I am not sure whether I can help you because I didn't use penelope and I don't know much about py.
One point is that the file name conventions for the Japanese dictionary (and therefore for the Chinese dictionary too) differ from the other dictionaries: if the word starts with (or consists in) a Chinese character the file name consists of only this one character. If however the word starts with something different (e.g full-width roman character) then the rules are the same as usual. I don't think penelope takes care of this. Therefore you might change the script accordingly or contact AlPe and ask whether he is interested in this. (Maybe he is rather following this thread
than this one).
Second point: I don't think changing the codepage improves the situation.
Which dictionary do you want to convert? Is it better than the cedict?