Extract words from strings without spaces or delimiters using R - r

Question

babybag - baby bag
badshelter - bad shelter
themoderncornerstore - the modern corner store
hamptonfamilyguidebook - hampton family guide book
Is there a way to use R to extract words from string that do not have spaces or other delimiters? I have a list of URLs and I am trying to figure out what words are included in the URLs.
input <- c("babybag", "badshelter", "themoderncornerstore", "hamptonfamilyguidebook")

Answer 1

Here is a naive approach that might give you inspiration, I used library hunspell but you could test substrings against any dictionary.
I start from the right, try every substring and keep the longest I can find in the dictionary, then change my starting position, it's quite slow so I hope you don't have 4 millions of those. hampton is not in this dictionary so it doesn't give the right result for the last one :
split_words <- function(x){
candidate <- x
words <- NULL
j <- nchar(x)
while(j !=0){
word <- NULL
for (i in j:1){
candidate <- substr(x,i,j)
if(!length(hunspell::hunspell_find(candidate)[[1]])) word <- candidate
}
if(is.null(word)) return("")
words <- c(word,words)
j <- j-nchar(word)
}
words
}
input <- c("babybag", "badshelter", "themoderncornerstore", "hamptonfamilyguidebook")
lapply(input,split_words)
# [[1]]
# [1] "baby" "bag"
#
# [[2]]
# [1] "bad" "shelter"
#
# [[3]]
# [1] "the" "modern" "corner" "store"
#
# [[4]]
# [1] "h" "amp" "ton" "family" "guidebook"
#
Here's a quick fix, adding words manually to the dictionary:
split_words <- function(x, additional = c("hampton","otherwordstoadd")){
candidate <- x
words <- NULL
j <- nchar(x)
while(j !=0){
word <- NULL
for (i in j:1){
candidate <- substr(x,i,j)
if(!length(hunspell::hunspell_find(candidate,ignore = additional)[[1]])) word <- candidate
}
if(is.null(word)) return("")
words <- c(word,words)
j <- j-nchar(word)
}
words
}
input <- c("babybag", "badshelter", "themoderncornerstore", "hamptonfamilyguidebook")
lapply(input,split_words)
# [[1]]
# [1] "baby" "bag"
#
# [[2]]
# [1] "bad" "shelter"
#
# [[3]]
# [1] "the" "modern" "corner" "store"
#
# [[4]]
# [1] "hampton" "family" "guidebook"
#
You can just cross fingers not to have any ambiguous expressions though. Note that "guidebook" is in one word in my output so we already have an edge case in your four examples.