I have had a similar problem to this before. What I did was I loaded up the Python shell, imported the module and then printed its __file__ attribute. From there I would just remove the folder or file that was being associated.
What you may want to look into is using virtualenv this system allows you to create a instance of python separate from your system. Any modules you install or use in this instance are self contained including the version of the module.
I keep all my projects now inside of there own contained virtualenv, which allows me to install and use whatever modules I want without worrying about screwing up modules from other projects.