skip_if_not_installed("metafor")
skip_if_not_installed("metadat")

d <- data.frame(
  estimate = c(0.111, 0.245, 0.8, 1.1, 0.03),
  std.error = c(0.05, 0.111, 0.001, 0.2, 0.01)
)
mydat <<- d
model <- metafor::rma(yi = estimate, sei = std.error, data = mydat)

test_that("model_info", {
  expect_true(model_info(model)$is_linear)
  expect_true(model_info(model)$is_meta)
})

test_that("find_formula", {
  expect_equal(
    find_formula(model),
    list(conditional = estimate ~ 1),
    ignore_attr = TRUE
  )
})

out <- get_data(model)
test_that("get_data", {
  expect_equal(
    out$estimate,
    c(0.111, 0.245, 0.8, 1.1, 0.03),
    tolerance = 1e-3
  )
  expect_identical(dim(out), as.integer(c(5, 3)))
  expect_identical(colnames(out), c("estimate", "std.error", "Weights"))
})

out <- get_data(model, source = "mf")
test_that("get_data, modelframe", {
  expect_identical(dim(out), as.integer(c(5, 3)))
  expect_identical(colnames(out), c("estimate", "std.error", "Weights"))
})

data(dat.bcg, package = "metadat")
dat <- metafor::escalc(
  measure = "RR", ai = tpos, bi = tneg, ci = cpos,
  di = cneg, data = dat.bcg
)
dat$alloc <- ifelse(dat$alloc == "random", "random", "other")
model <- metafor::rma(yi, vi,
  mods = ~alloc, data = dat, digits = 3,
  slab = author
)
test_that("get_data, modelframe", {
  expect_equal(
    find_formula(model),
    list(conditional = yi ~ alloc, dispersion = yi ~ alloc),
    ignore_attr = TRUE
  )
})