如何用R语言RSelenium包获取含tablist类div标签的页面谷歌地理坐标

尝试用R的RSelenium包获取目标页面的地理坐标（目标值：20º27'36.1"S 54º38'03.1"W），三次尝试均失败，代码如下：

library(rvest)
library(RSelenium)
library(httpuv)

port <- httpuv::randomPort()

rD <- rsDriver(browser = c("firefox"),
               verbose=TRUE,
               check = FALSE,
               port = port)

driver <- rD[["client"]]

urll <- "https://www.zapimoveis.com.br/lancamento/venda-apartamento-2-quartos-bairro-seminario-campo-grande-ms-46m2-id-2600496487/"
driver$navigate(urll)

politicas <- driver$findElement(using = "css",
                                value = "button.cookie-notifier__cta")
politicas$clickElement()

botaomapa <- driver$findElement(using = "xpath", "/html/body/main/div[1]/section/section/section[1]/button[2]")
botaomapa$clickElement()

#Attempt 1: using xpath from coordinates
coord <- driver$findElement(using="xpath", "/html/body/div/div/div/div[4]/div/div/div/div/div[1]/div")#errorrrrr

#Attempt 2: by botaomapa object
coord <- botaomapa$findElement(using="xpath", "/html/body/div/div/div/div[4]/div/div/div/div/div[1]/div")#errorrrr

#Attempt 3: by rvest package
readmap <- read_html(urll)
auxiliar <- readmap %>% html_elements("section")
auxiliar2 <- auxiliar%>%html_elements("#listing-map")
c1 <- readmap%>%html_nodes(xpath="/html/body/main/div[1]/section/section/section[2]/div/div[3]/article/iframe")#nothing
c2 <- auxiliar2%>%html_nodes(xpath="/html/body/main/div[1]/section/section/section[2]/div/div[3]/article/iframe")#nothing
c3 <- auxiliar2%>%html_nodes(xpath="/html/body/div/div/div/div[4]/div/div/div/div/div[1]/div")#nothing

1. 绝对XPATH不稳定：使用/html/body/...这种绝对路径，页面结构稍有变动就会失效，且无法定位到iframe内的元素
2. 未处理iframe：点击地图按钮后，地图内容加载在iframe中，直接在主页面上下文无法定位iframe内的元素
3. rvest局限性：rvest只能抓取初始静态HTML，无法获取动态加载的iframe内容，也无法模拟点击交互

步骤如下：

• 处理cookie弹窗后，点击地图按钮打开模态框
• 切换到地图所在的iframe上下文
• 定位坐标显示元素并提取文本
• 完成后关闭驱动

修正后的代码：

library(RSelenium)
library(httpuv)

# 初始化驱动
port <- httpuv::randomPort()
rD <- rsDriver(browser = "firefox", verbose = TRUE, check = FALSE, port = port)
driver <- rD[["client"]]

# 访问目标页面
urll <- "https://www.zapimoveis.com.br/lancamento/venda-apartamento-2-quartos-bairro-seminario-campo-grande-ms-46m2-id-2600496487/"
driver$navigate(urll)

# 处理cookie弹窗
Sys.sleep(2) # 等待页面加载
politicas <- driver$findElement(using = "css", value = "button.cookie-notifier__cta")
politicas$clickElement()

# 点击地图按钮打开模态框
Sys.sleep(1)
botaomapa <- driver$findElement(using = "css", value = "button[aria-label='Ver mapa']")
botaomapa$clickElement()

# 切换到地图iframe
Sys.sleep(2) # 等待模态框和iframe加载
iframe <- driver$findElement(using = "css", value = "iframe.listing-map__frame")
driver$switchToFrame(iframe)

# 定位并提取坐标文本
coord_element <- driver$findElement(using = "css", value = "div.mapboxgl-ctrl-geocoder--input")
coordinates <- coord_element$getElementText()[[1]]
print(coordinates) # 输出目标坐标：20º27'36.1"S 54º38'03.1"W

# 清理资源
driver$close()
rD$server$stop()

• 使用相对选择器（如button[aria-label='Ver mapa']）替代绝对XPATH，提升稳定性
• 加入Sys.sleep()等待元素加载，避免因页面渲染延迟导致的定位失败
• 必须切换到iframe上下文才能访问其内部元素，这是动态页面抓取的常见要点

内容的提问来源于stack exchange，提问作者Ivan Bezerra Allaman