1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
|
#' Create Counts of Patterns using Regular Expressions
#'
#' `step_count` creates a *specification* of a recipe
#' step that will create a variable that counts instances of a
#' regular expression pattern in text.
#'
#' @inheritParams step_pca
#' @inheritParams step_center
#' @param ... A single selector function to choose which variable
#' will be searched for the regex pattern. The selector should
#' resolve to a single variable. See [selections()] for more details.
#' @param pattern A character string containing a regular
#' expression (or character string for `fixed = TRUE`) to be
#' matched in the given character vector. Coerced by
#' `as.character` to a character string if possible.
#' @param normalize A logical; should the integer counts be
#' divided by the total number of characters in the string?.
#' @param options A list of options to [gregexpr()] that
#' should not include `x` or `pattern`.
#' @param result A single character value for the name of the new
#' variable. It should be a valid column name.
#' @param input A single character value for the name of the
#' variable being searched. This is `NULL` until computed by
#' [prep()].
#' @template step-return
#' @details
#'
#' # Tidying
#'
#' When you [`tidy()`][tidy.recipe()] this step, a tibble with columns
#' `terms` (the selectors or variables selected) and `result` (the
#' new column name) is returned.
#'
#' @template case-weights-not-supported
#'
#' @family dummy variable and encoding steps
#' @export
#' @examplesIf rlang::is_installed("modeldata")
#' data(covers, package = "modeldata")
#'
#' rec <- recipe(~description, covers) %>%
#' step_count(description, pattern = "(rock|stony)", result = "rocks") %>%
#' step_count(description, pattern = "famil", normalize = TRUE)
#'
#' rec2 <- prep(rec, training = covers)
#' rec2
#'
#' count_values <- bake(rec2, new_data = covers)
#' count_values
#'
#' tidy(rec, number = 1)
#' tidy(rec2, number = 1)
step_count <- function(recipe,
...,
role = "predictor",
trained = FALSE,
pattern = ".",
normalize = FALSE,
options = list(),
result = make.names(pattern),
input = NULL,
skip = FALSE,
id = rand_id("count")) {
if (!is.character(pattern)) {
rlang::abort("`pattern` should be a character string")
}
if (length(pattern) != 1) {
rlang::abort("`pattern` should be a single pattern")
}
valid_args <- names(formals(grepl))[-(1:2)]
if (any(!(names(options) %in% valid_args))) {
rlang::abort(paste0(
"Valid options are: ",
paste0(valid_args, collapse = ", ")
))
}
terms <- enquos(...)
if (length(terms) > 1) {
rlang::abort("For this step, only a single selector can be used.")
}
add_step(
recipe,
step_count_new(
terms = terms,
role = role,
trained = trained,
pattern = pattern,
normalize = normalize,
options = options,
result = result,
input = input,
skip = skip,
id = id
)
)
}
step_count_new <-
function(terms, role, trained, pattern, normalize, options, result, input, skip, id) {
step(
subclass = "count",
terms = terms,
role = role,
trained = trained,
pattern = pattern,
normalize = normalize,
options = options,
result = result,
input = input,
skip = skip,
id = id
)
}
#' @export
prep.step_count <- function(x, training, info = NULL, ...) {
col_name <- recipes_eval_select(x$terms, training, info)
check_type(training[, col_name], types = c("string", "factor", "ordered"))
if (length(col_name) > 1) {
rlang::abort("The selector should select at most a single variable")
}
step_count_new(
terms = x$terms,
role = x$role,
trained = TRUE,
pattern = x$pattern,
normalize = x$normalize,
options = x$options,
input = col_name,
result = x$result,
skip = x$skip,
id = x$id
)
}
bake.step_count <- function(object, new_data, ...) {
check_new_data(names(object$input), object, new_data)
if (length(object$input) == 0L) {
# Empty selection, but still return the new column
new_data[, object$result] <- if (object$normalize) NA_real_ else NA_integer_
return(new_data)
}
## sub in options
regex <- expr(
gregexpr(
text = getElement(new_data, object$input),
pattern = object$pattern,
ignore.case = FALSE,
perl = FALSE,
fixed = FALSE,
useBytes = FALSE
)
)
if (length(object$options) > 0) {
regex <- rlang::call_modify(regex, !!!object$options)
}
new_data[, object$result] <- vapply(eval(regex), counter, integer(1))
if (object$normalize) {
totals <- nchar(as.character(getElement(new_data, object$input)))
new_data[, object$result] <- new_data[, object$result] / totals
}
new_data
}
counter <- function(x) length(x[x > 0])
print.step_count <-
function(x, width = max(20, options()$width - 30), ...) {
title <- "Regular expression counts using "
print_step(x$input, x$terms, x$trained, title, width)
invisible(x)
}
#' @rdname tidy.recipe
#' @export
tidy.step_count <- function(x, ...) {
term_names <- sel2char(x$terms)
p <- length(term_names)
if (is_trained(x)) {
res <- tibble(
terms = term_names,
result = rep(x$result, p)
)
} else {
res <- tibble(
terms = term_names,
result = rep(na_chr, p)
)
}
res$id <- x$id
res
}
|
