Conversation
Address the actionable items from the CRAN-readiness review: 1. ASCII-ify all roxygen/comments in R/ (em-dash, arrows, approx, times, Greek letters, set/minus signs) and regenerate man/. This removes the non-ASCII characters that caused "LaTeX errors when creating PDF version" -- the Rd->LaTeX conversion now succeeds (verified; the local PDF step only fails for lack of pdflatex in the sandbox). 2. Add runnable @examples to pts() and the pts-methods group (forecast / print / fitted / residuals); `R CMD check` examples run OK. 3. Version -> 0.1.0 (DESCRIPTION, NEWS, python/pyproject.toml); Date updated. 4. .Rbuildignore: exclude .mypy_cache, .ruff_cache. 5. .Rbuildignore: exclude PYTHON_PORT.md, muse.Rcheck, *.tar.gz, .RData, .Rhistory, cran-comments.md. 6. src/Makevars: strip debug symbols on Linux -> installed size NOTE is now OK (18 MB -> under 5 MB). Documented the residual size in the new cran-comments.md. 9. Move Rcpp from Depends to Imports (already importFrom'd in NAMESPACE). 10. Date -> 2026-06-19. Title and Description expansion (items 7, 8) left for the maintainer. R CMD check --as-cran: no remaining package-level ERROR/WARNING/NOTE; the ones still shown are sandbox-environmental (no pdflatex / qpdf / tidy / V8, locale, clock skew). Tests and vignettes pass. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Close the gap where the Python front-end only accepted scalar non-seasonal ARMA orders. `orders` now takes length-2 ar/ma vectors for SARMA(p,q)(P,Q)_s, exactly as R does -- with one deliberate API choice: the seasonal lag comes from the separate top-level `lags`, NOT from `orders` (passing `orders["lags"]` raises a clear error). - translate.orders_to_uc(): port of R's .pts_orders_to_uc (minus the orders$lags branch) -- normalises ar/ma to per-lag vectors, derives lags = [1, seasonal_lag] for length-2, drops trailing zero blocks while keeping the non-seasonal block, validates non-negativity and the one-seasonal-lag limit. - translate.arma_spec(): builds the (arma_orders tuple, irregular string), 2-tuple/arma(p,q) or 5-tuple/arma(p,q,P,Q,s). - PTS.fit(): thread the per-lag vectors through pts_to_uc, the engine (_engine now takes the irregular-option string), the selector, and forecasting; $orders reports scalar ar/ma for non-seasonal and vectors for SARMA, like R's orders(). Parity: 20/20 pts specs match R within 1e-6 (worst 3.5e-9), including the new SARMA(1,0)(1,0)_12 / (0,1)(0,1)_12 / mixed cases. Forecast, simulate, and select all verified for SARMA. Functionality suite +4 tests (38 total). README updated to document seasonal orders + the lags-is-separate rule. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
The doc was R-only. Add the Python package now that it exists: - Package layout extended with src/python/musecpp2py.cpp, src/muse_compat.h, and the full python/ tree (pyproject/CMakeLists, src/muse/core/*.py, tests/). - New "Two front-ends, one engine" subsection: a table mapping the R and Python bindings (musecpp2R.cpp vs musecpp2py.cpp; .UCompC/.UCompARMAC vs _musecore.ucomp/ucomp_arma; pts() vs PTS; R CMD INSTALL vs scikit-build), and how the shared engine + muse_compat.h Rprintf shim + .Rbuildignore keep the two builds from colliding. - New "Python front-end" section: R<->Python module mapping table, the two engine entry points, deliberate differences (coef source, orders/lags API, no formula/auto-distribution, optional smooth/greybox deps), and the test split (functionality in CI, parity vs R out of CI). - Mental-model pointer to the above. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
"muse" is already registered on PyPI (a yanked 2015 stub), so the distribution is named "muse-py". The importable package is unchanged: pip install muse-py; import muse (wheel.packages still src/muse) -- the common pip-name != import-name pattern (cf. scikit-learn/sklearn). - pyproject: [project] name = "muse-py". - README: document the install/import mapping in the Installation section. Verified: wheel builds as muse_py-0.1.0, installs as "muse-py" (pip show), imports as `muse`, and fits. Nothing else changes -- import code, tests, CMake, the engine, and the R package (CRAN: muse) are untouched. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
…n smooth) Path filters (run each check only when the relevant code changes): - R-CMD-check.yaml: triggers on R/, src/ (shared engine) EXCEPT src/python/, man/, tests/, vignettes/, DESCRIPTION, NAMESPACE, the workflow itself. - python-check.yaml: triggers on python/, src/ (engine + src/python/), the workflow itself. (A src/ engine change runs both, as it should.) Release pipeline (build -> test -> draft -> manual publish), mirroring smooth's: - build-release-wheels.yml (manual / workflow_dispatch): cibuildwheel across Linux/macOS/Windows x cp310-cp313, smoke-tests each wheel (test_functionality.py), builds the sdist, and stages everything on a *draft* GitHub release (no tag, no PyPI). Re-running replaces the draft's assets. muse-specific delta vs smooth: installs system Armadillo (EPEL / brew / vcpkg) since muse uses the system Armadillo rather than a vendored copy. - release.yml (on release: published): downloads the staged assets and uploads them to PyPI (muse-py) via Trusted Publishing + PEP 740 attestations. So the flow is: dispatch build-release-wheels -> review the draft -> click "Publish release" -> release.yml ships to PyPI. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
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
1 participant
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.
No description provided.