tests/testthat/test-dfm_match.R

In koheiw/quanteda.core: Quantitative Analysis of Textual Data

context("test dfm_match")

test_that("test dfm_match", {

    txt <- c(doc1 = "aa bb BB cc DD ee",
             doc2 = "aa bb cc DD ee")
    mt <- dfm(txt, tolower = FALSE)

    mt_conf1 <- dfm_match(mt, c("aa", "zz", "xx", "bb"))
    expect_identical(
        featnames(mt_conf1),
        c("aa", "zz", "xx", "bb")
    )
    expect_identical(
        docnames(mt_conf1),
        c("doc1", "doc2")
    )
    expect_identical(
        colSums(mt_conf1),
        c("aa" = 2, "zz" = 0, "xx" = 0, "bb" = 2)
    )

    mt_conf2 <- dfm_match(mt, featnames(dfm("aa zz xx bb")))
    expect_identical(
        featnames(mt_conf2),
        c("aa", "zz", "xx", "bb")
    )
    expect_identical(
        docnames(mt_conf2),
        c("doc1", "doc2")
    )
    expect_identical(
        colSums(mt_conf2),
        c("aa" = 2, "zz" = 0, "xx" = 0, "bb" = 2)
    )

    expect_error(dfm_match(mt, c(TRUE, FALSE)),
                 "features must be a character vector")
    expect_error(dfm_match(mt, 1:3),
                 "features must be a character vector")
})

test_that("dfm_match works with padding", {
    toks <- tokens("aa bb !", padding = TRUE, remove_punct = TRUE)
    dfmt <- dfm(toks)
    expect_identical(
        featnames(dfm_match(dfmt, c("aa", "bb", "cc", ""))),
        c("aa", "bb", "cc", "")
    )
})