tests/testthat/test-data.R

In sbashevkin/LTMRdata: Data for the IEP Long Term Monitoring Survey Review

require(dplyr)
require(lubridate)
data<-bind_rows(LTMRdata::Baystudy, LTMRdata::Suisun, LTMRdata::FMWT, LTMRdata::DJFMP, LTMRdata::EDSM, LTMRdata::TMM, LTMRdata::SLS, LTMRdata::STN, LTMRdata::SKT, LTMRdata::Salvage)

test_that("All Lats are between 37 and 39 and all Longs are between -123 and -121", {
  expect_true(all((data$Latitude<39.3 & data$Latitude>37) | is.na(data$Latitude)))
  expect_true(all((data$Longitude<(-121) & data$Longitude>(-123)) | is.na(data$Longitude)))
})

test_that("Sample dates are formatted correctly", {
  expect_true(all(class(data$Date) %in% c("POSIXct","POSIXt")))
  expect_true(all(grepl("[0-9]{4}-[0-9]{2}-[0-9]{2}",as.character(data$Date))))
})

test_that("Sample times are formatted correctly", {
  expect_true(all(class(data$Datetime) %in% c("POSIXct","POSIXt")))
  # datetime_format <- "[0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}"
  # expect_true(all(grepl(datetime_format,as.character(data$Datetime)) |
  #                   is.na(data$Datetime)))
})

test_that("Tide has the expected value", {
  expect_true(all(data$Tide %in% c("Low Slack","Ebb","High Slack","Flood") |
                    is.na(data$Tide)))
})

test_that("Method has the expected value", {
  expect_setequal(unique(data$Method),
                    c('Midwater trawl', 'Otter trawl', 'Kodiak trawl',
                      'Beach seine', '20mm net', 'Oblique tow', 'STN net',
                      'Normal count', 'Second flush'))
})

test_that("Length_NA_flag has correct unique values", {
  expect_setequal(unique(data$Length_NA_flag), c(NA_character_, "Unknown length", "No fish caught"))
})

test_that("Length_NA_flag 'No fish caught' is applied correctly", {
  expect_setequal(filter(data, is.na(Length) & Count==0)$Length_NA_flag, "No fish caught") # 'No fish caught' should only be applied to cases where Length and Count are both NA
  expect_equal(nrow(filter(data, !(is.na(Length) & Count==0) & Length_NA_flag=="No fish caught")), 0) # 'No fish caught' should only be applied to cases where Length and Count are both NA
  expect_equal(nrow(filter(data, !is.na(Taxa) & Length_NA_flag=="No fish caught")), 0) # Taxa should be NA for all 'No fish caught'
  expect_equal(nrow(filter(data, is.na(Taxa) & Length_NA_flag!="No fish caught")), 0) # NA taxa rows should only exist when 'No fish caught'
})

test_that("Length_NA_flag 'Unknown length' is applied correctly", {
  expect_setequal(filter(data, is.na(Length) & Count>0)$Length_NA_flag, "Unknown length") # 'Unknown length' should only be applied when length is NA and Count>0
  expect_equal(nrow(filter(data, !(is.na(Length) & Count>0) & Length_NA_flag=="Unknown length")), 0) # 'Unknown length' should only be applied when length is NA and Count>0
})

# This is mostly checking for new species added to datasets
test_that("Combinations of Taxa, Count, and Length_NA_flag are as expected", {
  len_flag_values <- c("No fish caught","Unknown length", "Missing catch value")
  expect_true(all(data$Length_NA_flag %in% len_flag_values | is.na(data$Length_NA_flag)))

  ## If Length_NA_flag is NA, Count and Taxa should be present:
  sub_1 <- subset(data, is.na(Length_NA_flag))
  expect_true(sum(is.na(sub_1$Count)) == 0)
  expect_true(sum(is.na(sub_1$Taxa)) == 0)

  ## If Count is greater than 0, Taxa should be present:
  sub_2 <- subset(data, Count>0)
  expect_true(sum(is.na(sub_2$Taxa)) == 0)

  ## If Count is present, check Length_NA_flag:
  sub_3 <- subset(data, !is.na(Count))
  expect_true(all( (sub_3$Length_NA_flag %in% len_flag_values[1:2]) |
                     is.na(sub_3$Length_NA_flag) ))

  ## If Length_NA_flag is not missing, it should have one of the three values:
  sub_4 <- subset(data, !is.na(Length_NA_flag))
  expect_true(all(sub_4$Length_NA_flag %in% len_flag_values))
})

test_that("No zero counts exist in the dataset, except for instances of 'No fish caught'", {
  expect_equal(filter(data, Count==0), filter(data, Length_NA_flag=="No fish caught"))
})

test_that("No NA counts exist in the datast", {
  expect_equal(nrow(filter(data, is.na(Count))), 0)
})

test_that("No Tow volumes or Tow areas are negative", {
  expect_equal(nrow(filter(data, Tow_volume<0)), 0)
  expect_equal(nrow(filter(data, Tow_area<0)), 0)
})

test_that("No Tow volumes or Tow areas are 0", {
  expect_equal(nrow(filter(data, Tow_volume==0 & Source!="Salvage")), 0)
  expect_equal(nrow(filter(data, Tow_area==0)), 0)
})

test_that("No sample times are midnight, ignoring Salvage", {
  expect_equal(nrow(filter(data, Source!="Salvage" & hour(Datetime)==0 & minute(Datetime)==0)), 0)
})

test_that("No salinities are negative", {
  expect_equal(nrow(filter(data, Sal_surf<0)), 0)
  expect_equal(nrow(filter(data, Sal_bot<0)), 0)
})

test_that("No salinities are too high", {
  expect_equal(nrow(filter(data, Sal_surf>40)), 0)
  expect_equal(nrow(filter(data, Sal_bot>40)), 0)
})

test_that("No temperatures are negative", {
  expect_equal(nrow(filter(data, Temp_surf<0)), 0)
})

test_that("No temperatures are too high", {
  expect_equal(nrow(filter(data, Temp_surf>40)), 0)
})

rm(data)
gc()