Add single-case benchmark runner

[bernalde]

Summary

• Add benchmarks/run_single_case.py, a no-artifact single-case SciPy/Pyomo debug runner.
• Reuse current benchmark scenarios, adapters, grid-style dotted parameter overrides, discretization flags, warmstart, ramp constraints, and solver time guards.
• Wire solver verbosity through pyomo_adapter(..., tee=...) and document the new runner in benchmarks/README.md.
• Add help/parser and monkeypatched adapter-level coverage for the new runner.

Tests run

• python -m ruff check benchmarks/run_single_case.py benchmarks/adapters.py tests/test_benchmarks.py -> passed (All checks passed!)
• python -m ruff format --check benchmarks/run_single_case.py benchmarks/adapters.py tests/test_benchmarks.py -> passed (3 files already formatted)
• python -m pytest tests/test_benchmarks.py -q -> passed (17 passed in 5.64s)
• python -m pytest tests/ -n auto -v -m "not notebook and not pyomo" --ignore=tests/test_pyomo_models -> passed (254 passed in 469.66s)

Notes

• Did not run the Pyomo-marked live solver suite locally; it requires the optional IPOPT/IDAES stack and is treated as optional in CI. The new runner path is covered with monkeypatched adapter tests, and the adapter tee propagation is covered directly.
• Did not run notebook-marked tests; this change does not touch notebooks.

Closes #17

[bernalde]

Reviewed the PR diff, benchmark adapters, runner code, tests, and benchmark documentation. I did not find blocking issues, nonblocking issues that need an inline comment, or open questions.

Tests run:

• python -m ruff check benchmarks/run_single_case.py benchmarks/adapters.py tests/test_benchmarks.py -> passed
• python -m ruff format --check benchmarks/run_single_case.py benchmarks/adapters.py tests/test_benchmarks.py -> passed
• python -m pytest tests/test_benchmarks.py -q -> passed, 17 passed in 5.04s
• python -m pytest tests/ -n auto -v -m "not notebook and not pyomo" --ignore=tests/test_pyomo_models -> passed, 254 passed in 383.22s

I did not run the Pyomo-marked live solver suite because it requires the optional IPOPT/IDAES solver stack and is handled separately in CI. The PR is merge-ready from this review.

[bernalde]

Comment follow-up status:

• Commits pushed: none in this pass; branch was already up to date. Existing PR commit remains 6dcf222 Add single-case benchmark runner (#17).
• Main changes made: none. I fetched thread-aware review data and REST comment lists; there are no unresolved inline review threads and no top-level discussion comments to address.
• Tests run and results: no tests run in this pass because no code or documentation changes were made.
• Comments intentionally not addressed: the only review entry is the earlier self-authored review summary, which is informational and does not request changes.
• Remaining risks or follow-up items: none from PR comments.