replace_tokens: Replace Tokens

In trinker/textclean: Text Cleaning Tools

Description

Replace tokens with a single substring. This is much faster than mgsub if one wants to replace fixed tokens with a single value or remove them all together. This can be useful for quickly replacing tokens like names in string with a single value in order to reduce noise.

Usage

replace_tokens(x, tokens, replacement = NULL, ignore.case = FALSE, ...)

Arguments

x	A character vector.
tokens	A vector of token to be replaced.
replacement	A single character string to replace the tokens with. The default, NULL, replaces the tokens with nothing.
ignore.case	logical. If TRUE the case of the tokens will be ignored.
...	ignored.

Value

Returns a vector of strings with tokens replaced.

Note

The function splits the string apart into tokens for speed optimization. After the replacement occurs the strings are pasted back together. The strings are not guaranteed to retain exact spacing of the original.

See Also

mgsub, match_tokens

Examples

replace_tokens(DATA$state, c('No', 'what', "it's"))
replace_tokens(DATA$state, c('No', 'what', "it's"), "<<TOKEN>>")
replace_tokens(
    DATA$state, 
    c('No', 'what', "it's"), 
    "<<TOKEN>>", 
    ignore.case = TRUE
)

## Not run: 
## Now let's see the speed
## Set up data
library(textshape)
data(hamlet)
set.seed(11)
tokens <- sample(unique(unlist(split_token(hamlet$dialogue))), 2000)

tic <- Sys.time()
head(replace_tokens(hamlet$dialogue, tokens))
(toc <- Sys.time() - tic)


tic <- Sys.time()
head(mgsub(hamlet$dialogue, tokens, ""))
(toc <- Sys.time() - tic)


## Amp it up 20x more data
tic <- Sys.time()
head(replace_tokens(rep(hamlet$dialogue, 20), tokens))
(toc <- Sys.time() - tic)

## Replace names example

library(lexicon)
library(textshape)
nms <- gsub("(^.)(.*)", "\\U\\1\\L\\2", common_names, perl = TRUE)
x <- split_portion(
    sample(c(sample(grady_augmented, 5000), sample(nms, 10000, TRUE))), 
    n.words = 12
)
x$text.var <- paste0(
    x$text.var, 
    sample(c('.', '!', '?'), length(x$text.var), TRUE)
 )
replace_tokens(x$text.var, nms, 'NAME')

## End(Not run)

trinker/textclean documentation built on Nov. 3, 2021, 7:20 p.m.