Add adaptive horizon finder criteria

[geoffrey4444]

Proposed changes

Adds a base class and a simple example derived class for criteria for adjusting the resolution adaptively when finding an apparent horizon. A future PR will implement the actual criteria intended to be used in real binary-black-hole simulations. Another future PR (or set or PRs) will take care of implementing adaptivity in the apparent horizon finder.

The design and implementation are based on how adaptive mesh refinement criteria are implemented in spectre but enables the criteria to work with output of a horizon find (Strahlkorper and FastFlow::IterInfo) instead of an ElementId.

NOTE: I used Cursor generative AI (tab completion, review, fixing compiler errors) when preparing this PR. I examined all code suggestions, and to the best of my knowledge, this PR contains no code either copyrighted by someone else or used in violation of any code's license agreement. Please note in your review any concerns about this.

Upgrade instructions

Code review checklist

• The code is documented and the documentation renders correctly. Run
  make doc to generate the documentation locally into BUILD_DIR/docs/html.
  Then open index.html.
• The code follows the stylistic and code quality guidelines listed in the
  code review guide.
• The PR lists upgrade instructions and is labeled bugfix or
  new feature if appropriate.

Further comments

[geoffrey4444]

The clang-tidy failure did not come up when I ran the test locally, but I'll make the fix

[knelli2]

Looks good! Please add a fixup commit for these changes, and also fix the small conflict.

[knelli2]

Do we want to allow this as part of the "standard criteria"? This means it'll probably be available to users in the input file if this type alias is used in the metavars. I know this is what AMR does, but seems like this is just a testing criterion.

[geoffrey4444]

I'm fine not including this in the standard criteria list. If users really want it for some reason we can always add it later

[knelli2]

You can squash everything

[knelli2]

Got approval from @nilsvu on slack that I can just merge this without his review in the interest of time. Also test failure is unrelated