Example: Speak -> Spubeak, more info here
Don\'t give me a solution, but point me in the right direction or tell which which python library I could
You can use regular expressions for substitutions. See re.sub.
Example:
>>> import re
>>> re.sub(r'(e)', r'ub\1', 'speak')
'spubeak'
You will need to read the documentation for regex groups and so on. You will also need to figure out how to match different vowels instead of just the one in the example.
For some great ideas (and code) for using regular expressions in Python for a pronunciation dictionary, take a look at this link, which is one of the design pages for the Cainteoir project: http://rhdunn.github.com/cainteoir/rules.html
Cainteoir's text-to-speech rule engine design (which is not fully implemented yet) uses regular expressions. See also Pronunciation Dictionaries and Regexes, another article by the Cainteoir author.