Create a benchmark utility

[jmgq]

Inspired by the script provided by @pathway in #5 (examples/Terrain/scaletest01.php), I think it would be a good idea to have a benchmark utility that worked in more general cases. The benchmark utility should be able to:

1. Use any terrain defined in a text file, as opposed to a hard-coded terrain.
2. Customize the for loop (terrain size increase per loop iteration).
3. Adjust the terrain on each iteration of the for loop. The script in Scale #5 always use the full terrain array, but I think that the main terrain should be cropped to the desired size on each iteration.
4. Create a terrain generator. It should accept a size and a destination file path as parameters, as well as the minimum and maximum tile cost.
5. As an alternative (or complement) to item 1, use a random terrain. The downside is that we wouldn't be able to use the same terrain in two different benchmarks, so in this case, the for loop should be executed several times and then display averages. Or it could accept the RNG's seed as an optional parameter, that way the same terrain can be generated in different benchmarks, and there is no need to run the benchmark several times or calculate averages.
6. Use the symfony/Console component, as it is a very convenient library to create console commands, including utilities like progress bars, tables, and so on.
7. After the benchmark is run, display the used configuration and a table with the results.

[pathway]

Exciting!

We can then report for each run things like:

• memory usage (Im a bit concerned about that right now)
• timing

• of nodes expanded

...and any other metrics.

This will give us excellent visibility into the effects of any proposed changes to the algorithm or data structures.