Alignment with more typical Python standards.

[pablo-meier]

The song for this PR is Fantastic Baby, by BIGBANG.

This does a few things:

• Handle dependencies using a requirements.txt file, which is pretty standard for Python.
• Add formatting to the README so math looks a bit more like math.
• Puts the sources in their own directory.
• Converts the source files to be PEP8 compliant.
• Most dangerously, bumps the project from Python 2 to Python 3.

Most of the Py2 -> Py3 conversion was done with 2to3 suggestions, but without votes, I can't really test to ensure that the output is the same. I'll try to obtain some.

I also have questions about the actual algorithms we're implementing here — a lot of the README says "PageRank but not lol," and navigating some of the jargon isn't clear to me. I'm no stranger to graph theory, but some terms are still leaving me scratching my head.

Next steps would be to throw some tests on this.

[pablo-meier]

Additional changes:

• Use csv DictReader instead of our home-grown field extractor.
• Use defaultdict to generate our graph over checking key existence in a way prone to KeyErrors then manually assigning lists.