[Alex0420W]

… (v2.0)

First v2.x major-version release. Replaces selene's global coupling product (PSR-distance × ridge-distance multiplicative) with per-cell 2 km-radius EVA-disc PSR area integration, mirroring NASA's documented walking-EVA framing (Lawrence 2025 NTRS 20250008952; A3GT priorities at LPSC 2025; EVA-EXP-0070 Rev D). Cold-class threshold (110 K) is the standard water-ice stability ceiling from Paige et al. 2010 + Hayne et al. 2015. Weight slot is unchanged at 0.17 (same as v1.x's coupling slot) so the v1.9 → v2.0 comparison is methodology-only.

Adds Wueller per-launch-year illumination cross-check using the already-bundled SunDays25-32 columns from the Zenodo deposit's DBF. Of 55 matched sites at 20 m, 22 pass NASA's HLS continuous-lit threshold (≥ 10 Earth-days) for the 2028 Artemis IV launch window; 52 pass HLS in at least one launch year 2025-2032. Output schema lands at data/outputs/v2/wueller_comparison_with_launch_year.json.

20 m headline (default-weights, 8 active criteria with EVA-disc): 69 sites (was 69), 55 / 69 = 79.7 % within 5 km of an in-scope Wueller site (was 56 / 69 = 81.2 %), median matched-pair distance 1.84 km (was 1.69 km). 59 of 69 v2.0 sites are within 10 m of a v1.x site (identical re-pick); 7 reranked 0-5 km within their polygon; 3 reranked > 5 km — concentrated in N2, HW, MMP. Per-region site counts are unchanged from v1.9.

240 m baseline (full 240 m run, 70 sites): 56 / 70 = 80.0 % within 5 km, median 1.88 km, 45 / 73 = 61.6 % Wueller-side.

Continuous-lit illumination metric and rock-abundance safety filter were scoped for v2.0 but blocked on public data availability (PGDA hourly cube not distributed; all public Diviner rock-abundance products clip at ±60-70°, excluding the pole). Both deferred indefinitely pending data release. v2.0 ships with what's actually achievable from data already in the repo.

Summary

What this PR does, in 1–3 sentences.

Motivation

Why this change is worth making (link the issue if there is one).

Checklist

• ruff check . passes
• ruff format --check . passes
• pytest passes locally
• New behaviour has a test (or an explicit # noqa reason if not)
• Public-facing changes (CLI flags, criterion math, file layout) are
  reflected in the README or the relevant module docstring
• No real-data fixtures were committed; new tests run in CI without
  downloaded data (or are guarded with pytest.mark.skipif)

Scope notes

• Things this PR explicitly does not address (so reviewers know what
  to expect in follow-ups).