Include Tests

[hammy4815]

Tests are needed mainly for:

1. Assembly of different material cases
2. Solving and the curl relations

But could also be useful for other features currently being implemented.

[smartalecH]

We should test for accuracy and convergence too (rather than just code correctness).

IIRC, the original paper has a few different examples with effective indices, right? We can run the same problem at various resolutions and ensure they converge. (If we ever add subpixel smoothing, we can use the same tests to check).

[smartalecH]

There are a few more good tests here (which will require PML)

https://opg.optica.org/oe/fulltext.cfm?uri=oe-12-25-6165&id=81978

As a first test, it might be good to just use a few examples from Lukas's book (lots of replication already). ie a simple strip and rib waveguide, qTE and qTM, for a couple of different wavelengths.