Returns the tibble of conditions symbolize() flagged when it built
the model object. Each row is one warning with a code (machine
identifier), a severity (info / warn / error), a message (the
prose templated from inst/extdata/warning-templates.csv), and a
context string naming which slot values triggered the rule.
Today the system ships two checks:
few_groups_wald(warn) — Wald 95% CI used with a random-effect group of fewer than ~10 levels.small_n_wald(info) — Wald CI on a small dataset. Currently defined in the template CSV but reserved; no extractor fires it yet (the threshold is fit-dependent and not always informative).
New conditions are added by appending a row to
inst/extdata/warning-templates.csv and wiring the trigger in the
relevant drm_build_warnings() (or sibling for other classes).