Unicode friendly alphabetic pattern for python regex?

python, regex

Solution

[^\W\d]

Throw out non-word characters and throw out digits. Keep the rest.

Problem

I'm looking for a pattern equivalent to \w, and which doesn't match numeric pattern. I cannot use [a-zA-Z] because I would like it to match japanese kanjis as well. Is there a way to write something like [\w^[0-9]] ? Is there an equivalent of [:alpha:] in python regex?

Original source