Off by 1 error in create_rtree about mesa-geo

[Corvince]

Great find and bug descripton @bpmweel !

I guess thats why tests exist ;) May I reuse your code to base a test on it?

I would also love to hear more about your experience with mesa-geo so far, especially what was/is not so nice and should be improved. I honestly didn't expect anyone to use mesa-geo already and this is very motivating to keep me working on it.

[bpmweel]

Of course! Please go ahead and use the code to make a test.

So far using mesa-geo has been relatively good, this is my first time using mesa and mesa-geo so the most of the struggle is in learning a new framework I guess. You can check my fork over at https://github.com/DynaSlum/mesa-geo for some small changes I made:

1. The version is currently only a string, which pip or conda gives an error for. I suggest using something like 0.1.0-alpha or something to get some semantic version going.
2. In add_agent you use issubclass which has been deprecated in favor of isinstance
3. Updated some dependency versions

Another suggestion I would make is to have create_agents_from_GeoJSON return the list of created agents. When creating agents in this way you most likely want to add them to the Mesa scheduler, but now I have to manually fish out the new agents from the grid.