Passing a dataframe as a parameter from Shiny app to RMarkdown - r

Question

I'm still struggling with some aspects of a Shiny app I'm working on. The intention is that the user uploads a csv file of data, which is then processed to generate a report (from a .Rmd template), which the user can then download as an editable Word .doc.
The .Rmd works fine if I render it in a normal R session. However, if done from my Shiny app, I get the following error:
Warning: Error in unique: object 'report.data' not found
[No stack trace available]
report.data should be the dataframe produced by reading the input .csv file. Confusingly, the app does sometimes work (I think this occurs if report.data is already available in the global environment.).
I've tried defining the params in the header of the .Rmd file (see the commented out lines below.) - if I do this then the code runs without an error, but the resulting word document is blank, except for the title.
Can anyone see where I'm going wrong? Thank you, as ever, for your time in reading this and replying.
And apologies, I feel like I'm making a lot of threads asking for help with what seem to be quite basic things in Shiny, but I do search for similar questions and never find things that are quite right! But once I have these basic things in place I should be able to make progress by myself.
Code to generate a .csv file of example input for report.data:
library(dplyr)
set.seed(1234)
product1.parameter1.location1 <- data.frame(
result = rnorm(25, mean = 2.5, sd = 0.2),
product = c("Red Aeroplanes"),
parameter = c("Parameter 1"),
sample.no = c(1:25),
location = c("Factory 1")
)
product1.parameter1.location2 <- data.frame(
result = rnorm(25, mean = 2.6, sd = 0.1),
product = c("Red Aeroplanes"),
parameter = c("Parameter 1"),
sample.no = c(1:25),
location = c("Factory 2")
)
product1 <- rbind(product1.parameter1.location1, product1.parameter1.location2)
product2.parameter1.location1 <- data.frame(
result = rnorm(25, mean = 10, sd = 2),
product = c("Blue Trollies"),
parameter = c("Parameter 1"),
sample.no = c(1:25),
location = c("Factory 1")
)
product2.parameter1.location2 <- data.frame(
result = rnorm(25, mean = 9.5, sd = 0.75),
product = c("Blue Trollies"),
parameter = c("Parameter 1"),
sample.no = c(1:25),
location = c("Factory 2"))
product2.parameter1 <- rbind(product2.parameter1.location1, product2.parameter1.location2)
product2.parameter2.location1 <- data.frame(
result = rnorm(25, mean = 30, sd = 1.8),
product = c("Blue Trollies"),
parameter = c("Parameter 2"),
sample.no = c(1:25),
location = c("Factory 1")
)
product2.parameter2.location2 <- data.frame(
result = rnorm(25, mean = 25, sd = 0.75),
product = c("Blue Trollies"),
parameter = c("Parameter 2"),
sample.no = c(1:25),
location = c("Factory 2"))
product2.parameter2 <- rbind(product2.parameter2.location1, product2.parameter2.location2)
product2 <- rbind(product2.parameter1, product2.parameter2)
product3.parameter1.location1 <- data.frame(
result = rnorm(35, mean = 2, sd = 0.2),
product = c("Brown Carriages"),
parameter = c("Parameter 1"),
sample.no = c(1:35),
location = c("Factory 1")
)
product3.parameter1.location2 <- data.frame(
result = rnorm(35, mean = 1.9, sd = 0.15),
product = c("Brown Carriages"),
parameter = c("Parameter 1"),
sample.no = c(1:35),
location = c("Factory 2"))
product3.parameter1 <- rbind(product3.parameter1.location1, product3.parameter1.location2)
product3.parameter2.location1 <- data.frame(
result = rnorm(35, mean = 4, sd = 0.4),
product = c("Brown Carriages"),
parameter = c("Parameter 2"),
sample.no = c(1:35),
location = c("Factory 1")
)
product3.parameter2.location2 <- data.frame(
result = rnorm(35, mean = 3.8, sd = 0.5),
product = c("Brown Carriages"),
parameter = c("Parameter 2"),
sample.no = c(1:35),
location = c("Factory 2"))
product3.parameter2 <- rbind(product3.parameter2.location1, product3.parameter2.location2)
product3.parameter3.location1 <- data.frame(
result = rnorm(35, mean = 10, sd = 1.8),
product = c("Brown Carriages"),
parameter = c("Parameter 3"),
sample.no = c(1:35),
location = c("Factory 1")
)
product3.parameter3.location2 <- data.frame(
result = rnorm(35, mean = 10, sd = 2),
product = c("Brown Carriages"),
parameter = c("Parameter 3"),
sample.no = c(1:35),
location = c("Factory 2"))
product3.parameter3 <- rbind(product3.parameter3.location1, product3.parameter3.location2)
product3 <- rbind(product3.parameter1, product3.parameter2, product3.parameter3)
write.csv(product1, "product1.csv", row.names = FALSE)
write.csv(product2, "product2.csv", row.names = FALSE)
write.csv(product3, "product3.csv", row.names = FALSE)
report.data <- rbind(product1, product2, product3) %>% mutate(identifier = paste(product, parameter, sep = " "))
write.csv(report.data, "all.data.csv", row.names = FALSE)
The app.R code:
#
# This is a Shiny web application. You can run the application by clicking
# the 'Run App' button above.
#
# Find out more about building applications with Shiny here:
#
# http://shiny.rstudio.com/
#
library(shiny)
# Define UI for application that draws a histogram
ui <- fluidPage(
# Application title
titlePanel("R Shiny app"),
# Sidebar with file input
sidebarLayout(
sidebarPanel(
fileInput(
inputId = "file1",
label = "Select file(s)",
multiple = TRUE,
accept = NULL,
width = NULL,
buttonLabel = "Browse...",
placeholder = "No file(s) selected"
),
downloadButton("report", "Generate report")
),
# Show a plot of the generated distribution
mainPanel(
plotOutput("distPlot")
)
)
)
# Define server logic required to draw a histogram
server <- function(input, output) {
output$report <- downloadHandler(
reactive(file <- input$file1),
# For PDF output, change this to "report.pdf"
filename = "report.doc",
content = function(file) {
# Copy the report file to a temporary directory before processing it, in
# case we don't have write permissions to the current working dir (which
# can happen when deployed).
tempReport <- file.path(tempdir(), "wordreport.Rmd")
file.copy("wordreport.Rmd", tempReport, overwrite = TRUE)
# Knit the document, passing in the `params` list, and eval it in a
# child of the global environment (this isolates the code in the document
# from the code in this app).
params <- list(report.data = input$file1)
rmarkdown::render(tempReport, output_file = "wordreport.doc",
params = params,
envir = new.env(parent = globalenv()))
file.copy("wordreport.doc", file)
}
)
}
# Run the application
shinyApp(ui = ui, server = server)
The .Rmd file (with the lines relating to params declaration commented out):
---
title: "Comparison Report for [CATEGORY] in [MONTH/YEAR]"
output: word_document
toc: yes
#params:
#report.data: report.data
---
```{r setup, include=FALSE, comment = "", results = 'asis', echo = FALSE}
library(dplyr)
library(ggplot2)
library(purrr)
knitr::opts_chunk$set(echo = FALSE)
```
#report.data <- params$report.data
```
my_plot <- function(df) {
ggplot(df, aes(x = sample.no, y = result)) +
geom_point(aes(colour = location)) +
geom_hline(aes(yintercept = mean(result)), colour = "black", linetype = "dotted") +
geom_hline(aes(yintercept = mean(result) + 1.96 * sd(result)), colour = "red", linetype = "dashed") +
geom_hline(aes(yintercept = mean(result) - 1.96 * sd(result)), colour = "red", linetype = "dashed") +
theme_classic() +
theme(legend.title = element_blank()) +
labs(
title = paste0("Comparison for ", unique(df$identifier)),
x = "Sample number",
y = "Result") +
#caption = paste0("Caption here.")) +
expand_limits(y = 0) +
coord_cartesian(xlim = c(0, max(df$sample.no) + 2)) +
theme(
plot.caption=element_text(size=12, hjust = 0, margin = margin(t=20)),
plot.margin = margin(b=50)
)
}
```
```{r, comment = "", results = 'asis', echo = FALSE}
purrr::map(unique(report.data$identifier),
function(x) {
#section heading
cat("#", (x), "\n")
cat("\n\n")
# filter data before passing it to the plot function
report.data %>%
dplyr::filter(identifier == x) %>%
my_plot() %>% print()
cat("\n\n")
no.outofbounds <- report.data %>%
dplyr::filter(identifier == x) %>%
mutate(outofbounds = ifelse(result > mean(result)+1.96*sd(result)|result < mean(result)-1.96*sd(result), TRUE, FALSE)) %>%
dplyr::filter(outofbounds == TRUE) %>%
nrow()
ifelse(no.outofbounds > 0, paste(cat(no.outofbounds), " results greater than 1.96 standard deviations away from the mean."), "All results within 1.96 standard deviations of the mean.") %>%
cat()
cat("\n\n")
CV <- report.data %>%
dplyr::filter(identifier == x) %>%
summarise(CV = sd(result)/mean(result) * 100) %>%
round(2)
cat("\n\n")
paste("The all-site/factor CV for this parameter is ", CV, "%.") %>%
cat()
cat("\n\n")
cat("APPROVED/REJECTED.")
cat("\n\n")
}
) -> results
```

Answer 1

There are several issues with your code. I'll go over them one by one
Invalid parameter in downloadHandler()
You are passing an object of class reactive to the contentType parameter of downloadHandler().
downloadHandler(
reactive(file <- input$file1), ## <--- here
filename = "report.doc",
content = function(file) {
# ...
}
)
It seems that this messes up the whole logic of downloadHandler() and leads to "server error" messages on the client side with no errors or warnings from shiny.
This line needs to be removed in order to download files successfully
Reference the Rmd-parameter correctly
When you want to access the parameter from the Rmd report, you will need to use params$report.data. Just using report.data will lead to the following error: object 'report.data' not found.
---
output: word_document
params:
report.data: NULL
---
```{r}
report.data <- params$report.data
# ...
```
Fix the path to the generated file
You are knitting the Rmd inside the temporary directory, which is generally a good idea. However, getting the paths right is not always that easy. In your case, I used the following
rendered_report <- rmarkdown::render(
tempReport, output_file = "wordreport.doc",
params = params,
envir = new.env(parent = globalenv())
)
file.copy(rendered_report, file)
The reason your version didn't work is that the generated report is created inside the temporary directory alogside tmpReport. See the reference documentation of ?rmarkdown::render for more details.
I used the return value of rmarkdown::render() instead which holds an absolute path to the generated file. This is less error prone and especially useful if you do not know the file extension of the generated file in advance
Use read.csv to convert the uploaded file into a data.frame
Shiny doesn't automatically convert uploaded csv files into dataframes. You need to define a parsing logic to do that.
params <- list(report.data = read.csv(input$file1$datapath))
One final word
Try to get more organized with your coding projects and limit the scope of future SO questions to one issue at a time. Creating "minimal reproducible examples" might seem tedious at first, but there are several advantages in doing that
Other people can read the questions and answers and reuse them in their own projects easily without dissecting a wall of code
It is much easier to answer those questions. With questions like this, the SO community usually only provides comments because answering them properly requires a lot of effort
Minimizing and isolating problems is a skill that will help you to figure out issues in your future coding projects much more easily