Added python3 support and some simple tests and linting (+ TravisCI)

[chaosmail]

Hi guys,

thanks for this excellent library and the nice tutorial videos; these things make it really exciting to learn about Recurrent Neural Nets. I wanted to try Passage in Python 3 for the Crowdflower kaggle challenge (just to try it out, the dataset is too small I think), so I needed to add py3 support to it. Here is what I did:

• Fixed relative imports to support py3.
• Added simple test for TravisCI.
• Added register.py file to automatically generate README.rst and upload to pypi.
• Added pep8 linter and tox.ini config.

Not sure if this should go to the master or dev branch, so I am just dropping this here for the moment. If you have any question about the above changes, I would be happy to discuss in the comments.

Best, Christoph

[moneyatindico]

These are no longer applicable, closing