Perlin noise

[falkecarlsen]

• Find good lib for generating Perlin noise
• Use sumolib for more easily manipulating SUMO network
• Iterate over every edge-pair in network
• For each edge-pair apply perlin-noise dictated population proportional to stat-population
• Apply offset to Perlin noise generator to avoid (0,0) = 0 result
• Write unit-tests

Unit-tests and further work is delegated to later PRs as this is needed on master.

[NicEastvillage]

The additions are good but the PR seems unfinished. By that I don't mean unit testing, but a thourough clean up to make things more readable. Too many functions are just simple expressions that could be inlined.

Additionally, your provided perlin simplex noise function does consider scale, octaves, lucanarity, or persistance. We should find some good values for those either at calling site or as parameters to the whole tool.

[falkecarlsen]

Turns out that 'noise' does not support seeding but a stale PR exists which fixes this: caseman/noise#22.

Realistically, in this stage of development, simply offsetting and/or scaling would suffice as the resolution in 'noise' is very large.

[NicEastvillage]

Nice. Now that is better. I agree with your comments regarding #4, but it should be one of the next things we do.