tests/testthat/test-model_regression.R

In migraph: Inferential Methods for Multimodal and Other Networks

set.seed(123)
networkers <- manynet::ison_networkers %>% manynet::to_subgraph(Discipline == "Sociology")
netsenders <- manynet::to_unweighted(networkers)

test <- net_regression(weight ~ ego(Citations),
                     networkers, times = 50)
test_logit <- net_regression(. ~ ego(Citations),
                          netsenders, times = 50)

test_that("network_reg estimates correctly",{
  expect_s3_class(test, "netlm")
  expect_equal(top3(test$coefficients,3),
               c(26.811, -0.379, NA))
  expect_s3_class(test_logit, "netlogit")
  expect_equal(top3(test_logit$coefficients,3),
               c(0.146, -0.024, NA))
  expect_output(print(test), "Model summary statistics")
  expect_output(print(test_logit), "Model summary statistics")
})

test_that("network_reg tests correctly",{
  expect_equal(top3(test_logit$pgreqabs,2),
               c(0.8, 0.18, NA), tolerance = 0.1)
})

test_that("tidy works correctly for network_reg",{
  tidys <- tidy(test)
  expect_s3_class(tidys, "tbl_df")
  expect_equal(top3(tidys$estimate, 2), c(26.81, -0.38, NA))
  tidys <- tidy(test_logit)
  expect_output(print(tidys), "term")
})

test_that("glance works correctly for network_reg",{
  glances <- glance(test)
  expect_s3_class(glances, "tbl_df")
  expect_equal(top3(glances$r.squared, 3), c(0.021, NA, NA))
  glances <- glance(test_logit)
  expect_output(print(glances), "AIC")
})

test_that("multivariate QAP works",{
  expect_s3_class(net_regression(weight ~ ego(Citations) + alter(Citations) + sim(Citations), 
                                networkers, times = 10),
                 "netlm")
})

# test_that("specification advice appears",{
#   expect_message(net_regression(weight ~ same(Discipline), networkers, times = 1),
#                  "When testing for homophily")
# })

# plots <- plot(test)
# test_that("plot works correctly for network_reg",{
#   expect_s3_class(plots, "gg")
#   expect_equal(plots$labels$x, "Statistic")
# })