#' 'RCDMShinyViewer'
#'
#' RCDMShinyViewer function visualizes result of database analysis
#'
#'
#' @param Radiology_Occurrence_Table is a result of R-CDM : Radiology_Occurrence_Table<-radiologyOccurrenceTable(DICOMList)
#' @param Radiology_Image_Table is a result of R-CDM : Radiology_Image_Table<-radiologyImageTable(DICOMList)
#' @import dplyr
#' @import shiny
#' @import ggplot2
#' @import DT
#' @importFrom magrittr "%>%"
#'
#'
#' @return result of database analysis
#' @examples
#' Radiology_Occurrence_Table<-radiologyOccurrenceTable(DICOMList)
#' Radiology_Image_Table<-radiologyImageTable(DICOMList)
#' RCDMShinyViewer(Radiology_Occurrence_Table, Radiology_Image_Table)
#' @export
RCDMShinyViewer<-function(Radiology_Occurrence_Table, Radiology_Image_Table){
ui <- fluidPage(
titlePanel('Radiology-CDM'),
sidebarLayout(
sidebarPanel(
helpText('You can see reactive result of your database analysis by selecting Occurrence date, and modality'),
sliderInput(inputId = "dateTime",
"Occurrence date",
min = min(as.Date(Radiology_Occurrence_Table$studyDateTime)),
max = max(as.Date(Radiology_Occurrence_Table$studyDateTime)+1),
value = c(as.Date('2000-01-01'), as.Date('2015-01-01'))),
selectInput('mod', 'modality', choices = unique(Radiology_Occurrence_Table$modality), multiple = T),
helpText('After you select the occurrence date, and modality, you have to select protocol concept ID, and phase additionally. After that extract list of DICOM images files corresponding to the selected conditions by clicking extract button.'),
selectInput('Pro', 'select Radiology Protocol Concept ID', choices = unique(Radiology_Occurrence_Table$radiologyProtocolConceptId), multiple = T),
selectInput('Pha', 'select Radiology Phase Concept ID', choices = unique(Radiology_Image_Table$radiologyPhaseConceptId), multiple = T),
downloadButton('downloadData', 'Extract'),
width=3),
mainPanel(
fluidRow(
h2('Database Analysis'),
column(6,
h4('Count of occurrences'),
plotOutput(outputId = "modalityCount")),
column(6,
h4('Count of images'),
plotOutput(outputId = "imageCount")),
h2('Find images you want!'),
column(12,
h4('OMOP concept IDs in your database'),
DTOutput(outputId = "protocolConceptId"),
h3(textOutput('txt')))
),
width=9)
)
)
server <- function(input, output){
output$modalityCount <- renderPlot({
modality1<-Radiology_Occurrence_Table %>% select(modality)
barplot(table(factor(modality1$modality, levels=unique(Radiology_Occurrence_Table$modality))), border="#69b3a2", col="white", ylim=c(0, max(table(modality1))))
par(new=TRUE)
modality2<-Radiology_Occurrence_Table %>% filter(as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime)) %>% select(modality)
barplot(table(factor(modality2$modality, levels=unique(Radiology_Occurrence_Table$modality))), col=rgb(0.2,0.4,0.6,0.6), border="#69b3a2", ylim=c(0, max(table(modality1))))
par(new=TRUE)
modality3<-Radiology_Occurrence_Table %>% filter(as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime) & modality %in% c(input$mod)) %>% select(modality)
barplot(table(factor(modality3$modality, levels=unique(Radiology_Occurrence_Table$modality))), col=7, border="#69b3a2", ylim=c(0, max(table(modality1))))
par(new=TRUE)
})
output$imageCount <- renderPlot({
image1<-Radiology_Occurrence_Table %>% select(modality)
image2<-Radiology_Occurrence_Table %>% select(imageTotalCount)
barplot(table(factor(rep(image1$modality, image2$imageTotalCount), levels=unique(Radiology_Occurrence_Table$modality))), border="#69b3a2", col="white", ylim=c(0, max(table(rep(image1$modality, image2$imageTotalCount)))))
par(new=TRUE)
image3<-Radiology_Occurrence_Table %>% filter(as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime)) %>% select(modality)
image4<-Radiology_Occurrence_Table %>% filter(as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime)) %>% select(imageTotalCount)
barplot(table(factor(rep(image3$modality, image4$imageTotalCount), levels=unique(Radiology_Occurrence_Table$modality))), col=rgb(0.2,0.4,0.6,0.6), border="#69b3a2", ylim=c(0, max(table(rep(image1$modality, image2$imageTotalCount)))))
par(new=TRUE)
image5<-Radiology_Occurrence_Table %>% filter(as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime) & modality %in% c(input$mod)) %>% select(modality)
image6<-Radiology_Occurrence_Table %>% filter(as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime) & modality %in% c(input$mod)) %>% select(imageTotalCount)
barplot(table(factor(rep(image5$modality, image6$imageTotalCount), levels=unique(Radiology_Occurrence_Table$modality))), col=7, border="#69b3a2", ylim=c(0, max(table(rep(image1$modality, image2$imageTotalCount)))))
})
output$protocolConceptId <-renderDT({
RadiologyPlaybook<-unique(LoincRsnaRadiologyPlaybook[,c(2,3)])
RadiologyPlaybook<-data.frame(RadiologyPlaybook, row.names = NULL)
if(is.null(input$mod)==T){
dataframe1<-RadiologyPlaybook %>% filter(radiologyProtocolConceptId %in% Radiology_Occurrence_Table$radiologyProtocolConceptId)
dataframe5<-Radiology_Occurrence_Table %>% filter(radiologyProtocolConceptId %in% RadiologyPlaybook$radiologyProtocolConceptId & as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime)) %>% select(radiologyProtocolConceptId, imageTotalCount, radiologyOccurrenceId)
dataframe2<-split(dataframe5, as.character(dataframe5$radiologyProtocolConceptId))
dataframe2<-lapply(dataframe2, function(x){
return(data.frame(radiologyProtocolConceptId=unique(x$radiologyProtocolConceptId), Count_of_occurrences=nrow(x), Count_of_images=sum(x$imageTotalCount)))}
)
dataframe2<-do.call(rbind, dataframe2)
dataframe2<-data.frame(dataframe2, row.names = NULL)
answerDF<-merge(dataframe1, dataframe2, by='radiologyProtocolConceptId')
dataframe4<-Radiology_Occurrence_Table %>% filter(radiologyOccurrenceId %in% as.character(dataframe5$radiologyOccurrenceId)) %>% select(radiologyOccurrenceId, radiologyProtocolConceptId)
dataframe3<-Radiology_Image_Table %>% filter(radiologyOccurrenceId %in% as.character(dataframe4$radiologyOccurrenceId)) %>% select(radiologyOccurrenceId, radiologyPhaseConceptId) %>% group_by(radiologyOccurrenceId, radiologyPhaseConceptId) %>% count()
answer<-merge(dataframe4, dataframe3, by='radiologyOccurrenceId')
answer<-split(answer, as.character(answer$radiologyProtocolConceptId))
answer<-lapply(answer, function(x){
answer<-split(x, as.character(x$radiologyPhaseConceptId))
answer<-sapply(answer, function(y){
sum(y$n)
})
ImageCount<-c()
for (i in 1:length(answer)){
ImageCount<-c(ImageCount, sprintf('%s : %d', names(answer)[i], answer[i]))
}
ImageCount<-paste(ImageCount, collapse = ' / ')
return(data.frame(radiologyProtocolConceptId=unique(x$radiologyProtocolConceptId), counts=ImageCount))
})
answer<-do.call(rbind, answer)
answer<-merge(answerDF, answer, by='radiologyProtocolConceptId')
my_vals = answer$radiologyProtocolConceptId
my_colors = ifelse(my_vals %in% input$Pro,'orange','white')
datatable(answer) %>% formatStyle(
'radiologyProtocolConceptId',
target = 'row',
backgroundColor = styleEqual(my_vals, my_colors))
} else if (is.null(input$mod)==F) {
mods<-paste(input$mod, collapse ='|')
dataframe1<-RadiologyPlaybook %>% filter(radiologyProtocolConceptId %in% Radiology_Occurrence_Table$radiologyProtocolConceptId & grepl(mods, LongCommonName)==T)
dataframe5<-Radiology_Occurrence_Table %>% filter(radiologyProtocolConceptId %in% RadiologyPlaybook$radiologyProtocolConceptId & as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime)) %>% select(radiologyProtocolConceptId, imageTotalCount, radiologyOccurrenceId)
dataframe2<-split(dataframe5, as.character(dataframe5$radiologyProtocolConceptId))
dataframe2<-lapply(dataframe2, function(x){
return(data.frame(radiologyProtocolConceptId=unique(x$radiologyProtocolConceptId), Count_of_occurrences=nrow(x), Count_of_images=sum(x$imageTotalCount)))}
)
dataframe2<-do.call(rbind, dataframe2)
dataframe2<-data.frame(dataframe2, row.names = NULL)
answerDF<-merge(dataframe1, dataframe2, by='radiologyProtocolConceptId')
dataframe4<-Radiology_Occurrence_Table %>% filter(radiologyOccurrenceId %in% as.character(dataframe5$radiologyOccurrenceId)) %>% select(radiologyOccurrenceId, radiologyProtocolConceptId)
dataframe3<-Radiology_Image_Table %>% filter(radiologyOccurrenceId %in% as.character(dataframe4$radiologyOccurrenceId)) %>% select(radiologyOccurrenceId, radiologyPhaseConceptId) %>% group_by(radiologyOccurrenceId, radiologyPhaseConceptId) %>% count()
answer<-merge(dataframe4, dataframe3, by='radiologyOccurrenceId')
answer<-split(answer, as.character(answer$radiologyProtocolConceptId))
answer<-lapply(answer, function(x){
answer<-split(x, as.character(x$radiologyPhaseConceptId))
answer<-sapply(answer, function(y){
sum(y$n)
})
ImageCount<-c()
for (i in 1:length(answer)){
ImageCount<-c(ImageCount, sprintf('%s : %d', names(answer)[i], answer[i]))
}
ImageCount<-paste(ImageCount, collapse = ' / ')
return(data.frame(radiologyProtocolConceptId=unique(x$radiologyProtocolConceptId), Count_of_Images_of_each_Phase=ImageCount))
})
answer<-do.call(rbind, answer)
answer<-merge(answerDF, answer, by='radiologyProtocolConceptId')
my_vals = answer$radiologyProtocolConceptId
my_colors = ifelse(my_vals %in% input$Pro,'orange','white')
datatable(answer) %>% formatStyle(
'radiologyProtocolConceptId',
target = 'row',
backgroundColor = styleEqual(my_vals, my_colors)
)
}
})
output$txt <-renderText({
RadiologyPlaybook<-unique(LoincRsnaRadiologyPlaybook[,c(2,3)])
RadiologyPlaybook<-data.frame(RadiologyPlaybook, row.names = NULL)
answer1<-Radiology_Image_Table[, c('radiologyOccurrenceId', 'radiologyPhaseConceptId', 'dicomPath')]
answer1<-answer1 %>% filter(radiologyOccurrenceId %in% Radiology_Occurrence_Table$radiologyOccurrenceId)
answer2<-RadiologyPlaybook %>% filter(radiologyProtocolConceptId %in% Radiology_Occurrence_Table$radiologyProtocolConceptId)
answer2<-merge(Radiology_Occurrence_Table[, c('radiologyOccurrenceId', 'studyDateTime', 'modality', 'radiologyProtocolConceptId')], answer2, by='radiologyProtocolConceptId')
answer<-merge(answer1, answer2, by='radiologyOccurrenceId')
answer<-answer %>% filter(as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime) & grepl(paste(input$mod, collapse ='|'), LongCommonName)==T & radiologyProtocolConceptId %in% input$Pro & radiologyPhaseConceptId %in% input$Pha)
Counts<-nrow(answer)
sprintf("You have selected %d images! Now extract list of DICOM files by clicking Export button.", Counts)
})
output$downloadData <- downloadHandler(
filename = function() {
paste('List_of_DICOM_files.csv')
},
content = function(file) {
RadiologyPlaybook<-unique(LoincRsnaRadiologyPlaybook[,c(2,3)])
RadiologyPlaybook<-data.frame(RadiologyPlaybook, row.names = NULL)
answer1<-Radiology_Image_Table[, c('radiologyOccurrenceId', 'radiologyPhaseConceptId', 'dicomPath')]
answer1<-answer1 %>% filter(radiologyOccurrenceId %in% Radiology_Occurrence_Table$radiologyOccurrenceId)
answer2<-RadiologyPlaybook %>% filter(radiologyProtocolConceptId %in% Radiology_Occurrence_Table$radiologyProtocolConceptId)
answer2<-merge(Radiology_Occurrence_Table[, c('radiologyOccurrenceId', 'studyDateTime', 'modality', 'radiologyProtocolConceptId')], answer2, by='radiologyProtocolConceptId')
answer<-merge(answer1, answer2, by='radiologyOccurrenceId')
answer<-answer %>% filter(as.Date(studyDateTime) >= min(input$dateTime) & as.Date(studyDateTime) <= max(input$dateTime) & grepl(paste(input$mod, collapse ='|'), LongCommonName)==T & radiologyProtocolConceptId %in% input$Pro & radiologyPhaseConceptId %in% input$Pha)
write.csv(data.frame(answer$dicomPath, row.names = NULL), file)
})
}
shinyApp(ui=ui,server=server)
}
