> Oops, please try the attached new one.

Is it possible to get *all* possible input sequences?  For example, ˵
(shuo1) can also be input as `yue4' in chinese-tonepy, and this is
what your code currently returns.


    Werner