There's a fundamental question here: is the intent to basically create a Python compiler that uses C as a back-end, or to convert the program to C and maintain the C afterward?
Writing a compiler that produces (really ugly) C as its output probably isn't trivial -- a compiler rarely is, and generating code for Python will be more difficult than for a lot of other languages (dynamic typing, in particular, is hard to compile, at least to very efficient output). OTOH, at least the parser will be a lot easier than for some languages.
If by "translating", you mean converting Python to C that's readable and maintainable, that's a whole different question -- it's substantially more difficult, to put it mildly. Realistically, I doubt any machine translation will be worth much -- there are just too large of differences in how you normally approach problems in Python and C for there to be much hope of a decent machine translation.
