227 ensure passing tests form refactor rawdem

[rosepearson]

@jennan I've created a PR into your refactor branch to fix the breaking tests. Plesae note I am working a few days during the holidays when it is raiing in the mtns so I can go tramping the 3-4. Please do not feel to have to do anything until you're back at work after your RESTful restful holidays :).

• Re-add try/catch as otherwise crash when coarse DEM not overlapping the land/foreshore is used (this corner case isn't included in the tests.. perhaps I should add one at some point)
• Fix tests
  • Reverse mask logic for LiDAR foreshore as otherwise NaN values are set to zero instead of left to interpolation. This fixed all but one of the tests.
  • Updated the benchmark for test_dem_generation_westport_4.py after checking the scale and spatial distribution of numerical differences.

DESCRIPTION OF PR:

• [x] Make code change
• [x] Update tests
  • [x] Update / create new tests
  • [x] Ensure these tests have the expected behavour
  • [x] Test locally and ensure tests are passing
• [x] Ensure tests passing on GH Actions
• [x] Update documentation
  • [x] Doc strings
  • [x] Wiki
  • [x] Update package (version.py & pyproject.toml) version (if needed)

[jennan]

@rosepearson Thank you Rose for looking into this :-). I didn't realise that changing the masking logic from not (something and something_else) to (not something) or (not something else) would cause issues with NaNs. I have used this change in multiple places, I will double check this.