R Markdown无法同时编译HTML与运行Shiny代码求助

你的核心问题是**runtime: shiny的存在导致文档类型冲突**：当yaml中包含runtime: shiny时，文档是交互式Shiny文档，只能编译成带Shiny交互的HTML，无法作为普通静态HTML编译；去掉runtime: shiny后，普通HTML文档不支持Shiny的运行上下文，直接调用shinyApp()会报错。

下面提供两种针对性解决方案：

方案1：制作可编译的交互式Shiny文档（推荐，同时满足交互+HTML输出） #

保留runtime: shiny，但移除代码中的shinyApp(ui = ui, server = server)调用——因为启用runtime: shiny的R Markdown文档会自动识别并运行UI和server逻辑，无需手动启动Shiny应用。

修改后的完整代码： #

---
title: "Report"
author: "Name"
date: "`r Sys.Date()`"
output:
  html_document:
    df_print: paged
    runtime: shiny  # 保留这一行，启用Shiny交互
editor_options: 
  markdown: 
    wrap: 72
---

knitr::opts_chunk$set(echo = TRUE)
library(shiny)
library(leaflet)
library(rgdal)
library(sf)
library(dplyr)
library(RColorBrewer)
library(htmlwidgets)

Summary #

Text

Section #

Text

wave_data <- read_sf("~/path/Wave_data_32631.geojson")

wave_data <- st_transform(wave_data, crs = '+proj=longlat +datum=WGS84')

wave_data2 <- wave_data %>% 
   mutate_at(vars(An_mn_P_OD), ~ as.integer(round(.x))) 

bins <- c(0, 25, 50, 100, 150, 200, 300, 400, 500, 1000, 3100)
pal <- colorBin("RdYlBu", domain = wave_data2$Ave_Depth, bins = bins)

# 直接定义UI，无需包裹在shinyApp中
bootstrapPage(
   tags$style(type = "text/css", "html, body {width:100%;height:100%}"),
   leafletOutput("wave_data_map", width = "100%", height = "100%"),
   absolutePanel(top = 10, right = 10, 
                 sliderInput("wave_data_range", "Wave energy", 
                         min = 0, max = 75, 
                         value = c(min(wave_data2$An_mn_P_OD), max(wave_data2$An_mn_P_OD)), 
                         step = 5, 
                         round = 0.5,
                         dragRange = TRUE)
 )
)

# 直接定义server逻辑，无需手动调用shinyApp
server <- function(input, output, session) {
 wave_energy_output <- reactive({
    wave_data2 %>% 
       filter(An_mn_P_OD >= input$wave_data_range[1],
              An_mn_P_OD <= input$wave_data_range[2])
    # 移除无用的group_by(Lat)，否则会保留分组导致leaflet数据处理异常
 })

output$wave_data_map <- renderLeaflet({
   leaflet() %>% 
     addProviderTiles(providers$Esri.WorldTopoMap) %>% 
     setView(lng = -4.2026458, lat = 56.4906712, zoom = 5) %>% 
     addPolygons(data = wave_data2, 
                 weight = 1,
                 smoothFactor = 0.5,
                 color = "white", 
                 fillOpacity = 0.3, 
                 fillColor = pal(wave_data2$Ave_Depth)) %>% 
     addLegend(pal = pal, 
               values = wave_data2$Ave_Depth, 
               title = "Average depth",
               labFormat = labelFormat(suffix = "m"),
               opacity = 0.7, 
               position = "bottomright")
 })

 observeEvent(input$wave_data_range, {
  data <- wave_energy_output()
  leafletProxy("wave_data_map", data = data) %>% 
     clearShapes() %>% 
     addPolygons(data = data,
                 fillColor = pal(data$Ave_Depth),  # 这里改为data$Ave_Depth，避免依赖reactive重复调用
                 weight = 1,
                 smoothFactor = 0.5,
                 color = "white", 
                 fillOpacity = 0.5)
 })
}

Section #

Section #

编译后会生成**带交互功能的HTML文档**，既能正常显示静态文本，也能使用Shiny滑块交互地图。

---

## 方案2：制作普通静态HTML文档+可单独运行的Shiny代码

如果需要纯静态HTML文档，同时保留可单独运行的Shiny代码，只需移除yaml中的`runtime: shiny`，并将Shiny代码块设置为`eval=FALSE`（编译时不执行代码，仅显示代码内容），用户可复制代码到R环境单独运行。

### 修改后的yaml部分：
```yaml
---
title: "Report"
author: "Name"
date: "`r Sys.Date()`"
output:
  html_document:
    df_print: paged
    # 移除runtime: shiny
editor_options: 
  markdown: 
    wrap: 72
---

修改后的Shiny代码块： #

# 这里保留你原有的完整Shiny代码（包含shinyApp调用）
wave_data <- read_sf("~/path/Wave_data_32631.geojson")

wave_data <- st_transform(wave_data, crs = '+proj=longlat 
+datum=WGS84')

wave_data2 <- wave_data %>% 
   mutate_at(vars(An_mn_P_OD), ~ as.integer(round(.x))) 

bins <- c(0, 25, 50, 100, 150, 200, 300, 400, 500, 1000, 3100)
pal <- colorBin("RdYlBu", domain = wave_data2$Ave_Depth, bins = bins)

# Define UI for application

ui <- bootstrapPage(
   tags$style(type = "text/css", "html, body 
             {width:100%;height:100%}"),
   leafletOutput("wave_data_map", width = "100%", height = 
                    "100%"),
   absolutePanel(top = 10, right = 10, 
                 sliderInput("wave_data_range", "Wave energy", 
                         min = 0, max = 75, 
                         value = c(min(wave_data2$An_mn_P_OD), 
                                   max(wave_data2$An_mn_P_OD)), 
                         step = 5, 
                         round = 0.5,
                         dragRange = TRUE)
 )
 )

 # Define server logic

 server <- function(input, output, session) {
 wave_energy_output <- reactive({
    data <- wave_data2 %>% 
       filter(An_mn_P_OD >= input$wave_data_range[1],
            An_mn_P_OD <= input$wave_data_range[2])
   data %>% group_by(Lat)
 })

output$wave_data_map <- renderLeaflet({
   leaflet() %>% 
     addProviderTiles(providers$Esri.WorldTopoMap) %>% 
     setView(lng = -4.2026458, lat = 56.4906712, zoom = 5) %>% 
     addPolygons(data = wave_data2, 
                 weight = 1,
                 smoothFactor = 0.5,
                 color = "white", 
                 fillOpacity = 0.3, 
                 fillColor = pal(wave_data2$Ave_Depth)) %>% 
     addLegend(pal = pal, 
               values = wave_data2$Ave_Depth, 
               title = "Average depth",
               labFormat = labelFormat (suffix = "m"),
               opacity = 0.7, 
               position = "bottomright")
 })

 observeEvent(input$wave_data_range, {
  data <- wave_energy_output()
  leafletProxy("wave_data_map", data = data) %>% 
     clearShapes() %>% 
     addPolygons(data = data,
                 fillColor = pal(wave_energy_output()$Ave_Depth), 
                 weight = 1,
                 smoothFactor = 0.5,
                 color = "white", 
                 fillOpacity = 0.5)
 })

 }

 # Run the application 

 shinyApp(ui = ui, server = server)

编译后得到静态HTML文档，Shiny代码会以代码块形式展示，用户可复制到R中单独运行Shiny应用。

内容的提问来源于stack exchange，提问作者AlexaQuinoa