Question

I am interested in using selenium with R. I note that the various documentation is described here WebDriver (Selenium 2) API documentation. Has there been any work done on an implementation with R. How would I go about approaching this. In the documentation it notes about running a selenium server and one can query the api using Javascript. Any help would be much appreciated.

Answer 1

Selenium can be accessed using the JsonWireProtocol.

Firstly start up a Selenium server from the command line via:

java -jar selenium-server-standalone-2.25.0.jar

a new Firefox browser can be opened as follows:

library(RCurl)
library(RJSONIO)
library(XML)

baseURL<-"http://localhost:4444/wd/hub/"
server<-list(desiredCapabilities=list(browserName='firefox',javascriptEnabled=TRUE))

getURL(paste0(baseURL,"session"),
       customrequest="POST",
       httpheader=c('Content-Type'='application/json;charset=UTF-8'),
       postfields=toJSON(server))

serverDetails<-fromJSON(rawToChar(getURLContent('http://localhost:4444/wd/hub/sessions',binary=TRUE)))
serverId<-serverDetails$value[[1]]$id

Navigate to google.

getURL(paste0(baseURL,"session/",serverId,"/url"),
       customrequest="POST",
       httpheader=c('Content-Type'='application/json;charset=UTF-8'),
       postfields=toJSON(list(url="http://www.google.com")))

get the id of the search box

elementDetails<-fromJSON(rawToChar(getURLContent(paste0(baseURL,"session/",serverId,"/element"),
       customrequest="POST",
       httpheader=c('Content-Type'='application/json;charset=UTF-8'),
       postfields=toJSON(list(using="xpath",value="//*[@id=\"gbqfq\"]")),binary=TRUE))
       )

elementId<-elementDetails$value

search for a subject matter

rawToChar(getURLContent(paste0(baseURL,"session/",serverId,"/element/",elementId,"/value"),
       customrequest="POST",
       httpheader=c('Content-Type'='application/json;charset=UTF-8'),
       postfields=toJSON(list(value=list("\uE009","a","\uE009",'\b','Selenium api in R')))
       ,binary=TRUE))

return the search html

googData<-fromJSON(rawToChar(getURLContent(paste0(baseURL,"session/",serverId,"/source"),
       customrequest="GET",
       httpheader=c('Content-Type'='application/json;charset=UTF-8'),
       binary=TRUE
       ))
       )

get the suggested links

gxml<-htmlParse(googData$value)
urls<-unname(xpathSApply(gxml,"//*[@class='l']/@href"))

close the session

getURL(paste0(baseURL,"session/",serverId),
       customrequest="DELETE",
       httpheader=c('Content-Type'='application/json;charset=UTF-8')
       )