0
Da dies mein erster StackFlow-Beitrag ist, wenn ich irgendetwas nicht ordnungsgemäß gemacht habe, lassen Sie es mich bitte wissen. Ich arbeite an meiner zweiten ShinyApp. Die App soll einen Ordner lesen und alle * txt-Dateien für eine Analyse zurückgeben. Ich habe versucht, die shinyFiles-Anwendung und die shinyDirChoose-Funktion zu verwenden, um dies zu erreichen. Wenn ich getVolumes() benutze und meine App starte, kann ich keine Dateien oder Ordner meines Computers sehen. Ich bin sicher, dass mein Code auch einige andere Fehler hat, aber ich kann nicht einmal richtig zum Debuggen laufen, wenn ich den Ordner nicht richtig einlesen kann.shinyDirchoose GetVolumes() gibt keine Verzeichnisse zurück
library(ggplot2)
library(Cairo)
library(DT)
library(plyr)
library(shinythemes)
library(shinyFiles)
shinyApp(
ui = fluidPage(
theme = shinytheme("superhero"),
titlePanel(
fluidRow(
column(9, strong("Winnow Output")),
column(3, img(height = 150, width = 150, src = "spirit_black.jpg"))
)
),
sidebarLayout(
sidebarPanel(
fluidRow(
column(width=12,
shinyDirButton('directory', 'Folder select', 'Please select a folder'),
h3("Directory Selected"),
verbatimTextOutput("dirpath"),
div(class = "option_group",
radioButtons("plot_type", "Plot type",
c("Scatter", "Linear Rates", "AUC by MAE","True-Positives by False-Positives"), inline = FALSE),
conditionalPanel("input.plot_type === 'Scatter'",
selectInput("pvar1", "x-Var",
c("tp" = "tp",
"fp" = "fp"
),
selected = "tp"
),
selectInput("pvar2", "y-Var",
c("tp" = "tp",
"fp" = "fp"
),
selected = "fp"
)
),
conditionalPanel("input.plot_type === 'Linear Rates'",
selectInput("lvar1", "x-Var",
c("tpr" = "tpr",
"fpr" = "fpr"
),
selected = "tpr"
),
selectInput("lvar2", "y-Var",
c("tpr" = "tpr",
"fpr" = "fpr"
),
selected = "fpr"
)
),
conditionalPanel("input.plot_type === 'AUC by MAE'",
sliderInput("auc.min", "AUC axis minimum", min = 0, max = 2, value = .5,step=0.05),
sliderInput("auc.max", "AUC axis maximum", min = 0, max = 2, value = 1,step=0.05),
sliderInput("mae.min", "MAE axis minimum", min = 0, max = 2, value = 0,step=0.05),
sliderInput("mae.max", "MAE axis maximum", min = 0, max = 2, value = 1,step=0.05)
)
)#ends div
)#ends column
)#ends fluidRow
)#ends sidebarPanel
,
mainPanel(
tabsetPanel(
tabPanel("Plot",plotOutput("plot1")),
tabPanel("Table",
fluidRow(
column(4,
h4("means"),
tableOutput("meanresults")
),
column(4,
h4("sums"),
tableOutput("sumresults")
),
column(4,
h4("means"),
tableOutput("meanresults2")
)
)#end fluidRow
)#ends tablPale
)#ends tabsetPanel
)#ends mainPanel
)#ends sidebarLayout
),#ends ui = fluidPage
server = function(input, output) {
roots <- getVolumes()
folderInput1 <- reactive({
shinyDirChoose(input, 'directory', roots = roots, session = session,
restrictions = system.file(package = 'base'))
return(parseDirPath(volumes, input$directory))
})
output$directorypath = renderPrint({ folderInput1() })
files1 <- reactive({
list.files(path = folderInput1(), pattern = "*.txt", full.names = T)
})
nFiles1 <- reactive({ length(files1()) })
filenames <- reactive ({file_path_sans_ext(files1)})
myfiles <- reactive({lapply(files1, function(x) read.table(x, header=TRUE))})
tpmax<-function(list){
tps<-list()
for (i in 1:length(list)){
tps[[i]]<-list[[i]]$tp
}
y<-unlist(lapply(tps, max))
return(max(y))
}
ta<-reactive({tpmax(myfiles)})
#Determines minimum value of all true positives
tpmin<-function(list){
tps<-list()
for (i in 1:length(list)){
tps[[i]]<-list[[i]]$tp
}
y<-unlist(lapply(tps, min))
return(min(y))
}
tb<-reactive({tpmin(myfiles)})
#Determines median value of all true positives
tpmed<-function(list){
tps<-list()
for (i in 1:length(list)){
tps[[i]]<-list[[i]]$tp
}
y<-unlist(lapply(tps, median))
return(median(y))
}
tc<-reactive({tpmed(myfiles)})
#Determines maximum value of all false positives
fpmax<-function(list){
fps<-list()
for (i in 1:length(list)){
fps[[i]]<-list[[i]]$fp
}
y<-unlist(lapply(fps, max))
return(max(y))
}
fa<-reactive ({fpmax(myfiles)})
#Determines minimum value of all false positives
fpmin<-function(list){
fps<-list()
for (i in 1:length(list)){
fps[[i]]<-list[[i]]$fp
}
y<-unlist(lapply(fps, min))
return(min(y))
}
fb<-reactive({fpmin(myfiles)})
#Determines median value of all false positives
fpmed<-function(list){
fps<-list()
for (i in 1:length(list)){
fps[[i]]<-list[[i]]$fp
y<-unlist(lapply(fps, median))
return(median(y))
}
}
fc<-reactive({fpmed(myfiles)})
all.data<<-reactive ({do.call("rbind", myfiles)})
TPFP <- reactive ({ddply(all.data, .(tp, fp, filenames), summarize, count=length(filenames))})
output$plot1 = renderPlot({
if (input$plot_type == "Scatter") {
ggplot(all.data(),aes_string(x=input$pvar1,y=input$pvar2))+geom_point(color="firebrick")
} else if (input$plot_type == "Linear Rates") {
ggplot(all.data,aes_string(x=input$lvar1,y=input$lvar2))+geom_line(size = 1, alpha = 1)+
labs(title= "Comparison of Rates")
}
else if (input$plot_type == "AUC by MAE"){
plot(myfiles[[1]]$mae, myfiles[[1]]$auc, main="Plot of AUC by MAE", xlab="Mean Absolute Error (MAE)", ylab="Area under R-O Curve (AUC)",
pch=21, bg="black", xlim=c(input$mae.min, input$mae.max), ylim=c(input$auc.min,input$auc.max))
plotcol<-c("black")
if (length(myfiles) > 1){
#Create overlapping data plots to compare potentially by GWAS tool
#assuming that the length of the Winnow files is at least 2
for (i in 2:length(myfiles)){
points(myfiles[[i]]$mae, myfiles[[i]]$auc, main="Plot of AUC by MAE", xlab="Mean Absolute Error (MAE)", ylab="Area under R-O Curve (AUC)",
pch=21, bg=rainbow(i+1)[i], xlim=c(input$mae.min, input$mae.max), ylim=c(input$auc.min, input$auc.max))
plotcol[i]<-rainbow(i+1)[i]
}
}
}
else if (input$plot_type == "True-Positives by False-Positives"){
p <- ggplot(TPFP(), aes_string(x=fp, y=tp),environment=environment())
p2 <- p +
geom_rect(data=all.data[1,], aes(xmin=fc, xmax=fa, ymin=tc, ymax=ta),
alpha=0.2, fill="blue", linetype=0) +
geom_rect(data=all.data[1,], aes(xmin=fb, xmax=fc, ymin=tc, ymax=ta),
alpha=0.2,fill="green", linetype=0) +
geom_rect(data=all.data[1,], aes(xmin=fb, xmax=fc, ymin=tb, ymax=tc),
alpha=0.2, fill="blue", linetype=0) +
geom_rect(data=all.data[1,], aes(xmin=fc, xmax=fa, ymin=tb, ymax=tc),
alpha=0.2, fill="gray", linetype=0) +
theme(panel.background=element_rect(fill='white', colour='black')) +
theme(panel.grid.major=element_blank(), panel.grid.minor=element_blank()) +
geom_point(aes(colour=filenames, size=count)) +
scale_size_continuous(range=c(2,8)) +
xlab("False Positives") +
ylab("True Positives") +
ggtitle("False Positives by True Positves") +
xlim(0, fa) + ylim(0, ta)
print(p2)
}
})
output$meanresults = renderTable({apply(all.data()[,1:5],2,mean)},rownames=TRUE)
output$sumresults = renderTable(apply(all.data()[,6:9],2,sum),rownames=TRUE)
output$meanresults2 = renderTable(apply(all.data()[,10:16],2,mean),rownames=TRUE)