###############################
## Time Series Analysis - ui
## file: time_series_ui.R
###############################

# --- 1. 定义下拉菜单和选项常量 ---
ts_methods <- c("ma", "ets", "acf")
names(ts_methods) <- c(
  i18n$t("Moving Average"),
  i18n$t("Exp. Smoothing"),
  i18n$t("Autocorrelation")
)

# 绘图图层选项（统一用方法1）
ts_plot_opts <- c("original", "ma", "ets", "acf")
names(ts_plot_opts) <- c(
  i18n$t("Original Data"),
  i18n$t("Moving Average Line"),
  i18n$t("Exp. Smoothing Line"),
  i18n$t("ACF Plot")
)

# 聚合粒度选项（统一用方法1）
ts_agg_levels <- c("day", "week", "month", "quarter", "year")
names(ts_agg_levels) <- c(
  i18n$t("Day"), 
  i18n$t("Week"), 
  i18n$t("Month"), 
  i18n$t("Quarter"), 
  i18n$t("Year")
)

# 聚合函数选项（统一用方法1）
ts_agg_funs <- c("mean", "sum", "count")
names(ts_agg_funs) <- c(
  i18n$t("Mean (Average)"), 
  i18n$t("Sum (Total)"), 
  i18n$t("Count (Rows)")
)

# --- 2. 初始化参数 ---
# 读取后端 time_series 函数的默认参数
ts_args <- as.list(formals(time_series))

# --- 3. 打包用户输入 (ts_inputs) ---
ts_inputs <- reactive({
  # 处理数据过滤
  ts_args$data_filter <- if (input$show_filter) input$data_filter else ""
  ts_args$dataset <- input$dataset
  ts_args$envir <- r_data
  
  # 自动循环匹配 input$ts_xxx 到参数 xxx
  for (i in r_drop(names(ts_args))) {
    ts_args[[i]] <- input[[paste0("ts_", i)]]
  }
  ts_args
})

# --- 4. 动态 UI 组件定义 ---

# Y轴变量 (数值型)
output$ui_ts_var_num <- renderUI({
  isNum <- .get_class() %in% c("integer", "numeric", "ts")
  vars <- c("None" = "", varnames()[isNum])
  selectInput(
    inputId = "ts_var_num", label = i18n$t("Numeric variable:"),
    choices = vars, selected = state_single("ts_var_num", vars), multiple = FALSE
  )
})

# X轴变量 (日期/时间型/数值型)
output$ui_ts_var_date <- renderUI({
  all_vars <- varnames()
  # 自动检测日期类型的变量优先显示
  isDate <- .get_class() %in% c("Date", "POSIXt", "POSIXct")
  # 如果有日期变量，默认选中第一个
  def_sel <- if (any(isDate)) all_vars[isDate][1] else ""
  
  vars <- c("None" = "", all_vars)
  selectInput(
    inputId = "ts_var_date", label = i18n$t("Time variable:"),
    choices = vars, selected = state_single("ts_var_date", vars, def_sel), multiple = FALSE
  )
})

# --- 5. 侧边栏主面板 (UI Sidebar) ---
output$ui_time_series <- renderUI({
  req(input$dataset)
  tagList(
    wellPanel(
      # === 基础数据选择 ===
      uiOutput("ui_ts_var_num"),
      uiOutput("ui_ts_var_date"),
      
      # 改为垂直排列，确保下拉菜单可以正常弹出并被选择
      conditionalPanel(
        condition = "input.ts_var_date != ''",
        selectInput(
          inputId = "ts_agg_level", label = i18n$t("Interval:"), 
          choices = ts_agg_levels, 
          selected = state_single("ts_agg_level", ts_agg_levels, "month"),
          multiple = FALSE
        ),
        
        selectInput(
          inputId = "ts_agg_fun", label = i18n$t("Aggregation Method:"), 
          choices = ts_agg_funs, 
          selected = state_single("ts_agg_fun", ts_agg_funs, "mean"),
          multiple = FALSE
        )
      ),
      
      conditionalPanel(
        condition = "input.tabs_time_series == i18n$t('Summary')",
        h6("计算方法:"),
        
        # 选择计算方法
        checkboxGroupInput(
          inputId = "ts_methods", label = NULL,
          choices = ts_methods,
          selected = state_multiple("ts_methods", ts_methods, c("ma", "ets"))
        ),
        
        # MA 参数设置
        conditionalPanel(
          condition = "input.ts_methods.includes('ma')",
          numericInput("ts_ma_n", i18n$t("MA Window Size:"), 
                       value = state_init("ts_ma_n", 3), min = 2, step = 1)
        ),
        
        # 预测参数设置
        numericInput(
          inputId = "ts_fc_periods", label = i18n$t("Forecast Periods:"),
          value = state_init("ts_fc_periods", 0), min = 0, step = 1
        ),
      ),
      
      # === 绘图图层配置 ===
      conditionalPanel(
        condition = "input.tabs_time_series == i18n$t('Plot')",
        h6("图像:"),
        selectizeInput(
          inputId = "ts_plots", label = NULL,
          choices = ts_plot_opts,
          selected = state_multiple("ts_plots", ts_plot_opts, c("original", "ma", "ets")),
          multiple = TRUE,
          options = list(placeholder = "Select layers", plugins = list("remove_button", "drag_drop"))
        )
      )
    ),
    
    # 帮助按钮和报告生成按钮
    help_and_report(
      modal_title = "Time Series Analysis",
      fun_name = "time_series",
      help_file = inclMD(file.path(getOption("radiant.path.basics"), "app/tools/help/time_series.md"))
    )
  )
})

# --- 6. 定义图形高度 ---
ts_plot_height <- function() {
  # 如果包含 ACF 图，增加高度
  if ("acf" %in% input$ts_plots) 600 else 400
}

# --- 7. 注册主界面输出 (Main Panel) ---
output$time_series <- renderUI({
  # 注册打印和绘图输出
  register_print_output("summary_time_series", ".summary_time_series")
  register_plot_output("plot_time_series", ".plot_time_series", height_fun = "ts_plot_height")
  
  # 定义标签页
  ts_output_panels <- tabsetPanel(
    id = "tabs_time_series",
    tabPanel(i18n$t("Summary"), verbatimTextOutput("summary_time_series")),
    tabPanel(
      i18n$t("Plot"),
      download_link("dlp_time_series"),
      plotOutput("plot_time_series", height = "100%")
    )
  )
  
  # 构建 Radiant 标准 Tab 面板
  stat_tab_panel(
    menu = i18n$t("Basics > Time Series"),
    tool = i18n$t("Time Series Analysis"),
    tool_ui = "ui_time_series",
    output_panels = ts_output_panels
  )
})

# --- 8. 后端计算逻辑连接 ---

# 执行计算
.time_series <- reactive({
  tsi <- ts_inputs()
  tsi$envir <- r_data # 确保环境传递正确
  req(tsi$var_num)    # 必须选择了变量才执行
  do.call(time_series, tsi)
})

# 生成 Summary
.summary_time_series <- reactive({
  summary(.time_series())
})

# 生成 Plot
.plot_time_series <- reactive({
  validate(
    need(input$ts_plots, i18n$t("Please select at least one plot layer."))
  )
  withProgress(message = "Generating Plots...", value = 1, {
    plot(.time_series(), plots = input$ts_plots, shiny = TRUE)
  })
})

# --- 9. 报告生成 (Rmarkdown) ---
time_series_report <- function() {
  if (is.empty(input$ts_var_num)) return(invisible())
  
  # 判断是否生成图片代码
  if (length(input$ts_plots) == 0) {
    figs <- FALSE
    outputs <- c("summary")
    inp_out <- list("", "")
  } else {
    figs <- TRUE
    outputs <- c("summary", "plot")
    inp_out <- list("", list(plots = input$ts_plots, custom = FALSE))
  }
  
  update_report(
    inp_main = clean_args(ts_inputs(), ts_args),
    fun_name = "time_series",
    outputs = outputs,
    figs = figs,
    fig.width = 650,
    fig.height = ts_plot_height()
  )
}

# --- 10. 下载处理器 ---
download_handler(
  id = "dlp_time_series",
  fun = download_handler_plot,
  fn = function() paste0(input$dataset, "_time_series"),
  type = "png",
  caption = "Save time series plot",
  plot = .plot_time_series,
  width = 650,
  height = ts_plot_height
)

# --- 11. 事件监听 ---
observeEvent(input$time_series_report, {
  r_info[["latest_screenshot"]] <- NULL
  time_series_report()
})

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

observeEvent(input$modal_time_series_screenshot, {
  time_series_report()
  removeModal()
})