tests/testthat/test_data_reduction.R

In blongworth/amsdata: Tools for working with accelerator mass spectrometry data

context("Data Reduction")

data <- readRDS("CFAMS011421.rds")

context("Large Blank Correction")

test_that("LBC returns correct LBC Fm", {
        expect_equal(doLBC(1, 0, 1), 1)
        expect_equal(doLBC(0, 0, 1), 0)
        # ceylon should be no correction, but snics_results shows different cor val
        expect_true(dplyr::near(doLBC(data[1,3], data[1,8], 1.0398), data[1,3]))
        # C-8 OC true value is different at 0.000001 level
        expect_true(dplyr::near(doLBC(data[5,3], data[5,8], 1.0398), data[5,6]))
        # Small C-2 (lgblk from database)
        expect_true(dplyr::near(doLBC(data[7,3], data[7,8], 1.0398), data[7,6]))

})

test_that("LBC works with vectors", {
        expect_equal(doLBC(c(1, 0), c(0, 0), c(1,1)), c(1, 0))
        expect_equal(doLBC(data[,3], data[,8], 1.0398), data[,6])
})


test_that("LBC returns correct LBC Fm error", {
  expect_equal(doLBCerr(fmmeas = 1, fmblank = 0, fmstd = 1, fmmeaserr = 0, fmblankerr = 0), 0) # sqrt 0
    # ceylon should be no correction, but snics_results shows different cor val
    expect_true(dplyr::near(doLBCerr(data[1,3], data[1,8], 1.0398, data[1,18], data[1,9]), data[1,7]))
    # C-8 OC true value is different at 0.000001 level
    expect_true(dplyr::near(doLBCerr(data[5,3], data[5,8], 1.0398, data[5,18], data[5,9]), data[5,7]))
    # Small C-2 (lgblk from database)
    expect_true(dplyr::near(doLBCerr(data[7,3], data[7,8], 1.0398, data[7,18], data[7,9]), data[7,7]))
})

test_that("LBCerr works with vectors", {
  expect_equal(doLBCerr(data[,3], data[,8], 1.0398, data[,18], data[,9]), data[,7])
})

context("Mass Balance Correction")

# LBC leads to negative fm's, so testing inputs is tricky
# How should we handle these cases, data-wise?
test_that("doMBC checks for valid input", {
  expect_error(doMBC("blue", 0, 0, 1))
  #expect_error(doMBC(-1, 0, 0, 1))
  #expect_error(doMBC(1, 0, 0, 0))
  #expect_error(doMBC(0, 0, 1, 1))
})

test_that("doMBC returns NA when values are missing", {
  expect_equal(is.na(doMBC(data[1,6], data[1,12], data[1,14], data[1,16])), TRUE)

})

test_that("MBC returns correct MBC Fm", {
  expect_equal(doMBC(fmmeas = 1, fmblank = 1, massmeas = 1, massblank = 0), 1)
  expect_equal(doMBC(1, 0, 2, 1), 2)
  #expect_equal(doMBC(1, 0, 1, 1), 2) # test for handling of massmeas = massblank
  #expect_equal(doMBC(1, 0, 1, 2), 0) # test for massblank > massmeas
  expect_true(dplyr::near(doMBC(data[5,6], data[5,12], data[5,14], data[5,16]), data[5,10]))
})

test_that("MBC works with vectors", {
  expect_equal(doMBC(c(1,1), c(1,0), c(1,2), c(0,1)), c(1,2))
  expect_equal(doMBC(data[,6], data[,12], data[,14], data[,16]), data[,10])
})

test_that("MBC returns correct MBC Fm Err", {
  expect_equal(doMBCerr(1, 0, 1, 0, 0, 0, 0, 0), 0) # sqrt 0
  expect_equal(doMBCerr(data[5,6], data[5,12], data[5,14], data[5,16],
                     data[5,7], data[5,13], data[5,15], data[5,17]),
               data[5,11])
  expect_equal(doMBCerr(data[6,6], data[6,12], data[6,14], data[6,16],
                     data[6,7], data[6,13], data[6,15], data[6,17]),
               data[6,11])
})

test_that("MBCErr works with vectors", {
  expect_equal(doMBCerr(data[,6], data[,12], data[,14], data[,16],
                     data[,7], data[,13], data[,15], data[,17]),
               data[,11])
})

test_that("blankErr returns correct blank", {
  expect_equal(blankErr(1), NA) # can't calc error on one meas without error
  expect_equal(blankErr(1,1), 1)
  expect_equal(blankErr(4, 1), 2)
  expect_equal(blankErr(c(2,2)), 1)
})

test_that("d13C correction returns expected values", {
  expect_equal(calc_d13c(.0112372), 0) # can't calc error on one meas without error
  expect_equal(calc_d13c(.01102369), -19, tolerance = 0.0001) # OX-I
})