chore(release): release v0.4.4

[acgetchell]

• Bump crate, utility package, citation, and documentation metadata
• Publish the v0.4.4 changelog and release performance comparison
• Refresh reproducible LU-solve benchmarks and archived performance reports
• Include exact-conversion and LDLT performance improvements
• Harden version-sync checks and generated Markdown formatting

Summary by CodeRabbit

• New Release

  • Released version 0.4.4 with updated package metadata and release resources.

• Bug Fixes

  • Improved exact-arithmetic conversion failure classification and Bareiss scale alignment for exact solving.
  • Improved symmetric LDLᵀ factorization error handling for invalid/near-zero pivots.

• Performance

  • Updated LU-solve benchmarks to recompute per iteration (reducing batching effects at small sizes).
  • Refreshed performance reports and expanded comparisons across additional linear-algebra implementations.

• Documentation

  • Updated benchmarking, performance, releasing, and citation/changelog links for v0.4.4.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: e17e7163-bb35-48bd-aed2-32511843e57a

📥 Commits

Reviewing files that changed from the base of the PR and between 7019b82 and 2e1eb2e.

📒 Files selected for processing (6)

• CHANGELOG.md
• scripts/check_docs_version_sync.py
• scripts/postprocess_changelog.py
• scripts/tests/test_check_docs_version_sync.py
• scripts/tests/test_postprocess_changelog.py
• src/exact.rs

🚧 Files skipped from review as they are similar to previous changes (5)

• scripts/postprocess_changelog.py
• scripts/check_docs_version_sync.py
• scripts/tests/test_check_docs_version_sync.py
• CHANGELOG.md
• src/exact.rs

---

📝 Walkthrough

Walkthrough

The v0.4.4 release updates package metadata, documentation, benchmark measurement and reporting workflows, validation scripts, archived performance artifacts, and numerical exact-arithmetic and LDLT behavior.

Changes

Release and correctness updates

Layer / File(s)	Summary
Release metadata and documentation AGENTS.md, CHANGELOG.md, CITATION.cff, Cargo.toml, pyproject.toml, README.md, docs/archive/performance/*	Release metadata, links, changelog content, benchmark snapshots, and archived performance documentation are updated for v0.4.4.
Benchmark measurement and publication benches/vs_linalg.rs, docs/BENCHMARKING.md, scripts/criterion_dim_plot.py, scripts/bench_compare.py, scripts/tests/*, docs/assets/bench/*	LU-solve benchmarks use direct iter timing, publication commands select one metric, and provenance formatting and tests are updated.
Performance report regeneration docs/PERFORMANCE.md	The report adds v0.4.4 provenance, confidence intervals, and expanded nalgebra/faer comparisons across dimensions.
Release validation automation justfile, docs/RELEASING.md, scripts/check_docs_version_sync.py, scripts/postprocess_changelog.py, scripts/tests/*	Documentation version checks, benchmark current-tag validation, changelog formatting, and Dependabot footer stripping are added or revised.
Exact arithmetic and LDLT classification src/exact.rs, src/ldlt.rs, tests/vs_linalg_inputs.rs	Exact conversion reasons and Bareiss scale alignment are refined, LDLT pivot failures are centralized, and compatibility-dependent determinant conversion is adjusted.

Estimated code review effort: 4 (Complex) | ~45 minutes

Possibly related PRs

• acgetchell/la-stack#149 — Modifies related exact-arithmetic conversion and Bareiss scale handling.
• acgetchell/la-stack#70 — Modifies the benchmark comparison and provenance-reporting workflow.
• acgetchell/la-stack#129 — Modifies the overlapping LDLT factorization path.

Suggested labels:breaking change, documentation, enhancement

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title is concise and accurately summarizes this release-focused changeset for v0.4.4.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch release/v0.4.4

---

_{Comment @coderabbitai help to get the list of available commands.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 97.86%. Comparing base (e736c5f) to head (2e1eb2e).
✅ All tests successful. No failed tests found.

Additional details and impacted files

@@ Coverage Diff @@## main #188 +/- ##
==========================================
- Coverage 97.87% 97.86% -0.02% 
==========================================
Files 8 8 Lines 4947 4969 +22 ==========================================
+ Hits 4842 4863 +21 - Misses 105 106 +1 

Flag	Coverage Δ
unittests	97.86% <100.00%> (-0.02%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

[coderabbitai]

Actionable comments posted: 5

🧹 Nitpick comments (1)

scripts/tests/test_postprocess_changelog.py (1)

90-111: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Cover the postprocess_text() integration path.

These tests exercise _strip_dependabot_metadata() directly but do not verify the new call from postprocess_text(). Add an assertion through postprocess_text() so wiring and ordering regressions are detected.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@scripts/tests/test_postprocess_changelog.py` around lines 90 - 111, Add an
integration test in TestDependabotMetadata that invokes postprocess_text() with
Dependabot metadata and asserts the footer is removed while the expected
changelog content remains. Keep the existing direct _strip_dependabot_metadata()
coverage and ensure the assertion verifies the new call is wired into
postprocess_text() in the correct order.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In `@benches/vs_linalg.rs`:
- Around line 222-244: The three LU solve benchmarks use inconsistent input
ownership, so their timings include different copy costs. Update
la_stack_lu_solve, nalgebra_lu_solve, and faer_lu_solve to use the same
input-copying or borrowing protocol for both matrices and right-hand sides,
while preserving equivalent representative inputs. Regenerate the README and
release artifacts, and record comparable results using the same command,
features, inputs, and environment.
In `@CHANGELOG.md`:
- Around line 8-109: Regenerate the v0.4.4 section through the repository’s
prescribed changelog workflow instead of editing CHANGELOG.md directly. Preserve
the release entries while using the workflow’s generated formatting, provenance,
and reproducibility behavior.
In `@scripts/check_docs_version_sync.py`:
- Around line 261-265: Update _README_TAG_LINK_RE so version and revision
matches cannot terminate at a word boundary inside a longer tag; require the
token to end at “/” or an explicit non-tag-character boundary that excludes dots
and other semver/tag characters. Add a regression test covering dotted or
otherwise non-semver tags such as v1.2.3.4, ensuring they are not accepted as
v1.2.3.
In `@scripts/postprocess_changelog.py`:
- Around line 767-809: The _strip_dependabot_metadata function currently removes
matching YAML markers inside fenced code examples. Track fenced-code state while
scanning lines and only strip Dependabot metadata when the separator and
updated-dependencies block are outside fences; preserve all fenced content
unchanged. Add a regression test through postprocess_text covering a fenced YAML
example containing ---, updated-dependencies:, and ...
In `@src/exact.rs`:
- Around line 1298-1310: Update the public exact-solve documentation near the
existing independent-scaling description to state the conditional invariant:
matrix and RHS use a shared minimum scale when their independent scale exponents
differ by at most MAX_SHARED_SCALE_GAP_BITS (64), otherwise they retain
independent scales. Keep the documentation aligned with the logic in
ScaleExponent::for_decomposed and the matrix_scale/rhs_scale selection.
---
Nitpick comments:
In `@scripts/tests/test_postprocess_changelog.py`:
- Around line 90-111: Add an integration test in TestDependabotMetadata that
invokes postprocess_text() with Dependabot metadata and asserts the footer is
removed while the expected changelog content remains. Keep the existing direct
_strip_dependabot_metadata() coverage and ensure the assertion verifies the new
call is wired into postprocess_text() in the correct order.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 18ba6368-686b-487e-9b91-4734f9651f97

📥 Commits

Reviewing files that changed from the base of the PR and between e736c5f and 7019b82.

⛔ Files ignored due to path filters (4)

• Cargo.lock is excluded by !**/*.lock
• docs/assets/bench/vs_linalg_lu_solve_median.csv is excluded by !**/*.csv
• docs/assets/bench/vs_linalg_lu_solve_median.svg is excluded by !**/*.svg
• uv.lock is excluded by !**/*.lock

📒 Files selected for processing (25)

• AGENTS.md
• CHANGELOG.md
• CITATION.cff
• Cargo.toml
• README.md
• benches/vs_linalg.rs
• docs/BENCHMARKING.md
• docs/PERFORMANCE.md
• docs/RELEASING.md
• docs/archive/performance/README.md
• docs/archive/performance/v0.4.3-vs-v0.4.2.md
• docs/assets/bench/vs_linalg_lu_solve_median.provenance.json
• justfile
• pyproject.toml
• scripts/bench_compare.py
• scripts/check_docs_version_sync.py
• scripts/criterion_dim_plot.py
• scripts/postprocess_changelog.py
• scripts/tests/test_bench_compare.py
• scripts/tests/test_check_docs_version_sync.py
• scripts/tests/test_criterion_dim_plot.py
• scripts/tests/test_postprocess_changelog.py
• src/exact.rs
• src/ldlt.rs
• tests/vs_linalg_inputs.rs