Add convert ref #15

NAMESPACE

@@ -1,6 +1,8 @@
 # Generated by roxygen2: do not edit by hand
 
+S3method(convert,tokens_with_proximity)
 S3method(dfm,tokens_with_proximity)
 S3method(print,tokens_with_proximity)
 export(tokens_proximity)
+importFrom(quanteda,convert)
 importFrom(quanteda,dfm)

R/get_dist.R

@@ -78,6 +78,12 @@ tokens_proximity <- function(x, keywords, get_min = TRUE, valuetype = c("glob",
     return(toks)
 }
 
+.convert_df <- function(tokens_obj, proximity_obj, doc_id) {
+    data.frame("doc_id" = rep(doc_id, length(tokens_obj)),
+               "token" = tokens_obj,
+               "proximity" = proximity_obj)
+}
+
 #' @method print tokens_with_proximity
 #' @export
 print.tokens_with_proximity <- function(x, ...) {
@@ -88,6 +94,18 @@ print.tokens_with_proximity <- function(x, ...) {
     cat("keywords: ", quanteda::meta(x, field = "keywords"), "\n")
 }
 
+#' @method convert tokens_with_proximity
+#' @export
+#' @importFrom quanteda convert
+convert.tokens_with_proximity <- function(x, to = c("data.frame")) {
+    to <- match.arg(to)
+    purrr::list_rbind(
+               purrr::pmap(list(tokens_obj = as.list(x),
+                                       proximity_obj = quanteda::docvars(x, "proximity"),
+                                       doc_id = quanteda::docnames(x)),
+                           .convert_df)
+           )
+}
 
 #' Create a document-feature matrix
 #'

tests/testthat/test-tokens_dist.R

@@ -2,6 +2,10 @@ test_that("defensive", {
     expect_error(tokens_proximity("a", "a"), "x is not a")
 })
 
+test_that("edge cases", {
+    expect_error("" %>% tokens() %>% tokens_proximity("") %>% convert(), NA)
+})
+
 test_that(".resolve_keywords", {
     expect_equal(.resolve_keywords(c("abc", "def"), c("abcd", "defa"), valuetype = "fixed"), c("abc", "def"))
     expect_equal(.resolve_keywords(c("abc*", "def*"), c("abcd", "defa"), valuetype = "glob"), c("abcd", "defa"))
@@ -18,3 +22,9 @@ test_that("count_from", {
     "this is my life" %>% tokens() %>% tokens_proximity("my", count_from = -1) %>% docvars("proximity") -> res
     expect_equal(res$text1, c(1, 0, -1, 0))
 })
+
+test_that("convert", {
+    suppressPackageStartupMessages(library(quanteda))
+    "this is my life" %>% tokens() %>% tokens_proximity("my") %>% convert() -> res
+    expect_true(is.data.frame(res))
+})