tests/testthat/test-text_sub.R

In corpus: Text Corpus Analysis

context("text_sub")


test_that("'text_sub' errors for invalid 'start' arg", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!"),
                 filter = text_filter(drop_punct = TRUE))
              
    expect_error(text_sub(x, c(1, 2)),
                 "'start' length does not evenly divide argument length")

    expect_error(text_sub(x, c(1, 2, 3, 4)),
                 "'start' length does not evenly divide argument length")

    expect_error(text_sub(x, integer()),
                 "'start' length does not evenly divide argument length")

    expect_error(text_sub(x, "1"),
                 "'start' must be an integer vector or two-column matrix")

    expect_error(text_sub(x, matrix(1:3, 1, 3)),
                 "'start' must be an integer vector or two-column matrix")
})


test_that("'text_sub' errors for invalid 'end' arg", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!"),
                 filter = text_filter(drop_punct = TRUE))
              
    expect_error(text_sub(x, 1, c(1, 2)),
                 "'end' length does not evenly divide argument length")

    expect_error(text_sub(x, 1, c(1, 2, 3, 4)),
                 "'end' length does not evenly divide argument length")

    expect_error(text_sub(x, 1, integer()),
                 "'end' length does not evenly divide argument length")

    expect_error(text_sub(x, 1, "1"),
                 "'end' must be an integer vector")

    expect_error(text_sub(x, end = matrix(1:2, 1, 2)),
                 "'end' must be an integer vector")

    expect_warning(text_sub(x, start = matrix(1, 1, 2), end = 1),
        "'end' argument is ignored when 'start' is a two-column matrix")
})


test_that("'text_sub' can return whole text", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))
    expect_equal(text_sub(x), x)
    expect_equal(text_sub(x, 1, text_ntoken(x)), x)
    expect_equal(text_sub(x, 0, text_ntoken(x) + 1), x)
})


test_that("'text_sub' can get first token", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))
    y <- as_corpus_text(c("A ", "A \"", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))

    expect_equal(text_sub(x, 1, 1), y)
})


test_that("'text_sub' can get second token", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))
    y <- as_corpus_text(c("man, ", "canal\"?", "", "", NA),
                 filter = text_filter(drop_punct = TRUE))

    expect_equal(text_sub(x, 2, 2), y)
})


test_that("'text_sub' can get third token", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))
    y <- as_corpus_text(c("a ", "", "", "", NA),
                 filter = text_filter(drop_punct = TRUE))

    expect_equal(text_sub(x, 3, 3), y)
})


test_that("'text_sub' can get last token", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))
    y <- as_corpus_text(c("plan.", "canal\"?", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))

    expect_equal(text_sub(x, -1, -1), y)
})


test_that("'text_sub' can get second-to-last token", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))
    y <- as_corpus_text(c("a ", "A \"", "", "", NA),
                 filter = text_filter(drop_punct = TRUE))

    expect_equal(text_sub(x, -2, -2), y)
})


test_that("'text_sub' can get third-to-last token", {
    x <- as_corpus_text(c("A man, a plan.", "A \"canal\"?", "Panama!", "", NA),
                 filter = text_filter(drop_punct = TRUE))
    y <- as_corpus_text(c("man, ", "", "", "", NA),
                 filter = text_filter(drop_punct = TRUE))

    expect_equal(text_sub(x, -3, -3), y)
})


test_that("'text_sub' can be applied twice", {
    x <- paste(letters, collapse = " ")

    y <- text_sub(x, 5, 8)
    expect_equal(y, as_corpus_text(c("e f g h ")))

    z <- text_sub(y, 2, 3)
    expect_equal(z, as_corpus_text(c("f g ")))
})