Using Shiny App to dynamically visualise each iteration of string replacement? - r

Question

I would like to modify the input and display the output of some R code in a more interactive way. I think that this would be an ideal task for a Shiny app, but I am not very familiar with writing them. I have some R code that takes a string of text and iteratively changes it by adding letters or words at random locations:
library(tidyverse)
evolve_sentence <- function(sentence, arg2) {
chars <- str_split(sentence, "") %>% pluck(1)
if (runif(1) > 0.5) {
chars[sample(1:length(chars), 1)] <- sample(chars, 1)
}
sentence <- str_c(chars, collapse = "")
words <- str_split(sentence, " ") %>% pluck(1)
if (runif(1) > 0.9) {
words[sample(1:length(words), 1)] <- sample(words, 1)
}
sentence <- str_c(words, collapse = " ")
sentence
}
tbl_evolve <- tibble(iteration = 1:500, text = "I met a traveller from an antique land")
for (i in 2:500) {
tbl_evolve$text[i] <- evolve_sentence(tbl_evolve$text[i - 1])
}
tbl_evolve %>%
distinct(text, .keep_all = TRUE)
The output look like this:
1 I met a traveller from an antique land
2 I met a tIaveller from an antique land
4 I met a tIaveller from an antique lanr
5 I met a tIaveller from an fntique lanr
6 I met a tIaveller fromnan met lanr
I would love to present this in as a Shiny app where the input text and the probability of different types of changes can be specified by the user. For the latter that would be making the values in (runif(1) > 0.5) and (runif(1) > 0.9) specifiable by the user. I know this is possible in Shiny using insert UI and actionButton.
I am less sure if there is a way to dynamically show the output, so that the user can visually see each iteration of the code (with a defined time delay between each iteration?) rather than seeing all iterations output at once as once does with the existing code. I am open to different ways of dynamically visualising the output but I think ideally the user would see each iteration replaced by the next with a time delay.
I would also like a tab with the current output, each iteration being a row, so the user can go back and review each iteration.
Any advice on whether this is possible in Shiny or if I need a different tool would be much appreciated.

Answer 1

library(shiny)
library(tidyverse)
# Define UI for application that draws a histogram
ui <- fluidPage(
# Application title
titlePanel("Simple Testcase"),
# Sidebar with a slider input for number of bins
sidebarLayout(
sidebarPanel(
textInput("textinput", "Type text here"),
numericInput("p1", "Probability1", value = 0.5),
numericInput("p2", "Probability2", value = 0.9),
sliderInput("iteration", "Iterations", min = 20, max = 1000, step = 10, value = 100),
actionButton("calc", "Run Calculation!")
),
# Show a plot of the generated distribution
mainPanel(
tableOutput("ui")
)
)
)
# Define server logic required to draw a histogram
server <- function(session ,input, output) {
vals <- reactiveValues(counter = 0)
result <- eventReactive(input$calc, {
evolve_sentence <- function(sentence, arg2) {
chars <- str_split(sentence, "") %>% pluck(1)
if (runif(1) > input$p1) { # Value from numericinput p2
chars[sample(1:length(chars), 1)] <- sample(chars, 1)
}
sentence <- str_c(chars, collapse = "")
words <- str_split(sentence, " ") %>% pluck(1)
if (runif(1) > input$p2) { # Value from numericinput p2
words[sample(1:length(words), 1)] <- sample(words, 1)
}
sentence <- str_c(words, collapse = " ")
sentence
}
tbl_evolve <- tibble(iteration = 1:500, text = input$textinput)
for (i in 2:500) {
tbl_evolve$text[i] <- evolve_sentence(tbl_evolve$text[i - 1])
}
output <-tbl_evolve %>%
distinct(text, .keep_all = TRUE)
print(output)
output
})
output$ui <- renderTable({
df <- result()
invalidateLater(millis = 300, session)
vals$counter <- isolate(vals$counter) + 1
while(nrow(df) < vals$counter) {
vals$counter <- isolate(vals$counter) + 1
} #Prevent to add infinite empty columns.
for(i in 1:nrow(df)) {
newdf <- df[1:vals$counter,]
}
newdf
})
}
# Run the application
shinyApp(ui = ui, server = server)
How about this? To render the table we can set a reactiveValue, which gets updated after the invalidateLater function triggers. Take the value of the counter to subset your final dataset.