test_that("check_convergence runs correctly for valid approaches", {
  data("PimaIndiansDiabetes2", package = "mlbench")

  pima_data <- PimaIndiansDiabetes2 |>
    dplyr::mutate(
      diabetes = ifelse(diabetes == "pos", 1, 0),
      bmi = dplyr::case_when(
        mass < 25 ~ "Normal",
        mass >= 25 & mass < 30 ~ "Overweight",
        mass >= 30 ~ "Obese"
      ),
      bmi = factor(bmi, levels = c("Normal", "Overweight", "Obese")),
      age_cat = dplyr::case_when(
        age < 30 ~ "Young",
        age >= 30 & age < 50 ~ "Middle-aged",
        age >= 50 ~ "Older"
      ),
      age_cat = factor(age_cat, levels = c("Young", "Middle-aged", "Older"))
    )

  exposures <- c("bmi", "age_cat")
  outcome <- "diabetes"
  valid_approaches <- c("logit", "log-binomial", "robpoisson")

  for (approach in valid_approaches) {
    result_uni <- gtregression::check_convergence(pima_data,
                                                  exposures,
                                                  outcome,
                                                  approach = approach)
    expect_s3_class(result_uni, "data.frame")
    expect_true(all(c("Exposure", "Model", "Converged", "Max.prob.")
                    %in% names(result_uni)))
    expect_gt(nrow(result_uni), 0)
    expect_true(any(result_uni$Converged))

    result_multi <- gtregression::check_convergence(pima_data,
                                                    exposures,
                                                    outcome,
                                                    approach = approach,
                                                    multivariate = TRUE)
    expect_s3_class(result_multi, "data.frame")
    expect_equal(nrow(result_multi), 1)
    expect_true(result_multi$Converged)
  }

  data("quine", package = "MASS")

  quine_data <- quine |>
    dplyr::mutate(Days = as.numeric(Days))

  count_exposures <- c("Eth", "Sex", "Age", "Lrn")
  count_outcome <- "Days"

  for (approach in c("poisson", "negbin")) {
    result_uni <- gtregression::check_convergence(quine_data,
                                                  count_exposures,
                                                  count_outcome,
                                                  approach = approach)
    expect_s3_class(result_uni, "data.frame")
    expect_gt(nrow(result_uni), 0)
    expect_true(any(result_uni$Converged))

    result_multi <- gtregression::check_convergence(quine_data,
                                                    count_exposures,
                                                    count_outcome,
                                                    approach = approach,
                                                    multivariate = TRUE)
    expect_s3_class(result_multi, "data.frame")
    expect_equal(nrow(result_multi), 1)
    expect_true(result_multi$Converged)
  }

  # Errors
  expect_error(
    gtregression::check_convergence(data= quine_data,
                                    exposures= count_exposures,
                                    outcome= count_outcome,
                                    approach = "logit"),
    regexp = paste0(
      "This approach requires either a factor variable ",
      "or numeric variable coded as 0 and 1 \\(or 1 and 2\\)"
    )
  )

  # Empty data
  empty_data <- pima_data[0, ]
  expect_error(
    gtregression::check_convergence(empty_data,
                                    exposures,
                                    outcome),
    regexp = paste0("ll values in the outcome variable are missing")

)
}
)

test_that("check_convergence handles model fitting failure", {
  pima_data <- PimaIndiansDiabetes2 |>
    dplyr::mutate(
      diabetes = ifelse(diabetes == "pos", 1, 0),
      bmi = dplyr::case_when(
        mass < 25 ~ "Normal",
        mass >= 25 & mass < 30 ~ "Overweight",
        mass >= 30 ~ "Obese"
      ),
      bmi = factor(bmi, levels = c("Normal", "Overweight", "Obese")),
      age_cat = dplyr::case_when(
        age < 30 ~ "Young",
        age >= 30 & age < 50 ~ "Middle-aged",
        age >= 50 ~ "Older"
      ),
      age_cat = factor(age_cat, levels = c("Young", "Middle-aged", "Older"))
    )
  broken_data <- pima_data |>
    dplyr::mutate(all_one = factor("yes", levels = "yes"))

  result <- check_convergence(broken_data, exposures = "all_one",
                              outcome = "diabetes", approach = "logit")
  expect_false(result$Converged)
  expect_true(is.na(result$Max.prob.))
})