TODO_spaMM.md

ROADMAP FOR SPAMM SUPPORT

This document describes what should be done so that spaMM becomes supported by broom.mixed.

Basic setup

• fork and clone broom.mixed
• install broom.mixed dependencies
• check that R CMD check is clean

Note: 1 failed test for stanarm (commented) + 15 warnings in tests (untouched)

• add spaMM to DESCRIPTION
• create empty file for tests
• create empty file for tidiers

Implement the basic versions of the main methods

• implement basic version for tidy.HLfit()
• implement basic version for augment.HLfit()
• implement basic version for glance.HLfit()

See https://www.tidymodels.org/learn/develop/broom/ and the broom.mixed vignette for detailed instructions

Improve the main methods

See the dedicated spaMM vignette in this package

Checks / Tests

Make sure it works for:

• all common random effect structures
• all common families and links
• the 3 fitting functions (fitme(), HLfit(), corrHLfit())
• multivariate models?

See the dedicated spaMM vignette in this package

Polish

• cleanup test file
• uncomment added comment in test-stanarm.R
• merge latest broom.mixed
• store fitted models as rda in inst/extdata (produced in run_example.R)
• update documentation (see https://roxygen2.r-lib.org/articles/markdown.html)
• update examples
• remove special spaMM vignette
• update vignette
• update NEWS
• delete this file
• remove mention of this file in .Rbuildignore
• PR