inst/templates/tests/testthat/test-missing_data.R

In gorkang/jsPsychHelpeR: Standardize and automatize data preparation and analysis of jsPsych experiments created with jsPsychMaker

testthat::test_that('Tests if participants that completed the protocol are missing test results', {

  # DEBUG
  # targets::tar_load(c(DF_raw, DF_joined))
  
  
  # Name of test (should reflect the name of the file) ----------------------
  
  name_of_test = "missing_data"
  # cat(cli::style_underline(cli::col_yellow(paste0("\n\nRunning: ", cli::col_silver(name_of_test, paste(rep(" ", 40), collapse = " ")),"\n\n"))))
  
  white_list = c("AIM_01", "AIM_02", "AIM_04", "AIM_05", "AIM_06", "AIM_07", "AIM_08", "AIM_09", "AIM_10", "AIM_TramoIngreso_DIRd", "AIM_DIRt", 
                 "DEMOGR22_06",
                 "Report_001_1", "Report_001_2", "Report_alias_DIRd", "Report_SoloContacto_DIRd", "Report_EmailFuturo_DIRd",
                 "DMW_Vagabundeo_DIRd", "DMW_TipoVagabundeo_DIRd")
  
  # Test: missing data in standardized vars --------------------------------------------------------------
  
  # Selects all STDt, STDd, DIRt and DIRd scales
  # all_scales also used in "R/prepare_df_analysis.R"
  all_scales = grep(".*_DIRt$|.*_STDt$|.*_DIRd$|.*STDd$", names(DF_joined), value = TRUE, perl = TRUE)
  
  # Use max n of tasks as criteria for completed protocol
  ids_completed_protocol = 
    DF_raw |> 
    dplyr::count(id, experiment) |> 
    dplyr::count(id, name = "tasks") |>  
    dplyr::filter(tasks == max(tasks))
  
  missing_DF_temp = 
    DF_joined |> 
    dplyr::filter(id %in% ids_completed_protocol$id) |> # Filter participants that did not complete the protocol
    dplyr::select(id, dplyr::all_of(all_scales)) |> 
    dplyr::select(-matches(white_list)) %>%
    # dplyr::select(-matches(white_list)) |> 
    dplyr::mutate(NAs_id = rowSums(is.na(select(., matches(all_scales)))))
  
  # Get names
  DF_missing_vars = 
    missing_DF_temp |> 
    dplyr::mutate(dplyr::across(-id, is.na)) |>  # replace all NA with TRUE and else FALSE
    tidyr::pivot_longer(-id, names_to = "var") |>  # pivot longer
    dplyr::filter(value == TRUE) |> 
    dplyr::group_by(id) |>    # group by the ID
    dplyr::summarise(Missing_Variables = toString(var), .groups = "keep")
  
  missing_DF = 
    missing_DF_temp |> 
    dplyr::left_join(DF_missing_vars, by = "id") %>% 
    dplyr::select(id, NAs_id, Missing_Variables, colnames(.)[!complete.cases(t(.))]) |> 
    dplyr::filter(NAs_id > 0)
  
  missing_vars = 
    missing_DF |> dplyr::select(-id, -NAs_id, -Missing_Variables) |> names()
  
  missing_ids = missing_DF_temp |> 
    dplyr::filter(NAs_id > 0) |> 
    dplyr::arrange(id) |> 
    dplyr::pull(id)
  
  
  

  # Test: missings in any item ----------------------------------------------

  missingall_DF_temp = 
    DF_joined |> 
    dplyr::select(-dplyr::all_of(all_scales)) %>%
    dplyr::mutate(NAs_id = rowSums(is.na(select(., -matches(all_scales)))))
  
  # Get names
  DF_missingall_vars = 
    missingall_DF_temp |> 
    dplyr::mutate(dplyr::across(-id, is.na)) |>  # replace all NA with TRUE and else FALSE
    tidyr::pivot_longer(-id, names_to = "var") |>  # pivot longer
    dplyr::filter(value == TRUE) |> 
    dplyr::group_by(id) |>    # group by the ID
    dplyr::summarise(Missing_Variables = toString(var), .groups = "keep")
  
  missingall_DF = missingall_DF_temp |> 
    dplyr::left_join(DF_missingall_vars, by = "id") %>%
    dplyr::select(id, NAs_id, Missing_Variables, colnames(.)[!complete.cases(t(.))]) |> 
    dplyr::filter(NAs_id > 0)
  
  # Warning and log ---------------------------------------------------------
  
  if (length(missing_ids) > 0) {
    
    data.table::fwrite(missing_DF, here::here(paste0("outputs/tests_outputs/test-", name_of_test, ".csv")))
    data.table::fwrite(missingall_DF, here::here(paste0("outputs/tests_outputs/test-", name_of_test, "_ALL.csv")))
    
    cat(cli::col_red("\n", cli::style_underline("ERROR"), " in ", paste0("test-", name_of_test), "\n"),
        cli::col_red("  - Some of the participants are ", cli::style_underline("missing test results:")), missing_ids, "\n",
        cli::col_red("  - In the following vars:"), missing_vars, "\n",
        
        cli::col_green("  - # of Issues: "), cli::col_red(length(missing_ids)), "\n",
        cli::col_silver("  - DF with details stored in: ", paste0("'outputs/tests_outputs/test-", name_of_test, ".csv'"), "\n\n"))
    
  }
  
  # Actual expectations -------------------------------------------------------------
  
  testthat::expect_gt(ncol(missing_DF), 2) # Checks that we have some columns in the DF
  testthat::expect_length(missing_ids, 0) # Check that there are no id's with missing data

})