Add benchmarking

[MeBrei]

The benchmarking can be run with cargo bench.

So far, 4 different aspects are benchmarked.

• bench_osrank_naive_on_small_network tests performance with different randomly generated seeds for the rng (on a small sample graph). Of interest is not only that the average time of execution but also that the min and max do not vary too much
• bench_osrank_naive_on_sample_csv runs the full osrank_naive algorithm on a sample of the Rust network (1000 projects & 10000 contributions)
• bench_random_walk_on_csv & bench_rank_network run the respective parts of osrank_naive on the same sample set

[adinapoli-mndc]

Thanks @MeBrei , shaping up nicely! May I ask you the relation behind picking bencher to run the benchmarks? Maybe @xla or @cloudhead can help us here with some ecosystem discovery, but as far as I know there are 2 projects in the Rust ecosystem which might deserve some of our attention:

• criterion-rs - This seems to be a port of the popular Haskell's criterion library, and it has 800 starts on Github.

• hyperfine - Seems to have a staggering 3k starts on Github 😮

Have you explored any of these?

[cloudhead]

From what I know, criterion is the way to go.