Simplify defunct stubs and fix O(n^2) pagination concats

[thodson-usgs]

Summary

A few targeted simplifications and bug fixes turned up by a code review of recent commits.

Cleanup

• nwis.py: Reduce 5 defunct function stubs (get_qwdata, get_discharge_measurements, get_gwlevels, get_pmcodes, get_water_use) to minimal **kwargs signatures — they only raise NameError, so the full typed parameter lists were dead weight.
• samples.py: Replace the 230-line get_usgs_samples pass-through (which duplicated 22 parameters and their docstrings verbatim from waterdata.get_samples) with a small **kwargs wrapper.

Performance fixes

• nwis.py::_read_json: Fix O(n²) pd.concat in a per-site loop — now collects DataFrames in a list and concats once. Also drops an unnecessary str(record_json).replace("'", '"') → StringIO → pd.read_json round-trip, since record_json is already parsed Python data and pd.DataFrame(record_json) works directly.
• waterdata/utils.py::get_stats_data: Same collect-then-concat fix for the pagination loop — previously the concat was inside the while next_token: loop.

Bug fixes

• nwis.py::get_record: Restore the missing elif service == "peaks": dispatch branch. The docstring has long advertised "peaks" as a supported service but the branch was dropped in an old lint/format pass (#164), so get_record(service="peaks") has been raising TypeError: peaks service not yet implemented until now.
• Demo notebooks (Peaks, Ratings, Statistics, WaterUse under demos/hydroshare/): Fix a SyntaxError from an import line that got joined without a newline (from dataretrieval import nwisfrom dataretrieval import waterdata), apparently introduced by ruff-format commit #219.
• WaterUse_Examples.ipynb: Comment out two cells that still referenced a pennsylvania variable whose assignment had been commented out as # [Defunct].
• NWIS_demo_1.ipynb, WaterData_demo.ipynb: nb-clean pass (stripped stale version metadata only).

Test plan

• pytest tests/ — 116 passed (excluding the untracked ngwmn_test.py)
• Execute all 16 canonical demo notebooks via jupyter nbconvert --execute against live USGS APIs — 16/16 pass after the fixes (11/16 before)
• Sanity-check that the _read_json and get_stats_data paths are still exercised by existing tests (tests/nwis_test.py, tests/waterservices_test.py, tests/waterdata_test.py)

🤖 Generated with Claude Code