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

#' @tool_code ST_LOGISTIC_BINARY
#' @name 二元 Logistic 回归
#' @version 1.0.0
#' @description 二分类结局变量的多因素分析
#' @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$params
  outcome_var <- p$outcome_var
  predictors <- p$predictors  # 预测变量列表
  confounders <- p$confounders  # 混杂因素（可选）
  
  # ===== 参数校验 =====
  if (!(outcome_var %in% names(df))) {
    return(make_error(ERROR_CODES$E001_COLUMN_NOT_FOUND, col = outcome_var))
  }
  
  all_vars <- c(predictors, confounders)
  all_vars <- all_vars[!is.null(all_vars) & all_vars != ""]
  
  for (v in all_vars) {
    if (!(v %in% names(df))) {
      return(make_error(ERROR_CODES$E001_COLUMN_NOT_FOUND, col = v))
    }
  }
  
  if (length(predictors) == 0) {
    return(make_error(ERROR_CODES$E100_INTERNAL_ERROR, details = "至少需要一个预测变量"))
  }
  
  # ===== 数据清洗 =====
  original_rows <- nrow(df)
  
  # 移除所有相关变量的缺失值
  vars_to_check <- c(outcome_var, all_vars)
  for (v in vars_to_check) {
    df <- df[!is.na(df[[v]]), ]
  }
  
  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)"))
  }
  
  # ===== 结局变量检查 =====
  outcome_values <- unique(df[[outcome_var]])
  if (length(outcome_values) != 2) {
    return(make_error(ERROR_CODES$E003_INSUFFICIENT_GROUPS,
                      col = outcome_var, expected = 2, actual = length(outcome_values)))
  }
  
  # 确保结局变量是 0/1 或因子
  if (!is.factor(df[[outcome_var]])) {
    df[[outcome_var]] <- as.factor(df[[outcome_var]])
  }
  
  # 事件数统计
  event_counts <- table(df[[outcome_var]])
  n_events <- min(event_counts)
  n_predictors <- length(all_vars)
  
  log_add(glue("结局变量分布: {paste(names(event_counts), '=', event_counts, collapse=', ')}"))
  log_add(glue("事件数: {n_events}, 预测变量数: {n_predictors}"))
  
  # ===== 护栏检查 =====
  guardrail_results <- list()
  warnings_list <- c()
  
  # EPV 规则检查（Events Per Variable >= 10）
  epv <- n_events / n_predictors
  if (epv < 10) {
    warnings_list <- c(warnings_list, glue("EPV = {round(epv, 1)} < 10，模型可能不稳定"))
    log_add(glue("警告: EPV = {round(epv, 1)} < 10"))
  }
  
  # 样本量检查
  sample_check <- check_sample_size(nrow(df), min_required = 20, action = ACTION_BLOCK)
  guardrail_results <- c(guardrail_results, list(sample_check))
  
  guardrail_status <- run_guardrail_chain(guardrail_results)
  
  if (guardrail_status$status == "blocked") {
    return(list(
      status = "blocked",
      message = guardrail_status$reason,
      trace_log = logs
    ))
  }
  
  # ===== 构建模型公式 =====
  formula_str <- paste(outcome_var, "~", paste(all_vars, collapse = " + "))
  formula_obj <- as.formula(formula_str)
  log_add(glue("模型公式: {formula_str}"))
  
  # ===== 核心计算 =====
  log_add("拟合 Logistic 回归模型")
  
  model <- tryCatch({
    glm(formula_obj, data = df, family = binomial(link = "logit"))
  }, error = function(e) {
    log_add(paste("模型拟合失败:", e$message))
    return(NULL)
  }, warning = function(w) {
    warnings_list <<- c(warnings_list, w$message)
    log_add(paste("模型警告:", w$message))
    invokeRestart("muffleWarning")
  })
  
  if (is.null(model)) {
    return(map_r_error("模型拟合失败"))
  }
  
  # 检查模型收敛
  if (!model$converged) {
    warnings_list <- c(warnings_list, "模型未完全收敛")
    log_add("警告: 模型未完全收敛")
  }
  
  # ===== 提取模型结果 =====
  coef_summary <- summary(model)$coefficients
  
  # 计算 OR 和 95% CI
  coef_table <- data.frame(
    variable = rownames(coef_summary),
    estimate = coef_summary[, "Estimate"],
    std_error = coef_summary[, "Std. Error"],
    z_value = coef_summary[, "z value"],
    p_value = coef_summary[, "Pr(>|z|)"],
    stringsAsFactors = FALSE
  )
  
  coef_table$OR <- exp(coef_table$estimate)
  coef_table$ci_lower <- exp(coef_table$estimate - 1.96 * coef_table$std_error)
  coef_table$ci_upper <- exp(coef_table$estimate + 1.96 * coef_table$std_error)
  
  # 转换为列表格式（精简，不含原始系数）
  coefficients_list <- lapply(1:nrow(coef_table), function(i) {
    row <- coef_table[i, ]
    list(
      variable = row$variable,
      OR = round(row$OR, 3),
      ci_lower = round(row$ci_lower, 3),
      ci_upper = round(row$ci_upper, 3),
      p_value = round(row$p_value, 4),
      p_value_fmt = format_p_value(row$p_value),
      significant = row$p_value < 0.05
    )
  })
  
  # ===== 模型拟合度 =====
  null_deviance <- model$null.deviance
  residual_deviance <- model$deviance
  aic <- AIC(model)
  
  # Nagelkerke R²（伪 R²）
  n <- nrow(df)
  r2_nagelkerke <- (1 - exp((residual_deviance - null_deviance) / n)) / (1 - exp(-null_deviance / n))
  
  log_add(glue("AIC = {round(aic, 2)}, Nagelkerke R² = {round(r2_nagelkerke, 3)}"))
  
  # ===== 共线性检测（VIF） =====
  vif_results <- NULL
  if (length(all_vars) > 1) {
    tryCatch({
      if (requireNamespace("car", quietly = TRUE)) {
        vif_values <- car::vif(model)
        if (is.matrix(vif_values)) {
          vif_values <- vif_values[, "GVIF"]
        }
        vif_results <- lapply(names(vif_values), function(v) {
          list(variable = v, vif = round(vif_values[v], 2))
        })
        
        high_vif <- names(vif_values)[vif_values > 5]
        if (length(high_vif) > 0) {
          warnings_list <- c(warnings_list, paste("VIF > 5 的变量:", paste(high_vif, collapse = ", ")))
        }
      }
    }, error = function(e) {
      log_add(paste("VIF 计算失败:", e$message))
    })
  }
  
  # ===== 生成图表（森林图） =====
  log_add("生成森林图")
  plot_base64 <- tryCatch({
    generate_forest_plot(coef_table)
  }, 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: 二元 Logistic 回归
# 时间: {Sys.time()}
# ================================

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

# 模型拟合
model <- glm({formula_str}, data = df, family = binomial(link = "logit"))
summary(model)

# OR 和 95% CI
coef_summary <- summary(model)$coefficients
OR <- exp(coef_summary[, "Estimate"])
CI_lower <- exp(coef_summary[, "Estimate"] - 1.96 * coef_summary[, "Std. Error"])
CI_upper <- exp(coef_summary[, "Estimate"] + 1.96 * coef_summary[, "Std. Error"])
results <- data.frame(OR = OR, CI_lower = CI_lower, CI_upper = CI_upper, 
                      p_value = coef_summary[, "Pr(>|z|)"])
print(round(results, 3))

# 模型拟合度
cat("AIC:", AIC(model), "\\n")

# VIF（需要 car 包）
# library(car)
# vif(model)
')
  
  # ===== 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 = list(
      method = "Binary Logistic Regression (glm, binomial)",
      formula = formula_str,
      n_observations = nrow(df),
      n_predictors = n_predictors,
      coefficients = coefficients_list,
      model_fit = list(
        aic = jsonlite::unbox(round(aic, 2)),
        null_deviance = jsonlite::unbox(round(null_deviance, 2)),
        residual_deviance = jsonlite::unbox(round(residual_deviance, 2)),
        r2_nagelkerke = jsonlite::unbox(round(r2_nagelkerke, 4))
      ),
      vif = vif_results,
      epv = jsonlite::unbox(round(epv, 1))
    ),
    plots = if (!is.null(plot_base64)) list(plot_base64) else list(),
    trace_log = logs,
    reproducible_code = as.character(reproducible_code)
  ))
}

# 辅助函数：生成森林图
generate_forest_plot <- function(coef_table) {
  # 移除截距项
  plot_data <- coef_table[coef_table$variable != "(Intercept)", ]
  
  if (nrow(plot_data) == 0) {
    return(NULL)
  }
  
  plot_data$variable <- factor(plot_data$variable, levels = rev(plot_data$variable))
  
  p <- ggplot(plot_data, aes(x = OR, y = variable)) +
    geom_vline(xintercept = 1, linetype = "dashed", color = "gray50") +
    geom_point(size = 3, color = "#3b82f6") +
    geom_errorbarh(aes(xmin = ci_lower, xmax = ci_upper), height = 0.2, color = "#3b82f6") +
    scale_x_log10() +
    theme_minimal() +
    labs(
      title = "Forest Plot: Odds Ratios with 95% CI",
      x = "Odds Ratio (log scale)",
      y = "Variable"
    ) +
    theme(
      panel.grid.minor = element_blank(),
      axis.text.y = element_text(size = 10)
    )
  
  tmp_file <- tempfile(fileext = ".png")
  ggsave(tmp_file, p, width = 8, height = max(4, nrow(plot_data) * 0.5 + 2), dpi = 100)
  base64_str <- base64encode(tmp_file)
  unlink(tmp_file)
  
  return(paste0("data:image/png;base64,", base64_str))
}