tests/testthat/local-derive_synthetic.R

In synthACS: Synthetic Microdata and Spatial MicroSimulation Modeling for ACS Data

library(testthat)
library(synthACS)

context("local - derive synthetic microdata for geography set")

test_that("get correct results -- serial", {
  #-------------------------------
  #testthat::skip_on_cran()
  testthat::skip_on_covr()
  testthat::skip_on_travis()
  #-------------------------------
  ## parallel == FALSE
  split_ca_dat <- split(ca_dat, n_splits = 20)
  syn <- derive_synth_datasets(split_ca_dat[[1]], parallel= FALSE)
  
  # test class, structure, dimensions, etc
  expect_equal(class(syn), c("synthACS","list"))
  expect_true(is.synthACS(syn))
  expect_true(all(unlist(lapply(syn, is.macro_micro))))
  expect_equal(nrow(split_ca_dat[[1]]$estimates[[1]]), length(syn))
  expect_true(all(unlist(lapply(syn, length)) == 2))
  expect_true(all(unlist(lapply(syn, function(l) ncol(l[[2]]))) == length(split_ca_dat[[1]]$estimates) + 1))
  expect_true(all.equal(lapply(syn, function(l) names(l[[2]])),
                        replicate(length(syn), 
                                  c("age", "gender", "marital_status", "edu_attain", "emp_status", "nativity", 
                                    "pov_status", "geog_mobility", "ind_income", "race", "p"), simplify=FALSE), 
                        check.attributes= FALSE))
  
  # test total probabilities
  expect_true(all.equal(unlist(lapply(syn, function(l) sum(l[[2]]$p))), rep(1,3)
                        , tolerance= 1e-12, check.attributes= FALSE))
  
})

test_that("get correct results -- parallel", {
  #-------------------------------
  #testthat::skip_on_cran()
  testthat::skip_on_covr()
  testthat::skip_on_travis()
  #-------------------------------
  ## parallel == TRUE
  split_ca_dat <- split(ca_dat, n_splits = 20)
  syn <- derive_synth_datasets(split_ca_dat[[1]], parallel= TRUE)
  
  # test class, structure, dimensions, etc
  expect_equal(class(syn), c("synthACS","list"))
  expect_true(is.synthACS(syn))
  expect_true(all(unlist(lapply(syn, is.macro_micro))))
  expect_equal(nrow(split_ca_dat[[1]]$estimates[[1]]), length(syn))
  expect_true(all(unlist(lapply(syn, length)) == 2))
  expect_true(all(unlist(lapply(syn, function(l) ncol(l[[2]]))) == length(split_ca_dat[[1]]$estimates) + 1))
  expect_true(all.equal(lapply(syn, function(l) names(l[[2]])),
                        replicate(length(syn), 
                                  c("age", "gender", "marital_status", "edu_attain", "emp_status", "nativity", 
                                    "pov_status", "geog_mobility", "ind_income", "race", "p"), simplify=FALSE), 
                        check.attributes= FALSE))
  
  # test total probabilities
  expect_true(all.equal(unlist(lapply(syn, function(l) sum(l[[2]]$p))), rep(1,3), 
                        tolerance= 1e-12, check.attributes= FALSE))
  
})