tests/testthat/test-plot.R

In keyATM: Keyword Assisted Topic Models

# Read Data
data(keyATM_data_bills)
bills_dfm <- keyATM_data_bills$doc_dfm
bills_keywords <- keyATM_data_bills$keywords
bills_cov <- keyATM_data_bills$cov
bills_time_index <- keyATM_data_bills$time_index
keyATM_docs <- keyATM_read(bills_dfm)


#
# With intervals
#

# Base
base <- keyATM(docs = keyATM_docs,
               no_keyword_topics = 3,
               keywords = bills_keywords,
               model = "base",
               options = list(seed = 250, store_theta = TRUE, iterations = 12,
                              store_pi = 1, use_weights = 1, thinning = 1))


test_that("Plot modelfit: the base model", {
  p <- plot_modelfit(base)
  expect_s3_class(p, "keyATM_fig")
  skip_on_cran()
  expect_message(save_fig(p, paste0(tempdir(), "/test.pdf")), "Saving 7 x 7 in image")
  skip_on_os("linux") ; skip_on_cran()
  expect_equal(values_fig(p)$value[2], -9903822, tolerance = 0.00001)
})


test_that("Plot alpha: the base model", {
  p <- plot_alpha(base)
  expect_s3_class(p, "keyATM_fig")
  skip_on_cran()
  expect_message(save_fig(p, paste0(tempdir(), "/test.pdf")), "Saving 7 x 7 in image")
  skip_on_os("linux") ; skip_on_cran()
  expect_equal(values_fig(p)$alpha[2], 3.553541, tolerance = 0.00001)
  expect_identical(values_fig(p)$Topic[3], "3_Health")

  p <- plot_alpha(base, show_topic = c(1, 4))
  expect_identical(values_fig(p)$Topic[3], "1_Education")
})


test_that("Plot pi: the base model", {
  p <- plot_pi(base, ci = 0.95, method = "eti")
  p2 <- plot_pi(base, method = "hdi")
  expect_s3_class(p, "keyATM_fig")

  skip_on_cran()
  expect_message(save_fig(p, paste0(tempdir(), "/test.pdf")), "Saving 7 x 7 in image")

  skip_on_os("linux") ; skip_on_cran()
  expect_equal(as.numeric(values_fig(p)$Upper[2]), 0.03964407, tolerance = 0.00001)
  expect_equal(as.numeric(values_fig(p2)$Lower[2]), 0.03485987, tolerance = 0.00001)
})

test_that("Plot topic plot: the base model", {
  p <- plot_topicprop(base) ; expect_s3_class(p, "keyATM_fig")
  p <- plot_topicprop(base, order = "topicid") ; expect_s3_class(p, "keyATM_fig")
  p <- plot_topicprop(base, n = 5) ; expect_s3_class(p, "keyATM_fig")
  p <- plot_topicprop(base, show_topwords = FALSE) ; expect_s3_class(p, "keyATM_fig")
  p <- plot_topicprop(base, show_topic = 1:3, label_topic = paste0("T", 1:3)) ; expect_s3_class(p, "keyATM_fig")
  p <- plot_topicprop(base, show_topic = c(1, 3, 5), label_topic = paste0("T", c(1, 3, 5)), order = "topicid") ; expect_s3_class(p, "keyATM_fig")
  p <- plot_topicprop(base, show_topic = c(1, 4, 7)) ; expect_s3_class(p, "keyATM_fig")

  expect_error(plot_topicprop(base, show_topic = -1))
  expect_error(plot_topicprop(base, show_topic = 5:100))
  expect_error(plot_topicprop(base, label_topic = 5:100))
})

# Dynamic
dyn <- keyATM(docs = keyATM_docs,
              no_keyword_topics = 3,
              keywords = bills_keywords,
              model = "dynamic",
              model_settings = list(time_index = bills_time_index - 100,
                                    num_states = 5),
              options = list(seed = 250, verbose = FALSE, store_theta = TRUE, iterations = 20, thinning = 1))

test_that("Plot time series: with intervals", {
  p <- plot_timetrend(dyn) ; expect_s3_class(p, "keyATM_fig")
  p <- plot_timetrend(dyn, time_index_label = bills_time_index, method = "eti")
  p2 <- plot_timetrend(dyn, time_index_label = bills_time_index, method = "hdi", ci = 0.89)
  expect_s3_class(p, "keyATM_fig")

  skip_on_cran()
  expect_message(save_fig(p, paste0(tempdir(), "/test.pdf")), "Saving 7 x 7 in image")

  skip_on_os("linux") ; skip_on_cran()
  expect_equal(values_fig(p)$Lower[2], 0.04844414, tolerance = 0.00001)
  expect_equal(values_fig(p2)$Upper[2], 0.1310682, tolerance = 0.00001)

  p <- plot_topicprop(dyn) ; expect_s3_class(p, "keyATM_fig")
})

test_that("Plot alpha: the dynamic model", {
  skip_on_cran()
  p <- plot_alpha(dyn)
  expect_equal(tail(values_fig(p), 1)$alpha[1], 1.341934, tolerance = 0.00001)
  expect_identical(values_fig(p)$Topic[10], "2_Law")

  p <- plot_alpha(dyn, show_topic = c(1, 4))
  expect_identical(values_fig(p)$Topic[3], "1_Education")
})


#
# Without intervals
#

base <- keyATM(docs = keyATM_docs,
               no_keyword_topics = 3,
               keywords = bills_keywords,
               model = "base",
               options = list(seed = 250, store_theta = TRUE, iterations = 12,
                              store_pi = FALSE, use_weights = 1, thinning = 1))

test_that("Plot pi without interval", {
  p <- plot_pi(base)
  expect_s3_class(p, "keyATM_fig")

  skip_on_cran()
  expect_message(save_fig(p, paste0(tempdir(), "/test.pdf")), "Saving 7 x 7 in image")

  skip_on_os("linux") ; skip_on_cran()
  expect_equal(as.numeric(values_fig(p)$Probability[3]), 0.06074006, tolerance = 0.00001)
})


dyn <- keyATM(docs = keyATM_docs,
              no_keyword_topics = 3,
              keywords = bills_keywords,
              model = "dynamic",
              model_settings = list(time_index = bills_time_index - 100,
                                    num_states = 5),
              options = list(seed = 250, verbose = FALSE, store_theta = FALSE, iterations = 20, thinning = 1))

test_that("Time series: without intervals", {
  p <- plot_timetrend(dyn) ; expect_s3_class(p, "keyATM_fig")
  p <- plot_timetrend(dyn, time_index_label = bills_time_index) ; expect_s3_class(p, "keyATM_fig")
  skip_on_cran()
  expect_message(save_fig(p, paste0(tempdir(), "/test.pdf")), "Saving 7 x 7 in image")
  skip_on_os("linux") ; skip_on_cran()
  expect_equal(values_fig(p)$Proportion[2], 0.05016963, tolerance = 0.00001)
})