# Basic tests for gmm_survey estimation

test_that("gmm_survey runs with minimal inputs", {

  sim_data <- simulate_gmm_survey(
    n_individuals = 100,
    n_times = 4,
    n_classes = 2,
    seed = 111
  )

  fit <- gmm_survey(
    data = sim_data,
    id = "id",
    time = "time",
    outcome = "outcome",
    n_classes = 2,
    starts = 10,
    verbose = FALSE
  )

  expect_s4_class(fit, "SurveyMixr")
  expect_equal(fit@model_info$n_classes, 2)
  expect_true(fit@convergence_info$converged)
})

test_that("gmm_survey output has expected structure", {

  sim_data <- simulate_gmm_survey(
    n_individuals = 100,
    n_times = 4,
    n_classes = 2,
    seed = 222
  )

  fit <- gmm_survey(
    data = sim_data,
    id = "id",
    time = "time",
    outcome = "outcome",
    n_classes = 2,
    starts = 10,
    verbose = FALSE
  )

  # Check slots
  expect_true(length(fit@parameters$growth_parameters) == 2)
  expect_true(length(fit@parameters$class_proportions) == 2)
  expect_equal(ncol(fit@posterior_probs), 2)
  expect_equal(nrow(fit@posterior_probs), 100)

  # Check fit indices
  expect_true(!is.na(fit@fit_indices$loglik))
  expect_true(!is.na(fit@fit_indices$bic))
  expect_true(!is.na(fit@fit_indices$entropy))
  expect_true(fit@fit_indices$entropy >= 0 && fit@fit_indices$entropy <= 1)
})

test_that("gmm_survey handles survey design", {

  sim_data <- simulate_gmm_survey(
    n_individuals = 200,
    n_times = 4,
    n_classes = 2,
    design = "stratified_cluster",
    seed = 333
  )

  fit <- gmm_survey(
    data = sim_data,
    id = "id",
    time = "time",
    outcome = "outcome",
    n_classes = 2,
    strata = "stratum",
    cluster = "psu",
    weights = "weight",
    starts = 10,
    verbose = FALSE
  )

  expect_true(fit@survey_design$has_weights)
  expect_true(fit@survey_design$has_cluster)
  expect_true(fit@survey_design$has_strata)
})

test_that("S4 methods work correctly", {

  sim_data <- simulate_gmm_survey(
    n_individuals = 100,
    n_times = 4,
    n_classes = 2,
    seed = 444
  )

  fit <- gmm_survey(
    data = sim_data,
    id = "id",
    time = "time",
    outcome = "outcome",
    n_classes = 2,
    starts = 10,
    verbose = FALSE
  )

  # Test methods don't error
  expect_error(print(fit), NA)
  expect_error(summary(fit), NA)
  expect_error(coef(fit), NA)
  expect_error(logLik(fit), NA)
  expect_error(AIC(fit), NA)
  expect_error(BIC(fit), NA)

  # Test coef returns named vector
  coeffs <- coef(fit)
  expect_type(coeffs, "double")
  expect_true(length(coeffs) > 0)
  expect_true(!is.null(names(coeffs)))
})