test(wqp): parametrize the what_* tests + factor shared assertions

[thodson-usgs]

Draft. A test-only readability refactor — no production code, no coverage change.

What

The eight WQP what_* tests (test_what_sites, test_what_organizations, …)
were near-identical: each issued the same statecode=US:34 /
characteristicName=Chloride query against its own service endpoint and
repeated the same six assertions, differing only by:

• the service path segment (Station, Organization, Activity, …),
• the response fixture, and
• the expected df.size.

This collapses them into a single @pytest.mark.parametrized test_what_query
driven by a _WHAT_CASES table, and factors the repeated metadata block
(md.url / query_time / mock_header / comment) into a
_assert_wqp_metadata helper now shared with the get_results /
get_results_WQX3 tests as well.

Why it's safe

• Same coverage. All eight wrappers still run — each as its own parametrized
  case with a readable id, so a failure still points at the exact wrapper:

  test_what_query[what_sites]
  test_what_query[what_organizations]
  ... (8 total)
  

  Each case still asserts the correct service URL, parsed DataFrame, size, and
  metadata.
• −90 lines (+59 / −149), and adding a new what_* wrapper is now one row
  in _WHAT_CASES instead of a ~16-line copy-paste.
• 15 tests pass offline; ruff / mypy clean.

Scope

WQP only — the clearest case from the suite review. A similar parametrize-refactor
of nwis_test.py::TestMetaData (6 near-identical site_info cases) is a possible
follow-up, but it lives in a file currently being changed by #335, so I left it
out to avoid churn.

🤖 Generated with Claude Code