closes #109

closes #40

closes #85

Also moved the DSL macros into a separate file. Still need to write
tests for all of this.

This closes #47

This closes #24.

This commit includes a `sed` pass over all files to change all
occurrences of the former spelling "Persephone" to "Persefone":
`sed -ie "s/Persephone/Persefone/" *`

Multiple runs and parameter scanning works, but parallelising was (as
always) a bit trickier than expected... Will try to get this working
soon, but for now I need to focus on more important things.

This is a major commit with quite deep-reaching changes in the code
base. It enables the user to specify multiple values for a model
parameter, which will then be combined into a full-factorial
simulation experiment. Everything seems to work as intended, although
some refactoring would be good, and we still need to add parallelisation
for issue #42.

And yes, I coded all of this between 23:30 and 03:00 on a Saturday
night 😅 🙈

Using the GLOBAL_RNG introduces global state, which must be avoided
to preserve reproducibility. Therefore, all Persephone code must use
`model.rng` whenever calling `rand()`/`shuffle!()`/etc.

This is a major change to the codebase. I'm not entirely convinced yet
that it actually makes sense, but overall I think the benefits
outweigh the problems. Closes issue #37.

This is not yet in a working state.

This is a major commit, probably quite a bit too big...
However, all the pieces here depended on each other, so
I had to get it all working at the same time.

The model now includes:

	1. reading in map data
	2. creating an Agents.jl model object
	3. initialisation and stepping functions for the whole model
	4. a basic skeleton of nature.jl
	5. an example animal species (the Wolpertinger ;-) )