nursery_data$recommend <- as.character(nursery_data$recommend)
nursery_data$recommend[which(nursery_data$recommend=="not_recom")] <- "0"
nursery_data$recommend[which(nursery_data$recommend=="recommend")] <- "1"
nursery_data$recommend[which(nursery_data$recommend=="very_recom")] <- "2"
nursery_data$recommend[which(nursery_data$recommend=="priority")] <- "3"
nursery_data$recommend[which(nursery_data$recommend=="spec_prior")] <- "4"
nursery_data$recommend <- as.numeric(nursery_data$recommend)

# extract relevant columns
df = nursery_data[, c("finance", "children")]
# drop rows with missing values
df = df[complete.cases(df), ]
# convert finance to numeric
df = transform(df, finance = as.numeric(finance))
# subtract 1 to make 1s and 2S into 0a and 1s
df$finance <- df$finance -1
# formula to fit
formula = "finance ~ children"

test_that("acro_glm without initialising ACRO object first", {
  ac <<-NULL
  expect_error(acro_glm(formula=formula, data=df, family="probit"), "ACRO has not been initialised. Please first call acro_init()")
})

test_that("acro_glm with probit as a family works", {
  testthat::skip_on_cran()
  acro_init()
  model=acro_glm(formula=formula, data=df, family="probit")
  expect_s3_class(model,"statsmodels.iolib.summary.Summary")
})

test_that("acro_glm with logit as a family works", {
  testthat::skip_on_cran()
  acro_init()
  model=acro_glm(formula=formula, data=df, family="logit")
  expect_s3_class(model,"statsmodels.iolib.summary.Summary")
})

test_that("acro_glm through an error if the family is not recongnised", {
  testthat::skip_on_cran()
  acro_init()
  expect_error(acro_glm(formula=formula, data=df, family="mean"), "Invalid family. Options for family are: logit or probit")
})