[#149] Create initial version of 'partition' function

[rockbmb]

Closes #149. No tests, no benchmarks, no comments, and incorrect results. However, it works, and I'll be improving the code with these things. Ideally I reproduce the author's p(10^19) result.

• [x] Make it work with Double (or more generally, Floating).
• [ ] Improve the function so that p(n) does return a result with O(n^2) bits of precision. Some type-level programming will be required to return a fixed-point type whose precision is dependent on the input.
• [ ] Write tests that verify (should I put "verify" between quotes :smiley:) the function's correctness
• [ ] Write benchmarks
• [ ] Write better comments

[rockbmb]

Very stale and honestly, not really sure this is a good idea anymore.

Should this ever become a necessary feature, one can just come back to this PR and reuse whatever is possible.