[BUGFIX] Dimension update in LayoutOptimizationBase

[misi9170]

As described in #1033 , there was a section of code that was missed in the dimension reduction from 5D to 4D as part of the release of Floris v4. The missed code is very much an edge case, where the minimum yaw angle is set to be a stricly positive number (and vice versa, if the maximum yaw angle is set to be strictly negative).

This PR updates that section of code, and also adds an integration test that tests for these unusual situations.

Test results, if applicable

• Existing tests pass.
• New tests added (yaw_optimization_integration_test.py) fail without changes to YawOptimizationBase (with an "axis out of bounds" error due to indexing a non-existent dimension 3---see screenshot below).
• New tests pass with changes to YawOptimizationBase.

Notes

1. Since this is such an edge bug, I don't think we need to target main and release a patch. This can simply be bundled into the next minor release of FLORIS.
2. I'm also using this PR as an opportunity to fix a typo in the axis labels of one of the yaw optimization examples.

Screenshot of failing integration test prior to fixing dimensions:

[paulf81]

Looks good, reviewed the change and agree the new test covers the condition. All pass on my end.