data("sample_explan_data")
sample_explan_data$fakeblock <-  sample(1:3, nrow(sample_explan_data), replace = T)
sample_explan_data$fakeweights <-  sample(1:50, nrow(sample_explan_data), replace = T)
sample_explan_data$fakeabundance <-  sample(1:50, nrow(sample_explan_data), replace = T)
sample_explan_data_test <- dplyr::sample_n(sample_explan_data, 100)
sample_explan_data_train <- dplyr::sample_n(sample_explan_data, 100)


test_that("stops if no occ.data provided", {
  expect_error(
    brt_fit(
      response.col = "presence.absence",
      test.data = sample_explan_data_test,
      distribution = "bernoulli",
      varnames = colnames(sample_explan_data_test)[14:16]
    )
  )
})

test_that("stops if no distribution specified", {
  expect_error(
    brt_fit(
      occ.data = sample_explan_data_train,
      test.data = sample_explan_data_test,
      response.col = "presence.absence",
      varnames = colnames(sample_explan_data_train)[14:16]
    )
  )
})

test_that("stops if no varnames specified", {
  expect_error(
    brt_fit(occ.data = sample_explan_data_train,
      test.data = sample_explan_data_test,
      response.col = "presence.absence",
      distribution = "bernoulli"
    )
  )
})

test_that("stops if varnames not in occ.data", {
  expect_error(
    brt_fit(occ.data = sample_explan_data_train,
      test.data = sample_explan_data_test,
      response.col = "presence.absence",
      distribution = "bernoulli",
      varnames = c("fakevariable")
    )
  )
})

test_that("stops if block.col not in occ.data", {
  expect_error(
    brt_fit(occ.data = sample_explan_data_train,
      test.data = sample_explan_data_test,
      response.col = "presence.absence",
      distribution = "bernoulli",
      block.col = "fakecol",
      varnames = colnames(sample_explan_data_train)[14:16]
    )
  )
})

test_that("stops if response.col not in occ.data", {
  expect_error(
    brt_fit(occ.data = sample_explan_data_train,
      test.data = sample_explan_data_test,
      response.col = "fakecol",
      distribution = "bernoulli",
      varnames = colnames(sample_explan_data_train)[14:16]
    )
  )
})

test_that("If no block col provided, returns list of length 1", {
  results <- brt_fit(occ.data = sample_explan_data_train,
      test.data = sample_explan_data_test,
      response.col = "presence.absence",
      distribution = "bernoulli",
      varnames = colnames(sample_explan_data_train)[14:16]
    )
  expect_equal(length(results), 1)
})

test_that("Works if no block.col specified, and model parameters given", {
  results <- brt_fit(occ.data = sample_explan_data_train,
      response.col = "presence.absence",
      distribution = "bernoulli",
      n.trees = 1500,
      shrinkage = 0.1,
      interaction.depth = 4,
      varnames = colnames(sample_explan_data_train)[14:16]
    )
  expect_equal(length(results), 1)
})


test_that("Uses specified interaction.depth and weights", {
  n <- sample(1:4, 1)
  results <- brt_fit(occ.data = sample_explan_data_train,
      test.data = sample_explan_data_test,
      weights.col = "fakeweights",
      response.col = "presence.absence",
      distribution = "bernoulli",
      n.trees = 1500,
      shrinkage = 0.01,
      interaction.depth = n,
      varnames = colnames(sample_explan_data_train)[14:16]
    )
  expect_equal(results[[1]]$interaction.depth, n)
})

test_that("Works if block.col specified, and model parameters given", {
  results <- brt_fit(occ.data = sample_explan_data_train,
      response.col = "presence.absence",
      block.col = "fakeblock",
      distribution = "bernoulli",
      n.trees = 1500,
      shrinkage = 0.1,
      interaction.depth = 4,
      varnames = colnames(sample_explan_data_train)[14:16]
    )
  expect_equal(length(results), length(unique(sample_explan_data_train[, "fakeblock"])))
})

test_that("Works if block.col specified, and model parameters not given", {
  results <- brt_fit(occ.data = sample_explan_data_train,
      response.col = "presence.absence",
      block.col = "fakeblock",
      distribution = "bernoulli",
      varnames = colnames(sample_explan_data_train)[14:16]
    )
  expect_equal(length(results), length(unique(sample_explan_data_train[, "fakeblock"])))
})

## Test works with abundance data, not just binary

test_that(
  "Works if no block.col specified, and model parameters given for different family and data type",
  {
    results <- brt_fit(occ.data = sample_explan_data_train,
        test.data = sample_explan_data_test,
        response.col = "fakeabundance",
        distribution = "gaussian",
        varnames = colnames(sample_explan_data_train)[14:16]
      )
    expect_equal(length(results), 1)
  }
)