Skip to content

Bump the version to 2.2.0 - #194

Merged
OVVO-Financial merged 2 commits into
mainfrom
claude/nns-r14-parity-fidelity-rgag0e
Nov 9, 2026
Merged

OVVO-Financial merged 2 commits into
mainfrom
claude/nns-r14-parity-fidelity-rgag0e

Conversation

@OVVO-Financial

Copy link
Copy Markdown
Owner

Summary

Bumps ovvo-nns from 2.1.0 to 2.2.0. This release carries the R NNS 14.0 bit-level parity work merged in #193.

Changes

The version moves together everywhere it is recorded:

  • pyproject.toml
  • nns.__version__ in src/nns/__init__.py
  • the README's "Current version" row
  • the project entry in uv.lock, regenerated with uv lock; no other package changed

Checks

  • scripts/check_version_consistency.py: Version consistency OK: 2.2.0 in pyproject, nns.__version__, and README.
  • scripts/check_release_provenance.py --tag v2.2.0: OK. sync/nns_source.json already records R 14.0 (r_commit 7b07a837, r_version 14.0) and the NNS-core commit, so the tagged release gate passes.
  • pytest tests/tools: 13 passed.

Suggested release notes for v2.2.0

  • R 14.0 fidelity: results now match R NNS 14.0 to the last bit across the public API, with and without the native extension (verified on x86-64 Linux). See docs/parity.md.
  • Seeded results change: random_seed = s (seed for nns_causation) now reproduces R's set.seed(s) stream. Outputs saved from earlier seeded runs will differ.
  • New permutation p-values: nns_seas, nns_dep, nns_causation and causal_matrix.
  • Exact smoothing spline: nns_reg(smooth=True) now uses an exact port of R's smooth.spline. It raises R's errors where smooth.spline stops, instead of falling back to interpolation.
  • Smaller contract changes:
    • A numeric multivariate nns_reg result has no class.levels key.
    • nns_distance treats k as R does.
    • nns_rescale warns when all values are identical.

After merging

Per docs/releasing.md:

  1. Run the TestPyPI dry run (Actions → Release → publish: testpypi).
  2. Tag v2.2.0 and push the tag to publish to PyPI.

Generated by Claude Code

claude and others added 2 commits November 9, 2026 15:30
2.2.0 carries the R NNS 14.0 bit-level parity work merged in #193. The
version moves together in pyproject.toml, nns.__version__, the README's
current-version row and the uv.lock entry for the project, as
scripts/check_version_consistency.py requires. The release provenance in
sync/nns_source.json already records R 14.0 (r_commit 7b07a837), so a
v2.2.0 tag passes scripts/check_release_provenance.py.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Wws6NpuQ72C5sfEYVoVmGt
@OVVO-Financial
OVVO-Financial merged commit 440239f into main Nov 9, 2026
5 checks passed
OVVO-Financial pushed a commit that referenced this pull request Nov 13, 2026
…eproduction

inspect-r-api-update.yml regenerated the cache under setup-r's current R
release (4.6.2 on 13 November) with OpenBLAS on ubuntu-latest, not under
the environment docs/parity.md pins, so its cache moved results by an ulp
or more and could not pass the port's exact parity tests (#198). The job
now runs on ubuntu-24.04 with CRAN's R 4.5.0 build (4.5.0-3.2404.0, held),
the reference BLAS and LAPACK selected, and Ubuntu's Rcpp, RcppParallel
and jsonlite, and stops unless R reports exactly that. It regenerates
tests/invariants' entries as well as tests/parity's, so cache-only
verification does not miss them, and records the source archive's hash
in the manifest along with the commit.

r14-reproduction.yml was removed from main by the baseline job on #194
because it was not on the protected list. It is restored, pinned to the
current oracle (9c8bd21, NNS_14.0.tar.gz sha256 754dad65...), and added
to the list so unlabelled pull requests keep it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Wws6NpuQ72C5sfEYVoVmGt
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants