# Local data pre-processing
test_fit <- local({
  dta <- tern_ex_adtte[tern_ex_adtte$PARAMCD == "OS", ]
  survival::survfit(formula = Surv(AVAL, 1 - CNSR) ~ ARMCD, data = dta)
})

testthat::test_that("control_surv_med_annot works with default settings", {
  result <- control_surv_med_annot()

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("control_coxph_annot works with default settings", {
  result <- control_coxph_annot()

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("h_xticks works with default settings", {
  result <- h_data_plot(test_fit) %>%
    h_xticks()

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("h_xticks works with xticks number", {
  result <- h_data_plot(test_fit) %>%
    h_xticks(xticks = 100)

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("h_xticks works with xticks numeric", {
  expected <- c(0, 365, 1000)
  result <- h_data_plot(test_fit) %>%
    h_xticks(xticks = expected)

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("h_xticks returns error when xticks non-numeric", {
  testthat::expect_error(
    h_data_plot(test_fit) %>% h_xticks(xticks = TRUE)
  )
})

testthat::test_that("h_xticks works with max_time only", {
  result <- h_data_plot(test_fit) %>%
    filter(time <= 3000) %>%
    h_xticks(max_time = 3000)

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("h_xticks works with xticks numeric when max_time is not NULL", {
  expected <- c(0, 365, 1000)
  result <- h_data_plot(test_fit) %>%
    filter(time <= 1500) %>%
    h_xticks(xticks = expected, max_time = 1500)

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("h_xticks works with xticks number when max_time is not NULL", {
  result <- h_data_plot(test_fit) %>%
    filter(time <= 1500) %>%
    h_xticks(xticks = 500, max_time = 1500)

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("h_tbl_median_surv estimates median survival time with CI", {
  result <- h_tbl_median_surv(fit_km = test_fit)

  res <- testthat::expect_silent(result)
  testthat::expect_snapshot(res)
})

testthat::test_that("h_tbl_coxph_pairwise estimates HR, CI and pvalue", {
  df <- tern_ex_adtte %>%
    filter(PARAMCD == "OS") %>%
    mutate(is_event = CNSR == 0)
  variables <- list(tte = "AVAL", is_event = "is_event", arm = "ARMCD")
  result1 <- h_tbl_coxph_pairwise(
    df = df,
    variables = variables,
    control_coxph_pw = control_coxph()
  )

  res <- testthat::expect_silent(result1)
  testthat::expect_snapshot(res)

  result2 <- h_tbl_coxph_pairwise(
    df = df,
    variables = c(variables, list(strata = "SEX")),
    control_coxph_pw = control_coxph(pval_method = "wald", ties = "exact", conf_level = 0.99)
  )

  res <- testthat::expect_silent(result2)
  testthat::expect_snapshot(res)
})


# h_data_plot ----
testthat::test_that("h_data_plot works as expected", {
  data <- test_fit
  testthat::expect_silent(result <- h_data_plot(data))
  res <- names(result)

  testthat::expect_snapshot(res)
  testthat::expect_s3_class(result, "tbl_df")
})

testthat::test_that("h_data_plot respects the ordering of the arm variable factor levels", {
  data <- tern_ex_adtte %>%
    filter(PARAMCD == "OS") %>%
    mutate(ARMCD = factor(ARMCD, levels = c("ARM B", "ARM C", "ARM A"))) %>%
    survival::survfit(formula = Surv(AVAL, 1 - CNSR) ~ ARMCD, data = .)

  testthat::expect_silent(result <- h_data_plot(data))
  res <- levels(result$strata)

  testthat::expect_snapshot(res)
  testthat::expect_s3_class(result, "tbl_df")
  testthat::expect_s3_class(result$strata, "factor")
})

testthat::test_that("h_data_plot adds rows that have time 0 and estimate 1", {
  data <- test_fit
  testthat::expect_silent(result <- h_data_plot(data))
  result_corner <- result %>%
    filter(time == 0, estimate == 1)
  res <- result_corner$strata

  testthat::expect_snapshot(res)
  testthat::expect_true(with(
    result_corner,
    all(conf.high == 1) && all(conf.low == 1) && all(n.event == 0) && all(n.censor == 0)
  ))
})

# Deprecated Functions ----

fit_km <- tern_ex_adtte %>%
  filter(PARAMCD == "OS") %>%
  survival::survfit(formula = Surv(AVAL, 1 - CNSR) ~ ARMCD, data = .)
data_plot <- h_data_plot(fit_km = fit_km)
xticks <- h_xticks(data = data_plot)

testthat::test_that("h_ggkm, h_decompose_gg, h_grob_y_annot, and h_km_layout return deprecation warnings", {
  lifecycle::expect_deprecated(gg <- h_ggkm(
    data = data_plot,
    yval = "Survival",
    censor_show = TRUE,
    xticks = xticks, xlab = "Days", ylab = "Survival Probability",
    title = "tt",
    footnotes = "ff"
  ))

  lifecycle::expect_deprecated(withr::with_options(
    opts_partial_match_old,
    g_el <- h_decompose_gg(gg)
  ))

  lifecycle::expect_deprecated(h_grob_y_annot(ylab = g_el$ylab, yaxis = g_el$yaxis))
  lifecycle::expect_deprecated(h_km_layout(data = data_plot, g_el = g_el, title = "t", footnotes = "f"))
})

testthat::test_that("h_grob_median_surv return deprecation warning", {
  lifecycle::expect_deprecated(grob_surv <- fit_km %>% h_grob_median_surv())
})

testthat::test_that("h_grob_tbl_at_risk return deprecation warning", {
  annot_tbl <- summary(fit_km, times = xticks)

  strata_lst <- strsplit(sub("=", "equals", levels(annot_tbl$strata)), "equals")
  levels(annot_tbl$strata) <- matrix(unlist(strata_lst), ncol = 2, byrow = TRUE)[, 2]
  annot_tbl <- data.frame(
    n.risk = annot_tbl$n.risk,
    time = annot_tbl$time,
    strata = annot_tbl$strata
  )

  lifecycle::expect_deprecated(tbl <- h_grob_tbl_at_risk(data = data_plot, annot_tbl = annot_tbl, xlim = max(xticks)))
})

testthat::test_that("h_grob_coxph returns error when only one arm", {
  df <- tern_ex_adtte %>%
    filter(PARAMCD == "OS") %>%
    mutate(is_event = CNSR == 0)
  variables <- list(tte = "AVAL", is_event = "is_event", arm = "ARMCD")

  lifecycle::expect_deprecated(h_grob_coxph(df = df, variables = variables))
})