EDA_prototype(NOT USED).R

library(shiny)
library(shinydashboard)
library(shinydashboardPlus)
library(dplyr)
library(reactable)
library(GGally)
library(ggplot2)
inputData <- mtcars
shinyApp(
  ui = dashboardPage(
    dashboardHeader(),
    dashboardSidebar(),
    dashboardBody(
      shinydashboardPlus::box(
        title = "Description",
        status = "purple",
        solidHeader = TRUE,
        width = 12,
        fluidRow(
          column(
            width = 6,
            uiOutput('dataDimension')
          ),
          column(
            width = 6,
            uiOutput("missingData")
          )
        )
      ),
      shinydashboardPlus::box(
        title = "Variables",
        status = "purple",
        solidHeader = TRUE,
        width = 12,
        reactableOutput("reactOutput")
      ),
      shinydashboardPlus::box(
        title = "Correlation",
        status = "success",
        solidHeader = TRUE,
        width = 12,
        plotOutput("corplot")
      ),

      selectInput(
        inputId = "variableSelect",
          label = "variable",
          choices = colnames(inputData),
          selected = NULL
      ),
      shinydashboardPlus::box(
        title = "distribution",
        status = "success",
        solidHeader = TRUE,
        width = 12,
        plotOutput("distplot"),
        textOutput("normal"),
        htmlOutput("stats"),
        textOutput("outlier")
      ),
      shinydashboardPlus::box(
        title = "Correlation2",
        status = "success",
        solidHeader = TRUE,
        width = 12,
        selectInput(
          inputId = "var1",
          label = "variable 1",
          choices = c("", colnames(inputData)),
          selected = NULL
        ),
        selectInput(
          inputId = "var2",
          label = "variable 2",
          choices = c("", colnames(inputData)),
          selected = NULL
        ),
        plotOutput("corplot2")
      ),
    ),
    title = "EDA Module"
  ),
  server = function(input, output) {
    inputData <- mtcars
    obj <- board::brief(inputData)
    obj$unif <- ifelse(obj$unif, "True", NA)
    obj$uniq <- ifelse(obj$unif, "True", NA)

    output$corplot <- renderPlot(ggcorr(obj$cors))

    res <- data.frame(
      Name = obj$names,
      Cardinality = obj$cards,
      Zero = obj$zeros,
      Missing = obj$miss,
      isUniform = obj$unif,
      isUnique = obj$uniq
    )

    # updateSelectInput(
    #   inputId = "variableSelect",
    #   label = "variable",
    #   choices = colnames(inputData),
    #   selected = NULL
    # )

    output$dataDimension <- renderUI(
      descriptionBlock(
        header = paste0(obj$desc$nrow, " X ", obj$desc$ncol),
        numberIcon = icon("expand"),
        number = "Data Dimension",
        marginBottom = FALSE
      )
    )

    output$missingData <- renderUI(
      descriptionBlock(
        header = paste0(obj$desc$missingCellCount, "(", obj$desc$missingCellRatio, "%)"),
        numberIcon = icon("question"),
        number = "Missing Data",
        marginBottom = FALSE
      )
    )

    output$reactOutput <- renderReactable(
      reactable(
        res
      )
    )

    # updateSelectInput(
    #   inputId = "var1",
    #   label = "variable 1",
    #   choices = c("", colnames(inputData)),
    #   selected = NULL
    # )
    #
    # updateSelectInput(
    #   inputId = "var2",
    #   label = "variable 2",
    #   choices = c("", colnames(inputData)),
    #   selected = NULL
    # )

    observeEvent(input$var2, {
      if (input$var2 == "") {
        return(0)
      }
      output$corplot2 <- renderPlot(board::relation(inputData[, input$var1], inputData[, input$var2]))
    })

    observeEvent(input$variableSelect, {
      if(input$variableSelect == ''){
        return(0)
      }
      if (is.numeric(inputData[, input$variableSelect])) {
        output$distplot <- renderPlot(board::distribute(inputData[, input$variableSelect], name = input$variableSelect))

        ss <- board::describe(inputData[, input$variableSelect])
        ss <- paste0(
          "total count: ", ss$count,
          " mean : ", ss$m,
          " sd : ", ss$s,
          " min: ", ss$q0,
          "\n25%: ", ss$q1,
          " median: ", ss$q2,
          " 75%: ", ss$q3,
          " max: ", ss$q4
        )
        output$stats <- renderUI(HTML(ss))

        oo <- board::outlier(inputData[, input$variableSelect])
        oo <- paste0(
          ifelse(oo$lo > 0, paste0("value larger than ", oo$ov, " may consider as Outlier (", oo$lo, " exists)"), "large outlier not exists"),
          "\nAnd\n",
          ifelse(oo$lu > 0, paste0("value smaller than ", oo$uv, " may consider as Outlier (", oo$lu, " exists)"), "small outlier not exists")
        )

        output$outlier <- renderText(oo)

        if (board::isNormal(inputData[, input$variableSelect])) { # normality)
          output$normal <- renderText("Variable may normally distributed")
        } else {
          output$normal <- renderText("Variable may not normally distributed")
        }
      } else {
        output$distplot <- renderPlot(board::ggpie(inputData[, input$variableSelect]))
      }
    })
  }
)

# dominate -> category
# isNormal -> Numeric
# ditribute -> vis
# outlier -> numeric
# describe -> numeric