Split plotting.py

[eigenfoo]

The functionality of pyfolio can be seen as computing measures of performance/risk, and plotting such measures for easy visualization by humans. Right now, the computing is done in respective modules (e.g. pos.py, risk.py) and all plotting is handled in plotting.py.

As discussed in #391, the maintainers have decided to split up plotting.py, moving the plotting functions into their corresponding modules:

• pyfolio is modularized into separate tear sheets because that's how the user will consume the final product: tear sheet by tear sheet. It doesn't make sense to me to have the functionality of, say, the returns tear sheet split between two different modules.
• pyfolio plots much more than it computes: plotting.py clocks in at around 2000 lines of code, whereas most other modules barely make 500. Refactoring so that plotting functions immediately follow their corresponding compute functions would make pyfolio much easier to read.