pip install --user .
>>> from chinese_names import ChineseNames
>>> cn = ChineseNames()
>>> cn.parse("诸葛亮")
ChineseName(surname='诸葛', given_name='亮', p_male=0.9480300169983071, p_female=0.051969983001692865, p_name=1.3000188515699997e-07)
>>> cn.parse("Freddy Foreigner")
>>>
Note that p_name is the total name probability from a unigram character
model. Thus it is generally a rather low value. You should be able to get a
very rough approximation of the expected number of bearers, given naive
independence assumptions, by multiplying with population size. For the PRC:
>>> 1.412e9 * cn.parse("王伟").p_name
1399604.0484204555
>>> 1.412e9 * cn.parse("诸葛亮").p_name
183.56266184168396
Data is obtained from the ChineseNames R library under a CC-BY-NC-SA license.