Skip to contents

Contributing to mlmr

Source: CONTRIBUTING.md

Thank you for helping test mlmr.

The project is currently in an early public-beta stage. Feedback is especially useful when it is tied to a concrete model, dataset structure, or reporting workflow.

Useful Feedback

Please open a GitHub issue if you notice:

  • a model that should fit but fails;
  • confusing model-building language;
  • incorrect or unclear equations;
  • APA table formatting problems;
  • uploaded data that the app handles poorly;
  • diagnostics that are missing or hard to interpret;
  • places where the generated R code does not reproduce the app output.

Good Issue Reports

When possible, include:

  • the app tab or workflow step where the problem occurred;
  • the model structure you were trying to fit;
  • whether you used the example data or uploaded data;
  • screenshots of confusing output;
  • the generated R code if the problem involves reproducibility;
  • any warning or error text.

Development Notes

Before submitting changes, run:

R CMD build .
R CMD check --no-manual --no-build-vignettes mlmr_0.1.0.tar.gz

The bundled Shiny app in inst/app/ runs from mirrored copies of R/mlm_core.R, app.R, and www/style.css. After editing any of those files, refresh the copies:

Rscript dev/sync-app.R

Tests live in tests/testthat/ and can be run with devtools::test(). They cover model specification and validation, centering, formula generation, level-by-level equations and Tau matrices, APA tables, reproducible-code round trips, readiness checks, and app/package synchronization. tests/backend-tests.R keeps the original end-to-end smoke test.