Regenerate R parity cache for NNS 13.2#135
Merged
Merged
Conversation
…rity cache)
R's dy.d_ was reconciled in NNS 13.2 (dy.dx-style locally-adaptive step +
NNS.stack on the equal-weight synthetic regressor), and the parity cache was
regenerated accordingly. The shipped Python dy_d still used the pre-13.2
finite-difference engine, so the 8 test_dy_d_* parity tests diverged from the
new cache (e.g. mean-wrt gave 7.56 vs R's 2.667).
Point the public dy_d at the reconciled engine that already backed dy_d_best,
which matches R 13.2 to machine precision across mean/median/last/obs/apd and
the mixed-derivative corners. dy_d and dy_d_best now share one _dy_d_reconciled
implementation (differing only in the default messages value), and the dead
pre-13.2 engine (_dy_d_scalar and its band/mixed/aggregation helpers) is
removed. _dy_d_stack_estimates is kept as it is exercised directly by a test.
The apd+mixed invalid-input test now expects R's own message
("Mixed Derivatives are only for 2 IV"); the invariant (apd+mixed rejected) is
unchanged.
sync/nns_source.json advances r_commit/r_version to the 13.2 behavioral-truth
commit (ff7f4f4); the vendored native snapshot is unchanged because the
reconciliation touched only R sources.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01AxBV5HmXHY5nCUvW9SwDvb
PR #135's automation built NNS_13.2.tar.gz and the committed parity cache from R commit 8f236b9 ("NNS 13.2 Beta"), a descendant of the dy.d_ reconciliation merge. Record that commit as the behavioral-truth reference (was ff7f4f4) and clarify that re-syncing the vendored native snapshot to it is a separate NNS-core snapshot update. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AxBV5HmXHY5nCUvW9SwDvb
CI's mypy gate flagged the sole real error [no-any-return] at the shared engine's single-wrt return: the per-regressor reducer was typed Any inside spans, so returning its output leaked Any. Give it a concrete _DyDReducer alias. Annotation-only; no runtime change. (The same latent Any predates this branch in dy_d_best; the refactor moved it onto the mypy-checked path.) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01AxBV5HmXHY5nCUvW9SwDvb
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Automatic parity-cache refresh from the R source of truth.
OVVO-Financial/NNS8f236b97ba4612c93c09a949e7e14ec20dcb6c1413.2NNS_13.2.tar.gzNNS_13.2.zipfailurefailureThe exact R source package was installed and the committed parity cache was regenerated. Any remaining Python/R parity mismatches are retained in the workflow diagnostics and should be repaired against this R-authored baseline.