# unit testing for computing area under the curve

# single dyad for testing
one_rhyme <- NurseryRhymes_Prepped[NurseryRhymes_Prepped$Event_ID == "LittleLamb",]
one_rhyme <- droplevels(one_rhyme)


test_that("auc formats properly", {
  # computed previously
  expected_auc_df <- data.frame(
    Event_ID = "LittleLamb", Exchanges = 50, Talked_First = as.factor("Mary"),
    AUC_emo_anger_raw_Immediate = 1.485925, AUC_emo_anger_scaled50_Immediate = 1.485925,
    AUC_emo_anger_raw_Lag1 = 1.4556, AUC_emo_anger_scaled50_Lag1 = 1.4556
  )
  expected_auc_df$Talked_First <- factor(expected_auc_df$Talked_First, levels = c("Dave", "Mary"))
  # compare to computed df
  expect_equal(ConversationAlign:::compute_auc(one_rhyme), expected_auc_df)
})

# remove all but first three exchanges
one_rhyme_small <- one_rhyme[one_rhyme$Exchange_Count < 3, ]

test_that("auc fills with na when under 3 exchanges", {
  expected_small_auc_df <- data.frame(
    Event_ID = "LittleLamb", Exchanges = 2, Talked_First = as.factor("Mary"),
    AUC_emo_anger_raw_Immediate = as.double(NA), AUC_emo_anger_scaled50_Immediate = as.double(NA),
    AUC_emo_anger_raw_Lag1 = as.double(NA), AUC_emo_anger_scaled50_Lag1 = as.double(NA)
  )
  expected_small_auc_df$Talked_First <- factor(expected_small_auc_df$Talked_First, levels = c("Dave", "Mary"))
  # compare to computed df
  expect_equal(ConversationAlign:::compute_auc(one_rhyme_small), expected_small_auc_df)

})