Hello! I am trying to create an application (for my wedding 
) that reads in poetry (of different genres and languages, which you can choose from) and then generates new "poems" using Markov Chain sequences. It works great on my local machine, but the encoding breaks when printing out the generated poem; a problem when Finnish is littered with "ä" and "ö". The English language poems show nicely but I'd like to show the Finnish ones as well.
Curiously, the "ä" shows correctly on UI title panel. It only breaks down on textOutput.
Can I change deployment specifications or what can I do to have my printed poem show correctly? This is really annoying.
Here is the application hosted on shinyapps: https://lassewinter.shinyapps.io/talvipelto18/
And this is the code. Both ui.R and server.R files are saved with UTF-8 encoding on my local computer.
ui.R:
library(shiny)
ui <- fluidPage(
titlePanel("#Talvipelto18 Virallinen hääapplikaatio!"),
sidebarLayout(
sidebarPanel(
selectInput("tyylilaji", "Valitse runosi tyylilaji",
c("Moderni suomenkielinen"=1,
"Moderni englanninkielinen"=2,
"Klassinen englanninkielinen"=3,
"Klassinen suomenkielinen"=4,
"Shakespeare"=5))
),
mainPanel(
textOutput("runo")
)
)
)
server.R
#the official wedding App!
library(shiny)
library(gutenbergr)
library(dplyr)
library(markovchain)
library(RCurl)#collect books from Gutenberg
kirjat <- gutenberg_download(c(27899,23543,43521,27898,20883,47985), meta_fields = "author", strip = TRUE)
books <- gutenberg_download(c(45470,1141,5125,4009,37649,15311,2491), meta_fields = "author", strip = TRUE)
shakespeare <- gutenberg_download(1041, meta_fields = "author", strip = TRUE)
#read more poems from my dropbox account
modernitRunot <- readLines("newuserwasallowedtopostonly2links", skipNul = TRUE,
encoding = "latin1")modernitRunot <- gsub("ÿþ","",modernitRunot)
modernitRunot <- modernitRunot[nchar(modernitRunot) > 0]poems <- readLines("https://www.dropbox.com/s/qdd3cngd4d5hz1u/lovePoems.txt?raw=1", skipNul = TRUE,
encoding = "latin1")poems <- poems[nchar(poems) > 0]
#set encoding
Encoding(kirjat$text) <- "latin1"
Encoding(books$text) <- "latin1"
Encoding(shakespeare$text) <- "latin1"#clear empty rows
kirjat <- filter(kirjat, nchar(text)>0)
books <- filter(books, nchar(text) > 0)
shakespeare <- filter(shakespeare, nchar(text) > 0)server part starts
shinyServer(function(input, output){
output$runo <- renderText({
#choose which genre your generated poem reflects if(input$tyylilaji==1){ teksti <-modernitRunot }else if(input$tyylilaji==2){ teksti <- poems }else if(input$tyylilaji==3){ teksti <- paste(books$text, collapse = " ") }else if(input$tyylilaji==4){ teksti <- paste(kirjat$text, collapse = " ") }else{ teksti <- paste(shakespeare$text, collapse = " ") } #split text to terms terms <- unlist(strsplit(teksti, ' ')) terms <- terms[nchar(terms) > 0] #limit to max 2000 terms for quicker model fits if(length(terms) > 2000){ alku <- round(runif(1,1,length(terms)-2000),0) loppu <- alku + 2000 terms2 <- terms[alku:loppu] }else{ terms2 <- terms } #fit markov chain fit <- markovchainFit(data = terms2) #generate poem runo <- paste(markovchainSequence(n=20, markovchain=fit$estimate), collapse=' ') Encoding(runo) <- "latin1" runo})
})
Encoding problems, the bane of non-English coders...
