tests/testthat/test_prolific_scoring_stub.R

In umx: Structural Equation Modeling and Twin Modeling in R

library(testthat)
library(umx)

test_that("prolific_scoring_stub works correctly", {
	# 1. Create a dummy dataframe resembling a Qualtrics export
	dummyDf = data.frame(
		StartDate = c("2026-07-16", "2026-07-16"),
		IPAddress = c("1.1.1.1", "2.2.2.2"),
		QID123 = c("meta", "meta"),
		Age = c("25", "30"),
		Sex = c("Male", "Female"),
		AntiNatal_1 = c("Strongly disagree", "Somewhat agree"),
		AntiNatal_2 = c("Somewhat disagree", "Strongly agree"),
		AntiNatal_3 = c("Neither agree nor disagree", "Strongly agree"),
		SingleItem_1 = c("Option A", "Option B"),
		stringsAsFactors = FALSE
	)
	
	# Capture the printed messages using testthat's capture_messages
	output = capture_messages({
		returnedDf = prolific_scoring_stub(dummyDf, deleteJunk = TRUE)
	})
	
	# Verify that the age/sex conversion code is printed
	expect_true(any(grepl("dummyDf\\$Age = as.numeric\\(dummyDf\\$Age\\)", output)))
	expect_true(any(grepl("dummyDf\\$Sex = as.factor\\(dummyDf\\$Sex\\)", output)))
	
	# Verify scale options printed and sorted
	expect_true(any(grepl("AntiNatalSTR =", output)))
	# Verify the items are sorted negative -> neutral -> positive, specifically "Somewhat disagree" before "Neither agree nor disagree"
	expect_true(any(grepl("AntiNatalSTR = c\\(\"Strongly disagree\", \"Somewhat disagree\", \"Neither agree nor disagree\", \"Somewhat agree\", \"Strongly agree\"\\)", output)))
	
	# Verify the umx_score_scale call output
	expect_true(any(grepl("umx_score_scale\\(name = \"AntiNatal\"", output)))
	
	# Verify that SingleItem is not detected (since it is size 1)
	expect_false(any(grepl("SingleItemSTR", output)))
	
	# Verify deleteJunk removes StartDate, IPAddress, QID123
	expect_false("StartDate" %in% names(returnedDf))
	expect_false("IPAddress" %in% names(returnedDf))
	expect_false("QID123" %in% names(returnedDf))
	expect_true("AntiNatal_1" %in% names(returnedDf))
	
	# Test with deleteJunk = FALSE
	returnedDf2 = prolific_scoring_stub(dummyDf, deleteJunk = FALSE)
	expect_true("StartDate" %in% names(returnedDf2))
})

test_that("prolific_scoring_stub handles scales with digits in the base name", {
	# Create a dummy dataframe with Q1_2 through Q1_36
	colNames = c("StartDate", "IPAddress", paste0("Q1_", 2:36))
	dummyDf = as.data.frame(matrix("Strongly disagree", nrow = 2, ncol = length(colNames)))
	names(dummyDf) = colNames
	
	output = capture_messages({
		returnedDf = prolific_scoring_stub(dummyDf, deleteJunk = TRUE)
	})
	
	# Verify that it grouped them into a single scale "Q1" with items 2:36
	expect_true(any(grepl("Q1STR =", output)))
	# The score call should contain base = "Q1_" and pos = 2:36
	expect_true(any(grepl("umx_score_scale\\(name = \"Q1\", base = \"Q1_\", pos = 2:36", output)))
})