test_that("positive test tidy_pool_obj", {
  data("ADMI")
  set.seed(122)
  N_IMPUTATIONS <- 100
  BURN_IN <- 200
  BURN_BETWEEN <- 5

  # Convert key columns to factors
  ADMI$TRT <- factor(ADMI$TRT, levels = c("Placebo", "Drug A"))
  ADMI$USUBJID <- factor(ADMI$USUBJID)
  ADMI$AVISIT <- factor(ADMI$AVISIT)

  # Define key variables for ANCOVA analysis
  vars <- rbmi::set_vars(
    subjid = "USUBJID",
    visit = "AVISIT",
    group = "TRT",
    outcome = "CHG",
    covariates = c("BASE", "STRATA", "REGION") # Covariates for adjustment
  )

  # Specify the imputation method (Bayesian) - need for pool step
  method <- rbmi::method_bayes(
    n_samples = N_IMPUTATIONS,
    control = rbmi::control_bayes(
      warmup = BURN_IN,
      thin = BURN_BETWEEN
    )
  )

  # Perform ANCOVA Analysis on Each Imputed Dataset
  ana_obj_ancova <- analyse_mi_data(
    data = ADMI,
    vars = vars,
    method = method,
    fun = rbmi::ancova, # Apply ANCOVA
    delta = NULL # No sensitivity analysis adjustment
  )

  pool_obj_ancova <- rbmi::pool(ana_obj_ancova)

  # Just check the class of pool_obj_ancova before passing to tidy_pool_obj
  expect_s3_class(pool_obj_ancova, "pool")

  tidy_df <- tidy_pool_obj(pool_obj_ancova)

  # Test output
  # class check
  expect_s3_class(tidy_df, c("tbl_df", "tbl", "data.frame"))

  # column names check
  expect_named(
    tidy_df,
    c(
      "parameter",
      "description",
      "visit",
      "parameter_type",
      "lsm_type",
      "est",
      "se",
      "lci",
      "uci",
      "pval"
    )
  )

  # column types check
  expect_identical(
    lapply(tidy_df, class),
    list(
      parameter = "character",
      description = "character",
      visit = "character",
      parameter_type = "character",
      lsm_type = "character",
      est = "numeric",
      se = "numeric",
      lci = "numeric",
      uci = "numeric",
      pval = "numeric"
    )
  )

  # Columns created by tidy_pool_obj are:
  # - parameter_type
  # - lsm_type
  # - visit
  # - description
  expect_identical(
    tidy_df[, c('description', 'visit', 'parameter_type', 'lsm_type')],
    dplyr::as_tibble(
      data.frame(
        description = c(
          "Treatment Comparison",
          "Least Squares Mean for Reference at Week 24",
          "Least Squares Mean for Alternative at Week 24",
          "Treatment Comparison",
          "Least Squares Mean for Reference at Week 48",
          "Least Squares Mean for Alternative at Week 48"
        ),
        visit = c(
          "Week 24",
          "Week 24",
          "Week 24",
          "Week 48",
          "Week 48",
          "Week 48"
        ),
        parameter_type = c("trt", "lsm", "lsm", "trt", "lsm", "lsm"),
        lsm_type = c(NA, "ref", "alt", NA, "ref", "alt")
      )
    )
  )
})

test_that("tidy_pool_obj requires a pool object", {
  expect_true(class(mtcars) != "pool")

  expect_error(
    tidy_pool_obj(mtcars),
    class = "rbmiUtils_error_validation"
  )
})


# =============================================================================
# Additional comprehensive tests for tidy_pool_obj
# =============================================================================

test_that("tidy_pool_obj handles single visit data", {
  data("ADMI")
  set.seed(234)

  # Filter to single visit
  ADMI_single <- ADMI[ADMI$AVISIT == "Week 24", ]
  ADMI_single$TRT <- factor(ADMI_single$TRT, levels = c("Placebo", "Drug A"))
  ADMI_single$USUBJID <- factor(ADMI_single$USUBJID)
  ADMI_single$AVISIT <- factor(ADMI_single$AVISIT)

  vars <- rbmi::set_vars(
    subjid = "USUBJID",
    visit = "AVISIT",
    group = "TRT",
    outcome = "CHG",
    covariates = c("BASE", "STRATA", "REGION")
  )

  method <- rbmi::method_bayes(
    n_samples = 50,
    control = rbmi::control_bayes(warmup = 100, thin = 2)
  )

  ana_obj <- analyse_mi_data(
    data = ADMI_single,
    vars = vars,
    method = method,
    fun = rbmi::ancova
  )

  pool_obj <- rbmi::pool(ana_obj)
  tidy_df <- tidy_pool_obj(pool_obj)

  expect_s3_class(tidy_df, "tbl_df")
  expect_true(all(tidy_df$visit == "Week 24" | is.na(tidy_df$visit)))
})


test_that("tidy_pool_obj preserves numeric precision", {
  data("ADMI")
  set.seed(345)

  ADMI$TRT <- factor(ADMI$TRT, levels = c("Placebo", "Drug A"))
  ADMI$USUBJID <- factor(ADMI$USUBJID)
  ADMI$AVISIT <- factor(ADMI$AVISIT)

  vars <- rbmi::set_vars(
    subjid = "USUBJID",
    visit = "AVISIT",
    group = "TRT",
    outcome = "CHG",
    covariates = c("BASE", "STRATA", "REGION")
  )

  method <- rbmi::method_bayes(
    n_samples = 50,
    control = rbmi::control_bayes(warmup = 100, thin = 2)
  )

  ana_obj <- analyse_mi_data(
    data = ADMI,
    vars = vars,
    method = method,
    fun = rbmi::ancova
  )

  pool_obj <- rbmi::pool(ana_obj)
  tidy_df <- tidy_pool_obj(pool_obj)

  # Check numeric columns are numeric
  expect_type(tidy_df$est, "double")
  expect_type(tidy_df$se, "double")
  expect_type(tidy_df$lci, "double")
  expect_type(tidy_df$uci, "double")
  expect_type(tidy_df$pval, "double")
})


test_that("tidy_pool_obj correctly identifies lsm_ref parameters", {
  data("ADMI")
  set.seed(456)

  ADMI$TRT <- factor(ADMI$TRT, levels = c("Placebo", "Drug A"))
  ADMI$USUBJID <- factor(ADMI$USUBJID)
  ADMI$AVISIT <- factor(ADMI$AVISIT)

  vars <- rbmi::set_vars(
    subjid = "USUBJID",
    visit = "AVISIT",
    group = "TRT",
    outcome = "CHG",
    covariates = c("BASE", "STRATA", "REGION")
  )

  method <- rbmi::method_bayes(
    n_samples = 50,
    control = rbmi::control_bayes(warmup = 100, thin = 2)
  )

  ana_obj <- analyse_mi_data(
    data = ADMI,
    vars = vars,
    method = method,
    fun = rbmi::ancova
  )

  pool_obj <- rbmi::pool(ana_obj)
  tidy_df <- tidy_pool_obj(pool_obj)

  # Check that ref and alt are correctly identified
  ref_rows <- tidy_df[tidy_df$lsm_type == "ref" & !is.na(tidy_df$lsm_type), ]
  alt_rows <- tidy_df[tidy_df$lsm_type == "alt" & !is.na(tidy_df$lsm_type), ]

  expect_true(nrow(ref_rows) > 0)
  expect_true(nrow(alt_rows) > 0)
  expect_true(all(grepl("Reference", ref_rows$description)))
  expect_true(all(grepl("Alternative", alt_rows$description)))
})


test_that("tidy_pool_obj correctly identifies treatment comparison parameters", {
  data("ADMI")
  set.seed(567)

  ADMI$TRT <- factor(ADMI$TRT, levels = c("Placebo", "Drug A"))
  ADMI$USUBJID <- factor(ADMI$USUBJID)
  ADMI$AVISIT <- factor(ADMI$AVISIT)

  vars <- rbmi::set_vars(
    subjid = "USUBJID",
    visit = "AVISIT",
    group = "TRT",
    outcome = "CHG",
    covariates = c("BASE", "STRATA", "REGION")
  )

  method <- rbmi::method_bayes(
    n_samples = 50,
    control = rbmi::control_bayes(warmup = 100, thin = 2)
  )

  ana_obj <- analyse_mi_data(
    data = ADMI,
    vars = vars,
    method = method,
    fun = rbmi::ancova
  )

  pool_obj <- rbmi::pool(ana_obj)
  tidy_df <- tidy_pool_obj(pool_obj)

  # Check treatment comparison rows
  trt_rows <- tidy_df[tidy_df$parameter_type == "trt", ]

  expect_true(nrow(trt_rows) > 0)
  expect_true(all(grepl("Treatment Comparison", trt_rows$description)))
  expect_true(all(is.na(trt_rows$lsm_type)))
})


test_that("tidy_pool_obj output can be filtered by visit", {
  data("ADMI")
  set.seed(678)

  ADMI$TRT <- factor(ADMI$TRT, levels = c("Placebo", "Drug A"))
  ADMI$USUBJID <- factor(ADMI$USUBJID)
  ADMI$AVISIT <- factor(ADMI$AVISIT)

  vars <- rbmi::set_vars(
    subjid = "USUBJID",
    visit = "AVISIT",
    group = "TRT",
    outcome = "CHG",
    covariates = c("BASE", "STRATA", "REGION")
  )

  method <- rbmi::method_bayes(
    n_samples = 50,
    control = rbmi::control_bayes(warmup = 100, thin = 2)
  )

  ana_obj <- analyse_mi_data(
    data = ADMI,
    vars = vars,
    method = method,
    fun = rbmi::ancova
  )

  pool_obj <- rbmi::pool(ana_obj)
  tidy_df <- tidy_pool_obj(pool_obj)

  # Filter by visit should work
  week24_df <- tidy_df[tidy_df$visit == "Week 24", ]
  week48_df <- tidy_df[tidy_df$visit == "Week 48", ]

  expect_true(nrow(week24_df) > 0)
  expect_true(nrow(week48_df) > 0)
})


test_that("tidy_pool_obj errors with NULL input", {
  expect_error(tidy_pool_obj(NULL), class = "rbmiUtils_error_validation")
})


test_that("tidy_pool_obj errors with list input", {
  expect_error(tidy_pool_obj(list(a = 1, b = 2)), class = "rbmiUtils_error_validation")
})


# =============================================================================
# Tests for underscore-containing parameter names (HRD-01)
# =============================================================================

# Helper to build a proper mock pool object matching rbmi internal structure
make_mock_pool <- function(parameter_names, est, se, lci, uci, pval) {
  pars <- lapply(seq_along(parameter_names), function(i) {
    list(
      est = est[i],
      se = se[i],
      ci = c(lci[i], uci[i]),
      pvalue = pval[i]
    )
  })
  names(pars) <- parameter_names
  obj <- list(pars = pars)
  class(obj) <- "pool"
  obj
}


test_that("tidy_pool_obj handles parameter names with underscores", {
  mock_pool <- make_mock_pool(
    parameter_names = c(
      "trt_Week_24", "lsm_ref_Week_24", "lsm_alt_Week_24",
      "trt_Follow_Up", "lsm_ref_Follow_Up", "lsm_alt_Follow_Up"
    ),
    est = c(-2.5, 10.1, 7.6, -1.8, 9.5, 7.7),
    se = c(0.8, 0.5, 0.6, 0.9, 0.5, 0.6),
    lci = c(-4.1, 9.1, 6.4, -3.6, 8.5, 6.5),
    uci = c(-0.9, 11.1, 8.8, -0.0, 10.5, 8.9),
    pval = c(0.002, NA, NA, 0.05, NA, NA)
  )

  result <- tidy_pool_obj(mock_pool)

  # Verify underscore-containing visits are preserved intact
  expect_true("Week_24" %in% result$visit)
  expect_true("Follow_Up" %in% result$visit)

  # Verify parameter_type is correctly identified
  expect_equal(sum(result$parameter_type == "trt"), 2)
  expect_equal(sum(result$parameter_type == "lsm"), 4)

  # Verify lsm_type is correctly identified
  lsm_rows <- result[result$parameter_type == "lsm", ]
  expect_true(all(lsm_rows$lsm_type %in% c("ref", "alt")))

  # Verify trt rows have NA lsm_type
  trt_rows <- result[result$parameter_type == "trt", ]
  expect_true(all(is.na(trt_rows$lsm_type)))
})


test_that("tidy_pool_obj handles gcomp-style parameter names", {
  mock_pool <- make_mock_pool(
    parameter_names = c(
      "lsm_Placebo_Week 24", "lsm_Drug A_Week 24",
      "lsm_Placebo_Week 48", "lsm_Drug A_Week 48"
    ),
    est = c(0.30, 0.45, 0.35, 0.55),
    se = c(0.03, 0.04, 0.03, 0.04),
    lci = c(0.24, 0.37, 0.29, 0.47),
    uci = c(0.36, 0.53, 0.41, 0.63),
    pval = c(NA, NA, NA, NA)
  )

  result <- tidy_pool_obj(mock_pool)

  # Verify lsm_type captures arm names
  expect_true("Placebo" %in% result$lsm_type)
  expect_true("Drug A" %in% result$lsm_type)

  # Verify visit is correctly extracted
  expect_true("Week 24" %in% result$visit)
  expect_true("Week 48" %in% result$visit)
})


test_that("tidy_pool_obj handles gcomp-style trt parameter names", {
  mock_pool <- make_mock_pool(
    parameter_names = c(
      "trt_Drug A-Placebo_Week 24",
      "lsm_Placebo_Week 24", "lsm_Drug A_Week 24",
      "trt_Drug A-Placebo_Week 48",
      "lsm_Placebo_Week 48", "lsm_Drug A_Week 48"
    ),
    est = c(-0.10, 0.30, 0.20, -0.15, 0.35, 0.20),
    se = c(0.03, 0.03, 0.04, 0.03, 0.03, 0.04),
    lci = c(-0.16, 0.24, 0.12, -0.21, 0.29, 0.12),
    uci = c(-0.04, 0.36, 0.28, -0.09, 0.41, 0.28),
    pval = c(0.001, NA, NA, 0.001, NA, NA)
  )

  result <- tidy_pool_obj(mock_pool)

  # trt rows should have visit extracted (not the full comparison string)
  trt_rows <- result[result$parameter_type == "trt", ]
  expect_equal(trt_rows$visit, c("Week 24", "Week 48"))

  # lsm rows should have arm names as lsm_type
  expect_true("Placebo" %in% result$lsm_type)
  expect_true("Drug A" %in% result$lsm_type)
})


test_that("tidy_pool_obj errors with informative message for non-pool input", {
  expect_error(
    tidy_pool_obj(data.frame(x = 1)),
    class = "rbmiUtils_error_validation"
  )

  # Also verify the error inherits from the base class
  expect_error(
    tidy_pool_obj(data.frame(x = 1)),
    class = "rbmiUtils_error"
  )
})