TDAmapper Improvements

[peekxc]

Hi @paultpearson,

Love your package! I'm doing quite a bit of research into Mapper, and am in need of a fast Mapper implementation. I believe in reproducible research and prefer to stay open-source in the R world, so here I am!

I forked the package and rewrote the algorithmic portion using Reference Classes. A brief overview of the pull request is available here:

http://htmlpreview.github.io/?https://github.com/peekxc/TDAmapper/blob/master/mapper_ex.html

Notable changes:

• Internally use OOP-style R Reference classes in a similar way as used in the Python Mapper library
• Is much faster and more efficient.
• Handles all dimensional filter spaces uniformly, i.e. same as Python Mapper
• Addresses Issue 6 with actually a bit of more sophisticated solution: Bug in mapper2D$adjacency #6
• Addresses Issue 5 by allowing any clustering algorithm: Using different clustering method #5

This PR isn't comprehensive "CRAN"-ready, but I wanted to get your opinion on the whether the initial changes were something you wouldn't mind pulling in.