skip_on_cran()
skip_on_os(c("mac", "solaris"))
skip_if_not_installed("haven")
skip_if_not_installed("datawizard")

test_that("ggpredict, print", {
  # lm, linear regression ----

  data(efc, package = "ggeffects")
  efc$c172code <- datawizard::to_factor(efc$c172code)
  efc$e42dep <- datawizard::to_factor(efc$e42dep)
  efc$c82cop1 <- as.numeric(efc$c82cop1)
  fit <- lm(barthtot ~ c12hour + neg_c_7 + c82cop1 + e42dep + c161sex + c172code, data = efc)

  expect_message({
    junk <- capture.output(print(ggpredict(fit, terms = "c12hour")))
  })
  expect_silent({
    junk <- capture.output(print(ggpredict(fit, terms = "c12hour"), n = Inf))
  })
  ggpredict(fit, terms = "c172code")
  ggpredict(fit, terms = "c161sex")
  ggpredict(fit, terms = c("c12hour", "c172code"))
  ggpredict(fit, terms = c("c12hour", "c161sex"))
  ggpredict(fit, terms = c("e42dep", "c161sex"))
  ggpredict(fit, terms = c("e42dep", "c172code"))
  ggpredict(fit, terms = c("c12hour", "c172code", "c161sex"))
  ggpredict(fit, terms = c("e42dep", "c172code", "c161sex"))
  ggpredict(fit, terms = c("c12hour", "c172code", "e42dep"))
  ggpredict(fit, terms = c("c161sex", "c172code", "e42dep"))
  ggpredict(fit, terms = c("c12hour", "neg_c_7"))
  ggpredict(fit, terms = c("c12hour", "neg_c_7 [all]"))
  ggpredict(fit, terms = c("c12hour", "neg_c_7 [quart2]"))
  ggpredict(fit, terms = c("c12hour", "neg_c_7 [quart2]", "c161sex"))
  ggpredict(fit, terms = c("c12hour", "neg_c_7", "c161sex"))

  expect_snapshot(print(ggpredict(fit, terms = c("c12hour", "neg_c_7", "c161sex"))))
  expect_snapshot(print(ggpredict(fit, terms = c("c12hour", "neg_c_7", "c161sex")), n = Inf))

  out <- utils::capture.output(ggpredict(fit, terms = c("c12hour", "neg_c_7 [quart2]", "c82cop1")))
  expect_equal(
    out,
    c("# Predicted values of Total score BARTHEL INDEX", "", "neg_c_7: 9",
      "c82cop1: 1", "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     95.03 | 87.81, 102.26", "     45 |     91.98 | 84.67,  99.30",
      "     85 |     89.28 | 81.71,  96.84", "    170 |     83.52 | 74.96,  92.08",
      "", "neg_c_7: 9", "c82cop1: 2", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     94.45 | 88.65, 100.24",
      "     45 |     91.40 | 85.52,  97.28", "     85 |     88.69 | 82.53,  94.86",
      "    170 |     82.93 | 75.63,  90.24", "", "neg_c_7: 9", "c82cop1: 3",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     93.86 | 88.88,  98.85", "     45 |     90.82 | 85.77,  95.86",
      "     85 |     88.11 | 82.76,  93.46", "    170 |     82.35 | 75.76,  88.94",
      "", "neg_c_7: 9", "c82cop1: 4", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     93.28 | 88.18,  98.37",
      "     45 |     90.23 | 85.11,  95.35", "     85 |     87.52 | 82.13,  92.91",
      "    170 |     81.77 | 75.20,  88.34", "", "neg_c_7: 11", "c82cop1: 1",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     93.03 | 85.95, 100.11", "     45 |     89.98 | 82.82,  97.14",
      "     85 |     87.27 | 79.87,  94.68", "    170 |     81.52 | 73.11,  89.92",
      "", "neg_c_7: 11", "c82cop1: 2", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     92.44 | 86.73,  98.15",
      "     45 |     89.40 | 83.62,  95.18", "     85 |     86.69 | 80.63,  92.74",
      "    170 |     80.93 | 73.74,  88.13", "", "neg_c_7: 11", "c82cop1: 3",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     91.86 | 86.87,  96.85", "     45 |     88.81 | 83.78,  93.85",
      "     85 |     86.10 | 80.78,  91.43", "    170 |     80.35 | 73.81,  86.89",
      "", "neg_c_7: 11", "c82cop1: 4", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     91.28 | 86.07,  96.48",
      "     45 |     88.23 | 83.02,  93.44", "     85 |     85.52 | 80.06,  90.98",
      "    170 |     79.76 | 73.16,  86.37", "", "neg_c_7: 14", "c82cop1: 1",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     90.02 | 83.03,  97.01", "     45 |     86.98 | 79.92,  94.03",
      "     85 |     84.27 | 76.98,  91.56", "    170 |     78.51 | 70.24,  86.78",
      "", "neg_c_7: 14", "c82cop1: 2", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     89.44 | 83.70,  95.18",
      "     45 |     86.39 | 80.61,  92.18", "     85 |     83.68 | 77.64,  89.73",
      "    170 |     77.93 | 70.77,  85.08", "", "neg_c_7: 14", "c82cop1: 3",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     88.86 | 83.67,  94.04", "     45 |     85.81 | 80.61,  91.01",
      "     85 |     83.10 | 77.64,  88.56", "    170 |     77.34 | 70.72,  83.96",
      "", "neg_c_7: 14", "c82cop1: 4", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     88.27 | 82.74,  93.80",
      "     45 |     85.22 | 79.70,  90.74", "     85 |     82.52 | 76.78,  88.25",
      "    170 |     76.76 | 69.96,  83.56", "", "Adjusted for:", "*   e42dep =            independent",
      "*  c161sex =                   1.76", "* c172code = low level of education"
    )
    ,
    ignore_attr = TRUE
  )

  out <- utils::capture.output(ggpredict(fit, terms = c("c12hour", "neg_c_7", "c82cop1")))
  expect_equal(
    out,
    c("# Predicted values of Total score BARTHEL INDEX", "", "neg_c_7: 8",
      "c82cop1: 1", "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     96.03 | 88.71, 103.35", "     45 |     92.99 | 85.57, 100.40",
      "     85 |     90.28 | 82.61,  97.95", "    170 |     84.52 | 75.86,  93.18",
      "", "neg_c_7: 8", "c82cop1: 2", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     95.45 | 89.58, 101.32",
      "     45 |     92.40 | 86.44,  98.36", "     85 |     89.69 | 83.44,  95.94",
      "    170 |     83.94 | 76.55,  91.33", "", "neg_c_7: 8", "c82cop1: 3",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     94.86 | 89.84,  99.89", "     45 |     91.82 | 86.73,  96.91",
      "     85 |     89.11 | 83.71,  94.50", "    170 |     83.35 | 76.72,  89.99",
      "", "neg_c_7: 8", "c82cop1: 4", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     94.28 | 89.20,  99.36",
      "     45 |     91.23 | 86.12,  96.34", "     85 |     88.52 | 83.14,  93.91",
      "    170 |     82.77 | 76.19,  89.35", "", "neg_c_7: 11.8", "c82cop1: 1",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     92.23 | 85.19,  99.27", "     45 |     89.18 | 82.06,  96.29",
      "     85 |     86.47 | 79.11,  93.83", "    170 |     80.71 | 72.36,  89.07",
      "", "neg_c_7: 11.8", "c82cop1: 2", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     91.64 | 85.94,  97.34",
      "     45 |     88.60 | 82.83,  94.36", "     85 |     85.89 | 79.85,  91.92",
      "    170 |     80.13 | 72.96,  87.30", "", "neg_c_7: 11.8", "c82cop1: 3",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     91.06 | 86.04,  96.08", "     45 |     88.01 | 82.95,  93.07",
      "     85 |     85.30 | 79.96,  90.64", "    170 |     79.55 | 73.00,  86.09",
      "", "neg_c_7: 11.8", "c82cop1: 4", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     90.47 | 85.20,  95.75",
      "     45 |     87.43 | 82.15,  92.70", "     85 |     84.72 | 79.20,  90.24",
      "    170 |     78.96 | 72.32,  85.60", "", "neg_c_7: 15.7", "c82cop1: 1",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     88.32 | 81.31,  95.33", "     45 |     85.27 | 78.21,  92.34",
      "     85 |     82.57 | 75.28,  89.85", "    170 |     76.81 | 68.55,  85.06",
      "", "neg_c_7: 15.7", "c82cop1: 2", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     87.74 | 81.90,  93.58",
      "     45 |     84.69 | 78.81,  90.57", "     85 |     81.98 | 75.86,  88.10",
      "    170 |     76.22 | 69.02,  83.42", "", "neg_c_7: 15.7", "c82cop1: 3",
      "", "c12hour | Predicted |        95% CI", "-----------------------------------",
      "      0 |     87.15 | 81.77,  92.53", "     45 |     84.11 | 78.72,  89.49",
      "     85 |     81.40 | 75.77,  87.02", "    170 |     75.64 | 68.90,  82.38",
      "", "neg_c_7: 15.7", "c82cop1: 4", "", "c12hour | Predicted |        95% CI",
      "-----------------------------------", "      0 |     86.57 | 80.77,  92.36",
      "     45 |     83.52 | 77.75,  89.29", "     85 |     80.81 | 74.84,  86.78",
      "    170 |     75.06 | 68.08,  82.04", "", "Adjusted for:", "*   e42dep =            independent",
      "*  c161sex =                   1.76", "* c172code = low level of education"
    )
    ,
    ignore_attr = TRUE
  )

  out <- ggpredict(fit, terms = c("c161sex", "c172code", "e42dep"))
  expect_snapshot(print(out, group_name = TRUE))
  expect_snapshot(print(out, group_name = FALSE))
})

test_that("ggpredict, print factors", {
  skip_if_not_installed("emmeans")
  LEV <- c(
    "climate", "cutwelfare", "discipline", "freedom", "ineqincOK", "leader",
    "police", "politduty", "refugees", "Russia", "taxesdown", "worse-off"
  )
  n <- 100
  set.seed(1)
  data <- data.frame(
    bin_choice = sample(c(0, 1), size = n, replace = TRUE),
    Wshort = factor(sample(LEV, size = n, replace = TRUE), levels = LEV)
  )
  model.contcons <- glm(bin_choice ~ Wshort, data = data, family = binomial())

  pr <- ggemmeans(model.contcons, "Wshort [all]")
  expect_snapshot(print(pr))

  pr <- ggemmeans(model.contcons, "Wshort")
  expect_snapshot(print(pr))
})


test_that("ggpredict, collapse CI", {
  data(efc, package = "ggeffects")
  efc <- datawizard::to_factor(efc, c("c172code", "c161sex", "e42dep"))
  fit <- lm(barthtot ~ c161sex * c172code * e42dep + c160age, data = efc)
  pr <- suppressWarnings(ggpredict(fit, terms = c("c161sex", "c172code", "e42dep")))

  expect_snapshot(print(pr))
  expect_snapshot(print(pr, group_name = FALSE))
  expect_snapshot(print(pr, group_name = FALSE, collapse_ci = TRUE))
  expect_snapshot(print(pr, group_name = FALSE, collapse_ci = TRUE, ci_brackets = c("[", "]")))
  expect_snapshot(print(pr, group_name = TRUE, collapse_ci = TRUE, ci_brackets = c("[", "]")))

  fit <- lm(barthtot ~ e42dep + c160age, data = efc)
  pr <- ggpredict(fit, terms = "e42dep")
  expect_snapshot(print(pr, group_name = FALSE, collapse_ci = TRUE))
})


test_that("ggpredict, collapse tables", {
  data(iris)
  m <- lm(Sepal.Length ~ Species * Petal.Length, data = iris)
  expect_snapshot(print(ggpredict(m, c("Petal.Length", "Species")), collapse_tables = TRUE, n = 3))
})


test_that("ggpredict, ci-level", {
  data(iris)
  m <- lm(Sepal.Length ~ Species, data = iris)
  out <- ggpredict(m, "Species")
  expect_snapshot(print(out))
  out <- ggpredict(m, "Species", ci_level = 0.8)
  expect_snapshot(print(out))
})


test_that("ggpredict, weights", {
  skip_if_not_installed("MASS")
  data(housing, package = "MASS")
  m <- lm(Freq ~ Infl * Type * Sat, data = housing)
  expect_snapshot(print(ggaverage(m, c("Infl", "Type", "Sat")), collapse_tables = TRUE))
})