Release v2.7.0

[igerber]

Summary

• Bump version to 2.7.0 across __init__.py, pyproject.toml, rust/Cargo.toml
• Add [2.7.0] CHANGELOG entry with Added/Changed/Fixed sections covering all work since v2.6.1
• Add step 10b to submit-pr command to ensure PR ref exists after API creation

Methodology references (required if estimator / math changes)

• N/A — version bump only, no methodology changes in this PR

Validation

• Tests added/updated: No test changes
• N/A

Security / privacy

• Confirm no secrets/PII in this PR: Yes

---

Generated with Claude Code

[github-actions]

Overall Assessment

✅ Looks good

Executive Summary

• Methods affected: none. This diff changes release metadata and contributor workflow documentation only.
• I cross-checked the new method-specific changelog text against the methodology registry and did not find an undocumented methodology mismatch.
• The 2.7.0 version bump is internally consistent across Python and Rust package metadata.
• One non-blocking maintainability issue remains: the new PR-ref verification step checks the wrong remote in fork workflows and can falsely report that the PR ref is missing.
• No security or performance issues found.

Methodology

• Severity: None. Impact: No estimator implementation, weighting, variance/SE, identification assumption, or default behavior changes are introduced by this diff. The new release-note claims in CHANGELOG.md#L13, CHANGELOG.md#L14, CHANGELOG.md#L21, and CHANGELOG.md#L35 are consistent with existing registry entries for CallawaySantAnna WIF/sup-t at REGISTRY.md#L313, EfficientDiD at REGISTRY.md#L482, and TROP step-size/Lipschitz details at REGISTRY.md#L1400. Concrete fix: None.

Code Quality

• Severity: None. Impact: The version bump is synchronized across init.py#L182, pyproject.toml#L7, and Cargo.toml#L3. Concrete fix: None.

Performance

• Severity: None. Impact: No runtime or algorithmic code changed, so this PR does not introduce a performance regression in the library itself. Concrete fix: None.

Maintainability

• Severity: P2. Impact: The new step 10b is internally inconsistent for fork workflows. This command defines <base-remote> as upstream and <push-remote> as origin when a fork is present at submit-pr.md#L38, submit-pr.md#L39, and submit-pr.md#L40, but the new verification queries refs/pull/<number>/head on <push-remote> at submit-pr.md#L324. For forked PRs, that checks the fork instead of the target repository, so the ref check will false-negative and can trigger the unnecessary empty-commit fallback at submit-pr.md#L327. Concrete fix: Keep the push on <push-remote> at submit-pr.md#L319, but verify refs/pull/<number>/head on <base-remote> and say explicitly that PR refs live on the target repo.

Tech Debt

• Severity: None. Impact: No new deferred library tech debt is introduced by the version-bump files, and TODO.md does not mitigate or change the status of the P2 workflow issue above. Concrete fix: None.

Security

• Severity: None. Impact: No secrets, credentials, or risky command construction were added in the changed files. Concrete fix: None.

Documentation/Tests

• Severity: None. Impact: No library test gap is introduced by the version bumps themselves; the only documentation issue in scope is the fork-remote mismatch above. Concrete fix: None beyond correcting step 10b.