Too many warnings in examples

[pawelru]

I am checking our packages against warnings in examples and this package has too many of them. While warnings in general are not necessarily wrong it gives some bad experience when occured in examples. These should be small, simple and clean(!).

I do understand that sometimes this is intended / unavoidable but have a look at this and amend the example code to avoid it if possible.

At the moment of creating this issue a following list of files returns a warning but please check all again when doing this task.

• [x] analyze_variables.Rd - https://github.com/insightsengineering/formatters/pull/250
• [x] estimate_proportions.Rd - false positive from \insertCite
• [x] extract_rsp_biomarkers.Rd
• [x] fit_coxreg.Rd - no warning found (with most recent package versions installed)
• [x] fit_logistic.Rd - useful warning about model convergence
• [x] fit_rsp_step.Rd
• [x] h_ancova.Rd - no warning found (with most recent package versions installed)
• [x] h_logistic_regression.Rd - useful warning about model convergence
• [x] h_odds_ratio.Rd - no warning found (with most recent package versions installed)
• [x] h_prop_diff.Rd - false positive from \insertCite
• [x] h_prop_diff_test.Rd - false positive from \insertCite
• [x] h_proportions.Rd - false positive from \insertCite
• [x] h_response_subgroups.Rd - no warning found (with most recent package versions installed)
• [x] h_stack_by_baskets.Rd
• [x] odds_ratio.Rd - no warning found (with most recent package versions installed)
• [x] response_subgroups.Rd - false positive from \insertCite
• [x] strata_normal_quantile.Rd - false positive from \insertCite
• [x] summarize_logistic.Rd - useful warning about model convergence
• [x] tidy.glm.Rd - useful warning about model convergence

Code that I use to identify those:

withr::with_options(
  list(warn = 2),
  devtools::run_examples(run_donttest = TRUE, run_dontrun = FALSE)
)