tests/testthat/test_D_confound.r

In johnchower/oneD7: Gloo 1D7 analysis

library(dplyr)
glootility::connect_to_redshift()
RPostgreSQL::dbGetQuery(conn = redshift_connection$con
                        , statement = glootility::query_pa_flash_cat)
RPostgreSQL::dbGetQuery(conn = redshift_connection$con
                        , statement = glootility::query_user_flash_cat)

########## Test getConfounders #########
users_test <- 1:100
queryList_test <-  list(oneD7::query_confounder_use_case_sub
                        , oneD7::query_confounder_first_name_sub)

test_that("getConfounders returns results",{
  object_to_test <- getConfounders(users=NULL
                                   , queryList=queryList_test
                                   , runDate = 20170201)
  variable_names <- unique(object_to_test$variable)
  expect_is(object = object_to_test
            , class = 'data.frame')
  expect_gt(object = nrow(object_to_test)
                      , expected = 0)
  expect_equal(object = colnames(object_to_test)
               , expected = c('user_id', 'variable', 'value') )
  expect_equal(object = variable_names[order(variable_names)]
               , expected = c('account_type'
                              , 'first_name') )
})
test_that("getConfounders returns results on a subset",{
  object_to_test <- getConfounders(users=users_test
                                   , queryList=queryList_test
                                   , runDate = 20170201)
  variable_names <- unique(object_to_test$variable)
  expected_variable_names <- c('account_type'
                              , 'first_name')
  expect_is(object = object_to_test
            , class = 'data.frame')
  expect_gt(object = nrow(object_to_test)
                      , expected = 0)
  expect_equal(object = colnames(object_to_test)
               , expected = c('user_id', 'variable', 'value') )
  expect_equal(
    object = variable_names[order(variable_names)]
    , expected =  expected_variable_names[order(expected_variable_names)]
  )
})

queryList_test <-  list(oneD7::query_confounder_use_case_sub
                        , oneD7::query_confounder_oneD7_sub
                        , oneD7::query_confounder_FL_REVEAL_sub
                        , oneD7::query_confounder_belongs_to_cohort_sub
                        , oneD7::query_confounder_first_champ_sub)
test_that("getConfounders returns no nulls or duplicate rows.",{
  allUserConfounders <- oneD7::getConfounders(queryList = queryList_test
                                   , runDate = 20170201)
  has_duplicates <- allUserConfounders %>%
    group_by(user_id, variable) %>%
    summarise(duplicate_rows = length(unique(value)) - 1) 
  has_duplicates <- has_duplicates %>%
    {.$duplicate_rows} %>%
    {max(.) > 0}
  expect_true(object = !has_duplicates)
  allUserConfoundersWide <- tidyr::spread(data = allUserConfounders
                                          , key = 'variable'
                                          , value = 'value')
  naLengths <- lapply(allUserConfoundersWide
                      , FUN = function(x) sum(is.na(x)) )
  expect_equal(naLengths$account_type, 0)
  expect_equal(naLengths$belongs_to_cohort, 0)
  expect_equal(naLengths$connected_to_fl, 0)
  expect_equal(naLengths$connected_to_reveal, 0)
  expect_equal(naLengths$oned7, 0)
  expect_equal(naLengths$first_champ, 0)
})

dbDisconnect(conn = redshift_connection$con)