Bump version to 3.5.1

[igerber]

Summary

• Bump version 3.5.0 → 3.5.1 across pyproject.toml, rust/Cargo.toml, diff_diff/__init__.py, diff_diff/guides/llms-full.txt, and CITATION.cff (date-released: 2026-06-02).
• Promote the populated [Unreleased] CHANGELOG block to ## [3.5.1] - 2026-06-02 (fresh empty [Unreleased] retained on top) and add the [3.5.1] comparison link.

This is a release-cut PR — no source/methodology/test logic changes; only version strings and the CHANGELOG section header/link.

What's in 3.5.1 (already merged; documented in the [3.5.1] CHANGELOG block)

Patch release — enhancements, validations, and fixes to existing surfaces (no new estimator):

• SyntheticControl cross-validation + inverse-variance V-selection (ADH 2015 / Abadie 2021) (SyntheticControl: cv (out-of-sample) + inverse-variance V-selection (ADH 2015 / Abadie 2021) #523).
• Firpo & Possebom (2018) SCM-inference paper review on file — PR-A groundwork for CI-by-test-inversion (docs: Firpo & Possebom (2018) paper review — SCM CI by test inversion (PR-A) #524).
• HeterogeneousAdoptionDiD.fit() UX: extensive-margin warning + covariates= NotImplementedError (HAD fit(): extensive-margin warning + covariates= NotImplementedError #525).
• Covariate names colliding with reserved structural terms now raise ValueError instead of silently corrupting the design (Reject covariate names colliding with reserved structural terms (DifferenceInDifferences, MultiPeriodDiD, TwoWayFixedEffects) #518).
• EfficientDiD methodology-review-tracker promotion → Complete, with a sieve outcome-regression upgrade (EfficientDiD methodology validation (PR-B): sieve outcome-regression upgrade + Verified Components #521).
• CI + local AI PR-reviewer model upgraded gpt-5.4 → gpt-5.5 (Upgrade AI PR reviewer model: gpt-5.4 → gpt-5.5 (CI + local) #522).

Methodology references (required if estimator / math changes)

• Method name(s): N/A — no methodology/math code changes in this PR (version bump only). The underlying work landed in the feature PRs listed above and is documented in docs/methodology/REGISTRY.md.
• Paper / source link(s): N/A for this PR
• Any intentional deviations from the source (and why): None in this PR

Validation

• Tests added/updated: No test changes in this PR (release cut). Feature PRs Reject covariate names colliding with reserved structural terms (DifferenceInDifferences, MultiPeriodDiD, TwoWayFixedEffects) #518–HAD fit(): extensive-margin warning + covariates= NotImplementedError #525 carried their own test coverage, already merged and green on main.
• Backtest / simulation / notebook evidence (if applicable): N/A for this PR

Security / privacy

• Confirm no secrets/PII in this PR: Yes

---

Generated with Claude Code

[github-actions]

Overall Assessment

✅ Looks good. No unmitigated P0 or P1 findings.

Executive Summary

• The diff is release bookkeeping only: version/date updates plus promoting CHANGELOG.md [Unreleased] content to 3.5.1.
• No estimator code, inference logic, assumptions, defaults, weighting, or variance/SE code changed.
• Version surfaces are consistent across pyproject.toml:L7, diff_diff/__init__.py:L301, rust/Cargo.toml:L3, CITATION.cff:L10-L11, and diff_diff/guides/llms-full.txt:L5.
• Registry/TODO spot-checks align with the methodology claims named in the release notes; deferred SyntheticControl ADH follow-ups are tracked in TODO.md:L88.
• git diff --check reported no whitespace/errors.

Methodology

Severity	Impact	Concrete fix
N/A	No methodology-affecting code changed in this PR. The changelog references SyntheticControl/HAD methodology already documented in docs/methodology/REGISTRY.md:L1990-L2006 and docs/methodology/REGISTRY.md:L2933.	No action required.

Code Quality

Severity	Impact	Concrete fix
N/A	Metadata-only changes are straightforward and internally consistent.	No action required.

Performance

Severity	Impact	Concrete fix
N/A	No runtime code paths changed.	No action required.

Maintainability

Severity	Impact	Concrete fix
N/A	Release metadata appears synchronized across Python, Rust, citation, changelog, and LLM guide surfaces.	No action required.

Tech Debt

Severity	Impact	Concrete fix
N/A	No new deferred work introduced. Existing deferred SyntheticControl ADH tail items remain tracked in TODO.md:L88.	No action required.

Security

Severity	Impact	Concrete fix
N/A	No secrets, credentials, dependency changes, or executable logic introduced.	No action required.

Documentation/Tests

Severity	Impact	Concrete fix
N/A	Changelog and guide version were updated for the release. No tests are needed for this metadata-only release cut.	No action required.