问题
How do you justify text in R? By justify I mean that each line in a paragraph is exactly the same length (like when you justify in open office or excel). I have tried to find an option with strwrap and cat but have been unsuccessful.
## Get some sample text example from wikipedia api
library(httr)
library(xml2)
name <- "Invictus"
url <- URLencode(sprintf("https://en.wikisource.org/w/api.php?action=parse&prop=text&page=%s&format=json", name))
res <- read_html(content(GET(url))$parse$text[[1]])
string <- iconv(xml_text(xml_find_all(res, "//p"), trim=TRUE), "latin1", "ASCII", sub=" ")[1:2]
(string <- trimws(gsub('\\n|\\s{3,}', ' ', paste(string, collapse=" "))))
# [1] "Out of the night that covers me, Black as the pit from pole to pole, I thank whatever gods may be For my unconquerable soul. In the fell clutch of circumstance I have not winced nor cried aloud. Under the bludgeonings of chance My head is bloody, but unbow'd. Beyond this place of wrath and tears Looms but the Horror of the shade, And yet the menace of the years Finds and shall find me unafraid. It matters not how strait the gate, How charged with punishments the scroll, I am the master of my fate: I am the captain of my soul."
Some attempts using the aforementioned functions
## Using these I can get left/right/center justified text but not
## justified like in other text editing programs or newspapers.
width <- 30
cat(paste(strwrap(string, width=width), collapse='\n'))
## Or with cat
tokens <- strsplit(string, '\\s+')[[1]] # tokenise to pass to cat
out <- capture.output(cat(tokens, fill=width, sep=" ")) # strings <= width chars
cat(paste(out, collapse='\n'))
回答1:
Well, if there is no built-in way this works well enough for my purposes. Thanks for the comments above about how to use html styling as well.
justify <- function(string, width=getOption('width'),
fill=c('random', 'right', 'left')) {
strs <- strwrap(string, width=width)
paste(fill_spaces(strs, width, match.arg(fill)), collapse="\n")
}
fill_spaces <- function(lines, width, fill) {
tokens <- strsplit(lines, '\\s+')
res <- lapply(head(tokens, -1L), function(x) {
nspace <- length(x)-1L
extra <- width - sum(nchar(x)) - nspace
reps <- extra %/% nspace
extra <- extra %% nspace
times <- rep.int(if (reps>0) reps+1L else 1L, nspace)
if (extra > 0) {
if (fill=='right') times[1:extra] <- times[1:extra]+1L
else if (fill=='left')
times[(nspace-extra+1L):nspace] <- times[(nspace-extra+1L):nspace]+1L
else times[inds] <- times[(inds <- sample(nspace, extra))]+1L
}
spaces <- c('', unlist(lapply(times, formatC, x=' ', digits=NULL)))
paste(c(rbind(spaces, x)), collapse='')
})
c(res, paste(tail(tokens, 1L)[[1]], collapse = ' '))
}
cat(justify(string, width=40))
# Out of the night that covers me, Black
# as the pit from pole to pole, I thank
# whatever gods may be For my
# unconquerable soul. In the fell clutch
# of circumstance I have not winced nor
# cried aloud. Under the bludgeonings of
# chance My head is bloody, but unbow'd.
# Beyond this place of wrath and tears
# Looms but the Horror of the shade, And
# yet the menace of the years Finds and
# shall find me unafraid. It matters not
# how strait the gate, How charged with
# punishments the scroll, I am the master
# of my fate: I am the captain of my
# soul.
回答2:
@jenesaisquoi - great solution! But I found it does not work if there is a paragraph break with a space or if strs <- strwrap(string, width=width) returns one or fewer elements.
So, I found an improved version that first splits string by paragraph/line breaks and then applies the same logic:
justify = function(string, width = getOption('width'),
fill = c('random', 'right', 'left')) {
# Split text into paragraphs and remove trailing and leading white space.
paragraphs = gsub("^\\s+|\\s+$", "",
unlist(strsplit(x = string, split = "\n", fixed = TRUE)))
# NOTE: Empty elements are paragraphs break.
paragraphs = paragraphs[nchar(paragraphs) > 0]
formatted_text = lapply(paragraphs, function(paragraph){
strs = strwrap(paragraph, width = width)
paste(fill_spaces(strs, width, fill), collapse = "\n")
})
paste0(unlist(formatted_text, recursive = FALSE), collapse = "\n")
}
fill_spaces = function(lines, width, fill) {
tokens = strsplit(lines, '\\s+')
res = lapply(head(tokens, -1L), function(x) {
nspace = length(x) - 1L
extra = width - sum(nchar(x)) - nspace
reps = extra %/% nspace
extra = extra %% nspace
times = rep.int(if (reps > 0) reps + 1L else 1L, nspace)
if (extra > 0) {
if (fill == 'right') times[1:extra] = times[1:extra] + 1L
else if (fill == 'left')
times[(nspace - extra + 1L):nspace] = times[(nspace - extra + 1L):nspace] + 1L
else times[inds] = times[(inds <- sample(nspace, extra))] + 1L
}
spaces <- c('', unlist(lapply(times, formatC, x = ' ', digits = NULL)))
paste(c(rbind(spaces, x)), collapse = '')
})
c(res, paste(tail(tokens, 1L)[[1]], collapse = ' '))
}
nchar_per_line = 50
string = "Colin\'s practice outfits have reached a new level recently. It's difficult to determine the effect they are having on his teammates - whether they serve more as a distraction or a nice bit of comice releif, but it is clear they fuel Colin's fire and act as a motivator to him. \n\n On another note, Colin's high fives per 36 have been through the roof recently. It really seems like something he's been focusing on in practice lately, as well as putting extra reps in at the gym. \n\n Keep an eye out for Colin, could be an interesting 10 day pickup down the line."
cat(justify(string, width = nchar_per_line))
# Colin's practice outfits have reached a new level
# recently. It's difficult to determine the effect
# they are having on his teammates - whether they
# serve more as a distraction or a nice bit of
# comice releif, but it is clear they fuel Colin's
# fire and act as a motivator to him.
# On another note, Colin's high fives per 36 have
# been through the roof recently. It really seems
# like something he's been focusing on in practice
# lately, as well as putting extra reps in at the
# gym.
# Keep an eye out for Colin, could be an
# interesting 10 day pickup down the line.
来源：https://stackoverflow.com/questions/34710597/justify-text-in-r