############################################
##  Multigroup Difference Analysis (ANOVA/KW) - UI
##  对齐单独检验的UI设计：简洁+严格校验+统一风格
############################################
## 1. 翻译标签（对齐单独检验的i18n逻辑，保持术语一致）
mda_norm_type <- c("overall", "by_group")
names(mda_norm_type) <- c(i18n$t("Overall (Whole variable)"),
                          i18n$t("By Group (Each level separately)"))

mda_plots <- c("norm_qq", "norm_hist", "homo_box")
names(mda_plots) <- c(i18n$t("Normality: Q-Q Plot"),
                      i18n$t("Normality: Histogram"),
                      i18n$t("Homogeneity: Boxplot by Group"))

mda_test_methods <- c("anova", "kw")
names(mda_test_methods) <- c(i18n$t("Analysis of Variance (ANOVA)"),
                             i18n$t("Kruskal-Wallis Test"))

## 2. 函数形参
mda_args <- as.list(formals(mda))
mda_args <- mda_args[names(mda_args) %in% c("dataset", "var", "group", "normality_type","test_method")]

## 3. 输入收集
mda_inputs <- reactive({
  req(input$dataset)
  # 基础参数
  inputs <- list(
    dataset = input$dataset,
    var = input$mda_var,
    group = input$mda_group,
    normality_type = input$mda_normality_type,
    test_method = input$mda_test_method,
    envir = r_data
  )
  # 校验参数完整性
  for (arg in names(mda_args)) {
    if (is.null(inputs[[arg]]) || inputs[[arg]] == "") inputs[[arg]] <- mda_args[[arg]]
  }
  inputs
})

## 4. 因变量选择
output$ui_mda_var <- renderUI({
  req(input$dataset)
  current_data <- get_data(input$dataset, envir = r_data)
  is_num <- sapply(current_data, function(col) is.numeric(col) || is.ts(col))
  num_vars <- names(is_num)[is_num]
  
  if (length(num_vars) == 0) {
    return(div(class = "alert alert-warning", i18n$t("No numeric variables in dataset. Please select another dataset.")))
  }
  
  # 提取变量类型并组合标签
  var_types <- sapply(current_data[, num_vars, drop = FALSE], function(col) class(col)[1])
  choices <- setNames(nm = paste0(num_vars, " {", var_types, "}"), object = num_vars)
  
  selectInput(
    inputId = "mda_var",
    label = i18n$t("Dependent variable:"),
    choices = c("None" = "", choices),
    selected = state_single("mda_var", num_vars),
    multiple = FALSE
  )
})

## 5. 分组变量选择
output$ui_mda_group <- renderUI({
  req(input$dataset)
  current_data <- get_data(input$dataset, envir = r_data)
  is_group <- sapply(current_data, function(col) is.factor(col) || is.character(col))
  group_candidates <- names(is_group)[is_group]
  
  valid_groups <- character(0)
  for (grp in group_candidates) {
    grp_vals <- current_data[[grp]]
    valid_levels <- length(unique(grp_vals[!is.na(grp_vals)]))
    if (valid_levels >= 2) {
      valid_groups <- c(valid_groups, grp)
    }
  }
  
  if (length(valid_groups) == 0) {
    return(div(class = "alert alert-warning", i18n$t("No valid grouping variables (need ≥2 levels).")))
  }
  
  #提取变量类型并组合标签
  group_types <- sapply(current_data[, valid_groups, drop = FALSE], function(col) class(col)[1])
  choices <- setNames(nm = paste0(valid_groups, " {", group_types, "}"), object = valid_groups)
  
  selectInput(
    inputId = "mda_group",
    label = i18n$t("Grouping variable:"),
    choices = choices,  
    selected = state_single("mda_group", valid_groups),
    multiple = FALSE
  )
})


## 6. 正态性检验类型选择
output$ui_mda_normality_type <- renderUI({
  selectInput(
    inputId = "mda_normality_type",
    label = i18n$t("Normality test:"),
    choices = mda_norm_type,
    selected = state_single("mda_normality_type", mda_norm_type, "overall"),
    multiple = FALSE
  )
})

## 7. 主UI
output$ui_mda <- renderUI({
  req(input$dataset)
  tagList(
    wellPanel(
      # Summary标签页
      conditionalPanel(
        condition = "input.tabs_mda == 'Summary'",
        uiOutput("ui_mda_var"),
        uiOutput("ui_mda_group"),
        uiOutput("ui_mda_normality_type"),
        radioButtons(
          inputId = "mda_test_method",
          label = i18n$t("Select test method:"),
          choices = mda_test_methods,
          selected = state_single("mda_test_method", mda_test_methods, "anova"),
          inline = FALSE
        )
      ),
      # Plot标签页
      conditionalPanel(
        condition = "input.tabs_mda == 'Plot'",
        selectizeInput(
          inputId = "mda_plots",
          label = i18n$t("Select plots:"),
          choices = mda_plots,
          selected = state_multiple("mda_plots", mda_plots, "norm_qq"),  # 默认选QQ图
          multiple = TRUE,
          options = list(
            placeholder = i18n$t("Select plot types"),
            plugins = list("remove_button", "drag_drop")  
          )
        )
      )
    ),
    # 帮助与报告
    help_and_report(
      modal_title = i18n$t("Multigroup Difference Analysis (ANOVA/KW)"),
      fun_name = "mda",
      help_file = inclMD(file.path(getOption("radiant.path.basics"),
                                   "app/tools/help/mda.md"))
    )
  )
})

## 8. 图表尺寸
mda_plot_dims <- reactive({
  req(.mda())
  plot_count <- length(input$mda_plots)
  group_count <- if (.mda()$normality_type == "by_group") {
    length(levels(.mda()$plot_obj$homo$data[[.mda()$group]]))
  } else {
    1
  }
  base_subplot_height_px <- 350
  total_height_px <- base_subplot_height_px * plot_count * group_count
  total_height_px <- min(total_height_px, 2000)
  total_height_px <- max(total_height_px, 400)
  
  list(
    width = 700,
    height = total_height_px
  )
})

mda_plot_width <- function() mda_plot_dims()$width
mda_plot_height <- function() mda_plot_dims()$height

## 9. 输出面板
output$mda <- renderUI({
  # 注册输出
  register_print_output("summary_mda", ".summary_mda")
  register_plot_output("plot_mda", ".plot_mda",
                       height_fun = "mda_plot_height")
  
  # 标签页
  mda_panels <- tabsetPanel(
    id = "tabs_mda",
    tabPanel(
      title = i18n$t("Summary"),
      value = "Summary",
      verbatimTextOutput("summary_mda", placeholder = TRUE) 
    ),
    tabPanel(
      title = i18n$t("Plot"),
      value = "Plot",
      download_link("dlp_mda"),  # 下载按钮
      plotOutput("plot_mda", height = "100%"),
      style = "margin-top: 10px;"
    )
  )
  
  # 整合到Radiant标准面板
  stat_tab_panel(
    menu = i18n$t("Basics > Means"),
    tool = i18n$t("Multigroup Difference Analysis(ANOVA/Kruskal-Wallis)"),
    tool_ui = "ui_mda",
    output_panels = mda_panels
  )
})

## 10. 可用性检验
mda_available <- reactive({
  req(input$dataset)
  current_data <- get_data(input$dataset, envir = r_data)
  
  # 1. 校验因变量
  if (not_available(input$mda_var) || !input$mda_var %in% colnames(current_data)) {
    return(i18n$t("Please select a valid numeric dependent variable."))
  }
  
  # 2. 校验分组变量
  if (not_available(input$mda_group) || !input$mda_group %in% colnames(current_data)) {
    return(i18n$t("Please select a valid grouping variable."))
  }
  
  # 3. 校验分组变量水平
  group_vals <- current_data[[input$mda_group]]
  valid_levels <- length(unique(group_vals[!is.na(group_vals)]))
  if (valid_levels < 2) {
    return(i18n$t("Grouping variable has <2 valid levels. Choose another."))
  }
  
  # 4. 校验有效样本
  valid_n <- sum(!is.na(current_data[[input$mda_var]]) & !is.na(current_data[[input$mda_group]]))
  if (valid_n < 5) {  # 最小样本量校验
    return(i18n$t(paste("Valid samples are too few (n=", valid_n, "). Need at least 5.", sep = "")))
  }
  
  "available"  # 所有校验通过
})

## 11. 计算核心
.mda <- reactive({
  req(mda_available() == "available")
  do.call(mda, mda_inputs())
})

.summary_mda <- reactive({
  req(mda_available() == "available")
  summary(.mda())
})

.plot_mda <- reactive({
  req(mda_available() == "available")
  validate(need(input$mda_plots, i18n$t("Please select at least one plot type first.")))
  # 进度提示
  withProgress(message = i18n$t("Generating plots..."), value = 0.5, {
    p <- plot(.mda(), plots = input$mda_plots, shiny = TRUE)
    setProgress(value = 1)
  })
  p
})

## 12. 下载与截图
# 图表下载
download_handler(
  id = "dlp_mda",
  fun = function(file) {
    # 1. 校验图表对象
    plot_obj <- .plot_mda()
    width_in <- mda_plot_width() / 96  
    height_in <- mda_plot_height() / 96  
    
    ggsave(
      filename = file,
      plot = plot_obj,
      width = width_in,
      height = height_in,
      device = "png",
      dpi = 300,
      limitsize = FALSE,  
      bg = "white" 
    )
  },
  fn = function() paste0(input$dataset, "_mda_plots"),
  type = "png",
  caption = i18n$t("Save plots")
)

# 报告生成
mda_report <- function() {
  req(mda_available() == "available")
  figs <- length(input$mda_plots) > 0
  # 报告结构
  update_report(
    inp_main = clean_args(mda_inputs(), mda_args),
    fun_name = "mda",
    inp_out = if (figs) list("", list(plots = input$mda_plots)) else list(""),
    outputs = if (figs) c("summary", "plot") else "summary",
    figs = figs,
    fig.width = mda_plot_width(),
    fig.height = mda_plot_height()
  )
}

# 截图功能
observeEvent(input$mda_report, {
  r_info[["latest_screenshot"]] <- NULL
  mda_report()
})

observeEvent(input$mda_screenshot, {
  r_info[["latest_screenshot"]] <- NULL
  radiant_screenshot_modal("modal_mda_screenshot")
})

observeEvent(input$modal_mda_screenshot, {
  mda_report()
  removeModal()
})