Support for irregular grids

[jaylorch]

This adds support for irregular grids, as found in the examples shape and statue_park_loop.

It changes the interface to use Point and Vector instead of generic tuples. This should also help with a planned addition of support for hexagonal grids, and perhaps even later support for three-dimensional grids. All examples have been updated to use the new interface, including in the tutorial. (However, I haven't tested the tutorial, as I'm not sure the right way to do that.)

Note: There are a few "too-many" pylint warnings, which I'm not sure what to do about. Perhaps we want to disable these types of warnings globally?

[obijywk]

I'll disable R0902 (too-many-instance-attributes)... that one seems annoying to work around.

For R0914 (too-many-locals) I usually think of that as an indication that we're doing too much inside of a single function, and should introduce some helper functions. It's usually pretty easy to fix by doing that. So I think I'd prefer to leave that on, and we can refactor the examples that are running into this limit to not do everything inside of main(). I think let's merge this for now, though, and fix that in a follow-up (#6 ).

In general this looks great! Thanks for fixing up all the code to support this. I made a few personal-style-preference-ish cleanups (and fixed a few typing comment mistakes) that I'll push as I merge this.

I created a few more issues (#4 #5 ) for cleanups I think we should make as a follow-up to this. The mostly-mechanical changes you made to the examples to support this are the right way to start, but I think we can now go back and simplify the code for many of the examples now that they have a locations list to work with.

[jaylorch]

Thanks! My next plan is to replace the location set (which is currently passed in as just a list of points) with an abstract Grid class. This way, using subclasses RectangularGrid, FlatToppedHexGrid, and PointyToppedHexGrid, it would represent a not-necessarily-rectangular grid. The idea is that the Grid would have its own methods for adjacency, rotation, etc. So that seems like a good place to put the print routine to fix issue #4.