Hanjo Jo'burg Odendaal Hanjo Jo'burg Odendaal - 2 months ago 8
R Question

scrape website with hidden csrf token at login with R

As part of a late night project I am working on out of interest, I am trying to scrape my Uber trip data off of their website.

I have had a look at the code of the login page on


https://login.uber.com/login


and have seen that they use POST method in their form setup as follows:

<form method="post" class="form" novalidate="">
<input type="hidden" name="_csrf_token" value="1452201446-01-hujzoBTxkYPrJessd6zQwnD2ZOFxMOVgIYN8iXntr6c=">
<input type="hidden" data-js="access-token" name="access_token">
<a href="#" class="btn btn--full btn--facebook" data-js="facebook-connect">
<span class="push--ends flush">Continue with Facebook</span>
</a>
<p class="primary-font primary-font--semibold text-uber-white background-line push--top push--bottom">
<span>or use email</span>
</p>

<div class="form-group push-tiny--top flush--bottom">
<input type="email" name="email" class="text-input square--bottom " placeholder="Email Address" value="" id="email">
</div>
<div class="form-group push--bottom">
<input type="password" name="password" class="text-input square--top " placeholder="Password" id="password">
</div>


What I have read up is that one needs to send csrf token along when trying to scrape

library(RCurl)
library(XML)

URL_str<-"https://login.uber.com/login"
URL_str2<-"https://riders.uber.com/trips"

email<-"exampleMail@gmail.com"
pass<-"thisisalong"
token<- "1452201446-01-hujzoBTxkYPrJessd6zQwnD2ZOFxMOVgIYN8iXntr6c="

params <- list('email' = email,
'password' = pass,
'_csrf_token'=token)

URL_doc = postForm(URL_str2, style="POST",
.params=params)


If I now try and scrape the site, I get

ERROR: FORBIDDEN


I have seen some examples in python with similar websites. Can the same be done in R?

Answer

The final answer ended up being quite simple:

library(rvest)
library(RCurl)
library(XML)

session <-html_session("https://login.uber.com/login")

email<-"exampleMail@gmail.com"
pass<-"thisisalong"

#Handling the html_form
form<-html_form(session)[[1]]
form<-set_values(form, email=email, password=pass)

#Found that unless I explicitly assigned the value to empty it created errors
form$url<-""

session_open<-submit_form(session,form)

session_open<-jump_to(session_open,URL_str2)

From here on in its straight forward pulling tables using html_table and isolating nodes using html_nodes. Hope this helps

Comments