Define as phrase for dictionary2

quanteda · Mar 26, 2022 · 1533fca · 1533fca
1 parent 1b28792
commit 1533fca
Show file tree

Hide file tree

Showing 5 changed files with 27 additions and 16 deletions.
diff --git a/NAMESPACE b/NAMESPACE
@@ -59,6 +59,7 @@ S3method(as.list,tokens)
 S3method(as.matrix,dfm)
 S3method(as.phrase,collocations)
 S3method(as.phrase,default)
+S3method(as.phrase,dictionary2)
 S3method(as.phrase,list)
 S3method(as.phrase,tokens)
 S3method(as.tokens,default)

diff --git a/R/phrases.R b/R/phrases.R
@@ -77,6 +77,12 @@ phrase.collocations <- function(x, separator = " ") {
     as.phrase(x)
 }
 
+#' @noRd
+#' @export
+as.phrase.dictionary2 <- function(x) {
+  phrase(unlist(x, use.names = TRUE), separator = " ")
+}
+
 #' @noRd
 #' @export
 as.phrase.list <- function(x) {

diff --git a/tests/testthat/test-collocations.R b/tests/testthat/test-collocations.R
@@ -165,27 +165,26 @@ test_that("kwic works as expected with and without collocations phrases", {
                            stringsAsFactors = FALSE)
     class(coll_tri) <- c("collocations", "data.frame")
 
-    suppressWarnings({
+
     expect_equal(
-        as.data.frame(kwic(txt, coll_bi))$keyword,
+        as.data.frame(kwic(toks_uni, coll_bi))$keyword,
         c("a b", "e g", "g h",
           "a b", "e g", "g h")
     )
     expect_equal(
-        as.data.frame(kwic(txt, coll_tri))$keyword,
+        as.data.frame(kwic(toks_uni, coll_tri))$keyword,
         c("e g h", "e g h")
     )
 
     expect_equal(
-        as.data.frame(kwic(txt, as.phrase(coll_bi)))$keyword,
+        as.data.frame(kwic(toks_uni, as.phrase(coll_bi)))$keyword,
         c("a b", "e g", "g h",
           "a b", "e g", "g h")
     )
     expect_equal(
-        as.data.frame(kwic(txt, as.phrase(dict_bi)))$keyword,
+        as.data.frame(kwic(toks_uni, as.phrase(dict_bi)))$keyword,
         c("a b", "a b")
     )
-    })
 
     expect_equal(nrow(kwic(toks_uni, coll_bi)), 6)
     expect_equal(nrow(kwic(toks_uni, coll_tri)), 2)

diff --git a/tests/testthat/test-default-methods.R b/tests/testthat/test-default-methods.R
@@ -131,12 +131,12 @@ test_that("kwic default works", {
 test_that("phrase default works", {
     expect_error(
         phrase(TRUE),
-        "phrase\\(\\) only works on character.*dictionary2"
+        "phrase\\(\\) only works on character, collocations, dictionary2, list, tokens objects"
     )
 
     expect_error(
         as.phrase(TRUE),
-        "phrase\\(\\) only works on collocations, list, tokens objects"
+        "phrase\\(\\) only works on collocations, dictionary2, list, tokens objects"
     )
 })
 

diff --git a/tests/testthat/test-phrases.R b/tests/testthat/test-phrases.R
@@ -22,19 +22,17 @@ test_that("test phrase for character", {
 })
 
 test_that("test phrase for dictionaries", {
-    dict <- dictionary(list(country = c("United States"),
+    dict1 <- dictionary(list(country = c("United States"),
                             institution = c("Congress", "feder* gov*")))
     expect_equivalent(
-        phrase(dict),
+        phrase(dict1),
         list(c("united", "states"), c("congress"), c("feder*", "gov*"))
     )
-})
-
-test_that("test phrase for dictionaries", {
-    dict <- dictionary(list(country = c("United States"),
-                            institution = c("Congress", "feder* gov*")))
+    dict2 <- dictionary(list(country = c("United+States"),
+                            institution = c("Congress", "feder*+gov*")),
+                        separator = "+")
     expect_equivalent(
-        phrase(dict),
+        phrase(dict2),
         list(c("united", "states"), c("congress"), c("feder*", "gov*"))
     )
 })
@@ -63,6 +61,13 @@ test_that("test as.phrase", {
         suppressWarnings(phrase(col)),
         list(c("a", "b"), c("b", "c"), c("c", "d"), c("a", "b", "c"))
     )
+    dict <- dictionary(list(country = c("United+States"),
+                            institution = c("Congress", "feder*+gov*")),
+                        separator = "+")
+    expect_equivalent(
+      phrase(dict),
+      list(c("united", "states"), c("congress"), c("feder*", "gov*"))
+    )
 })
 
 test_that("helper functions for phrase() work", {