AIclinicalresearch/r-statistics-service/tools/wilcoxon.R

#' @tool_code ST_WILCOXON
#' @name Wilcoxon 符号秩检验
#' @version 1.0.0
#' @description 配对样本的非参数检验（配对 T 检验的替代方法）
#' @author SSA-Pro Team

library(glue)
library(ggplot2)
library(base64enc)

run_analysis <- function(input) {
  # ===== 初始化 =====
  logs <- c()
  log_add <- function(msg) { logs <<- c(logs, paste0("[", Sys.time(), "] ", msg)) }

  on.exit({}, add = TRUE)

  # ===== 数据加载 =====
  log_add("开始加载输入数据")
  df <- tryCatch(
    load_input_data(input),
    error = function(e) {
      log_add(paste("数据加载失败:", e$message))
      return(NULL)
    }
  )

  if (is.null(df)) {
    return(make_error(ERROR_CODES$E100_INTERNAL_ERROR, details = "数据加载失败"))
  }
  log_add(glue("数据加载成功: {nrow(df)} 行, {ncol(df)} 列"))

  p <- input$params
  before_var <- p$before_var
  after_var <- p$after_var

  # ===== 参数校验 =====
  if (!(before_var %in% names(df))) {
    return(make_error(ERROR_CODES$E001_COLUMN_NOT_FOUND, col = before_var))
  }
  if (!(after_var %in% names(df))) {
    return(make_error(ERROR_CODES$E001_COLUMN_NOT_FOUND, col = after_var))
  }

  # ===== 数据清洗 =====
  original_rows <- nrow(df)
  df <- df[!is.na(df[[before_var]]) & !is.na(df[[after_var]]), ]

  # 确保数值型
  if (!is.numeric(df[[before_var]])) {
    df[[before_var]] <- as.numeric(as.character(df[[before_var]]))
    df <- df[!is.na(df[[before_var]]), ]
  }
  if (!is.numeric(df[[after_var]])) {
    df[[after_var]] <- as.numeric(as.character(df[[after_var]]))
    df <- df[!is.na(df[[after_var]]), ]
  }

  removed_rows <- original_rows - nrow(df)
  if (removed_rows > 0) {
    log_add(glue("数据清洗: 移除 {removed_rows} 行缺失值 (剩余 {nrow(df)} 行)"))
  }

  # ===== 护栏检查 =====
  guardrail_results <- list()
  warnings_list <- c()

  sample_check <- check_sample_size(nrow(df), min_required = 5, action = ACTION_BLOCK)
  guardrail_results <- c(guardrail_results, list(sample_check))
  log_add(glue("配对样本量: N = {nrow(df)}, {sample_check$reason}"))

  guardrail_status <- run_guardrail_chain(guardrail_results)

  if (guardrail_status$status == "blocked") {
    return(list(status = "blocked", message = guardrail_status$reason, trace_log = logs))
  }

  # ===== 计算差值 =====
  diff_values <- df[[after_var]] - df[[before_var]]

  # 检查差值方差（容差比较避免浮点精度问题）
  if (isTRUE(sd(diff_values) < .Machine$double.eps^0.5)) {
    return(make_error(ERROR_CODES$E007_VARIANCE_ZERO, col = paste(after_var, "-", before_var)))
  }

  # ===== 核心计算 =====
  log_add("执行 Wilcoxon 符号秩检验")

  result <- tryCatch({
    wilcox.test(df[[before_var]], df[[after_var]], paired = TRUE, conf.int = TRUE)
  }, error = function(e) {
    log_add(paste("Wilcoxon 检验失败:", e$message))
    return(NULL)
  })

  if (is.null(result)) {
    return(map_r_error("Wilcoxon 符号秩检验计算失败"))
  }

  method_used <- result$method
  log_add(glue("V = {result$statistic}, P = {round(result$p.value, 4)}"))

  # ===== 效应量: r = Z / sqrt(N) =====
  n_pairs <- nrow(df)
  z_approx <- qnorm(result$p.value / 2)
  r_effect <- abs(z_approx) / sqrt(n_pairs)

  r_interpretation <- if (r_effect < 0.1) "微小" else if (r_effect < 0.3) "小" else if (r_effect < 0.5) "中等" else "大"

  # ===== 描述统计 =====
  before_vals <- df[[before_var]]
  after_vals <- df[[after_var]]

  desc_stats <- list(
    before = list(
      variable = before_var,
      n = length(before_vals),
      mean = round(mean(before_vals), 3),
      sd = round(sd(before_vals), 3),
      median = round(median(before_vals), 3),
      q1 = round(quantile(before_vals, 0.25), 3),
      q3 = round(quantile(before_vals, 0.75), 3)
    ),
    after = list(
      variable = after_var,
      n = length(after_vals),
      mean = round(mean(after_vals), 3),
      sd = round(sd(after_vals), 3),
      median = round(median(after_vals), 3),
      q1 = round(quantile(after_vals, 0.25), 3),
      q3 = round(quantile(after_vals, 0.75), 3)
    ),
    difference = list(
      mean = round(mean(diff_values), 3),
      sd = round(sd(diff_values), 3),
      median = round(median(diff_values), 3)
    )
  )

  output_results <- list(
    method = method_used,
    statistic_V = jsonlite::unbox(as.numeric(result$statistic)),
    p_value = jsonlite::unbox(as.numeric(result$p.value)),
    p_value_fmt = format_p_value(result$p.value),
    pseudomedian = if (!is.null(result$estimate)) jsonlite::unbox(round(as.numeric(result$estimate), 4)) else NULL,
    conf_int = if (!is.null(result$conf.int)) round(as.numeric(result$conf.int), 4) else NULL,
    effect_size = list(
      r = jsonlite::unbox(round(r_effect, 4)),
      interpretation = r_interpretation
    ),
    descriptive = desc_stats
  )

  # ===== 生成图表 =====
  log_add("生成配对变化图")
  plot_base64 <- tryCatch({
    generate_paired_plot(df, before_var, after_var, diff_values)
  }, error = function(e) {
    log_add(paste("图表生成失败:", e$message))
    NULL
  })

  # ===== 生成可复现代码 =====
  original_filename <- if (!is.null(input$original_filename) && nchar(input$original_filename) > 0) {
    input$original_filename
  } else {
    "data.csv"
  }

  reproducible_code <- glue('
# SSA-Pro 自动生成代码
# 工具: Wilcoxon 符号秩检验
# 时间: {Sys.time()}
# ================================

library(ggplot2)

# 数据准备
df <- read.csv("{original_filename}")
before_var <- "{before_var}"
after_var <- "{after_var}"

# 数据清洗
df <- df[!is.na(df[[before_var]]) & !is.na(df[[after_var]]), ]

# Wilcoxon 符号秩检验
result <- wilcox.test(df[[before_var]], df[[after_var]], paired = TRUE, conf.int = TRUE)
print(result)

# 描述统计
cat("Before: median =", median(df[[before_var]]), "\\n")
cat("After:  median =", median(df[[after_var]]), "\\n")
cat("Diff:   median =", median(df[[after_var]] - df[[before_var]]), "\\n")
')

  # ===== 构建 report_blocks =====
  blocks <- list()

  # Block 1: 描述统计
  desc_kv <- list()
  desc_kv[["配对样本量"]] <- as.character(n_pairs)
  desc_kv[[paste0(before_var, " Median [Q1, Q3]")]] <- as.character(glue("{desc_stats$before$median} [{desc_stats$before$q1}, {desc_stats$before$q3}]"))
  desc_kv[[paste0(after_var, " Median [Q1, Q3]")]] <- as.character(glue("{desc_stats$after$median} [{desc_stats$after$q1}, {desc_stats$after$q3}]"))
  desc_kv[["差值 Median"]] <- as.character(desc_stats$difference$median)
  blocks[[length(blocks) + 1]] <- make_kv_block(desc_kv, title = "样本概况")

  # Block 2: 检验结果
  kv_result <- list(
    "方法" = method_used,
    "V 统计量" = as.character(round(as.numeric(result$statistic), 1)),
    "P 值" = output_results$p_value_fmt,
    "效应量 r" = as.character(output_results$effect_size$r),
    "效应量解释" = r_interpretation
  )
  if (!is.null(output_results$pseudomedian)) {
    kv_result[["伪中位数"]] <- as.character(output_results$pseudomedian)
  }
  if (!is.null(output_results$conf_int)) {
    kv_result[["95% 置信区间"]] <- sprintf("[%.4f, %.4f]", output_results$conf_int[1], output_results$conf_int[2])
  }
  blocks[[length(blocks) + 1]] <- make_kv_block(kv_result, title = "Wilcoxon 符号秩检验结果")

  # Block 3: 图表
  if (!is.null(plot_base64)) {
    blocks[[length(blocks) + 1]] <- make_image_block(plot_base64,
      title = paste("配对变化:", before_var, "→", after_var),
      alt = "配对样本前后变化图")
  }

  # Block 4: 结论摘要
  sig_text <- if (result$p.value < 0.05) "差异具有统计学意义" else "差异无统计学意义"
  direction <- if (mean(diff_values) > 0) "升高" else "降低"
  conclusion <- glue(
    "Wilcoxon 符号秩检验结果：V = {round(as.numeric(result$statistic), 1)}，P {output_results$p_value_fmt}。",
    "配对样本从 **{before_var}** 到 **{after_var}** 的变化{sig_text}",
    "（中位数{direction} {abs(desc_stats$difference$median)}，效应量 r = {output_results$effect_size$r}，{r_interpretation}效应）。"
  )
  blocks[[length(blocks) + 1]] <- make_markdown_block(conclusion, title = "结论摘要")

  # ===== 返回结果 =====
  log_add("分析完成")

  return(list(
    status = "success",
    message = "分析完成",
    warnings = if (length(warnings_list) > 0) warnings_list else NULL,
    results = output_results,
    report_blocks = blocks,
    plots = if (!is.null(plot_base64)) list(plot_base64) else list(),
    trace_log = logs,
    reproducible_code = as.character(reproducible_code)
  ))
}

# 辅助函数：配对变化图（差值直方图 + 配对连线图）
generate_paired_plot <- function(df, before_var, after_var, diff_values) {
  # 配对连线图
  n <- nrow(df)
  plot_df <- data.frame(
    id = rep(1:n, 2),
    time = rep(c("Before", "After"), each = n),
    value = c(df[[before_var]], df[[after_var]])
  )
  plot_df$time <- factor(plot_df$time, levels = c("Before", "After"))

  p <- ggplot(plot_df, aes(x = time, y = value)) +
    geom_line(aes(group = id), alpha = 0.3, color = "gray60") +
    geom_point(aes(color = time), size = 2, alpha = 0.6) +
    stat_summary(fun = median, geom = "point", shape = 18, size = 5, color = "red") +
    stat_summary(fun = median, geom = "line", aes(group = 1), color = "red", linewidth = 1.2) +
    theme_minimal() +
    labs(
      title = paste("Paired Change:", before_var, "→", after_var),
      x = "",
      y = "Value"
    ) +
    scale_color_manual(values = c("Before" = "#3b82f6", "After" = "#ef4444")) +
    theme(legend.position = "none")

  tmp_file <- tempfile(fileext = ".png")
  ggsave(tmp_file, p, width = 6, height = 5, dpi = 100)
  base64_str <- base64encode(tmp_file)
  unlink(tmp_file)

  return(paste0("data:image/png;base64,", base64_str))
}