v1.6.8

[jhollway]

Description

Models

• Substantially accelerated net_regression() (by about 35x), resolving the CRAN NOTE about the running time of the predict() examples
  • Permutations now operate directly on the matrices rather than coercing to and from a network object each time, deferring to manynet::to_permuted() again once manynet gains a method for matrix input
  • Vectorising the matrix list for each fit now uses base operations rather than constructing a data frame
  • Results are unchanged: for a given seed, the same permutations are drawn and the same coefficients, test statistics, and p-values are returned

Checklist:

• Documentation
  • Longer functions are commented inline or broken down into helper functions to help debugging
• PR form
  • PR description above and the NEWS.md file are aligned

[codecov]

Codecov Report

❌ Patch coverage is 68.57143% with 11 lines in your changes missing coverage. Please review.
✅ Project coverage is 61.22%. Comparing base (010efde) to head (f5c6ff9).
⚠️ Report is 16 commits behind head on main.

Files with missing lines	Patch %	Lines
R/class_makes.R	0.00%	4 Missing ⚠️
R/class_models.R	0.00%	4 Missing ⚠️
R/tutorial_run.R	57.14%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #320      +/-   ##
==========================================
+ Coverage   60.51%   61.22%   +0.71%     
==========================================
  Files           9        9              
  Lines         704      717      +13     
==========================================
+ Hits          426      439      +13     
  Misses        278      278              

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

This PR releases migraph 1.6.8 and focuses on significantly speeding up net_regression() permutations to address CRAN NOTE runtime concerns around the predict() examples, while keeping statistical results unchanged for a given seed.

Changes:

• Optimized net_regression() permutation path by permuting matrices directly (with a fallback to manynet::to_permuted() once a matrix method exists).
• Reworked vectorise_list() to use base matrix assembly instead of dplyr::bind_cols(), reducing per-permutation overhead.
• Added tests to lock in permutation equivalence and delegation behavior; updated docs/examples and versioning (DESCRIPTION/NEWS).

Reviewed changes

Copilot reviewed 6 out of 7 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
R/model_regression.R	Switches permutation implementation to a fast matrix-based path; rewrites vectorisation to avoid tibble/data-frame overhead.
tests/testthat/test-model_regression.R	Adds regression tests ensuring the new permutation shortcut matches manynet::to_permuted() and that delegation logic behaves as intended.
R/model_predict.R	Updates examples with guidance about using larger times for publication-ready results.
man/predict.Rd	Keeps rendered documentation examples aligned with the updated roxygen examples.
NAMESPACE	Removes now-unused dplyr::bind_cols import.
DESCRIPTION	Bumps package version to 1.6.8.
NEWS.md	Adds 1.6.8 release notes describing the performance improvements and invariance claims.

Files not reviewed (1)

• man/predict.Rd: Generated file

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.