fails on double 'n'

[garfieldnate]

In both traditional and modified Hepburn romanization, "annai" should be rendered as あんない, but romkan renders it as あんあい.

import romkan
romkan.to_hiragana("annai")
'あんあい'