Python- automate navigation through websites (Part 1)

Hi, There are many cases where we need to do the same thing over and over again on websites, for instance downloading episodes of a series manually from a website, checking mails every 10 mins waiting for some important mails, logging into your company or college website everyday just for the sake of logging in. For the impatient people out there, python comes to rescue 🙂

Python and its libraries can be used very effectively to navigate through websites, fill redundant forms and a lot more. A lot of cool things can be built where the website lacks an API. Some examples are: An automatic files downloader from a website, automated login and filling forms, booking tickets automatically, automatic payments using your credit card or debit card. So here, I am going to discuss a few libraries and some tips to use them effectively.

For checking out the content of websites, beautiful soup and urllib/urllib2 are libraries to look at. I have written about them here regarding some basic navigation https://crondev.wordpress.com/2014/06/15/use-python-to-download-files-from-websites/

Here we are looking at python ways of interacting with website forms. One of the difficult tasks is to login into websites.

Doing it the hard way- Using urllib2

Kind of a difficult and complex way, I would say. We will need to use cookielib for this task. So, let’s import it

  import cookielib

Now, we need to create a cookie jar to keep all cookies.
  cj = cookielib.CookieJar()
  opener = urllib2.build_opener(urllib2.HTTPCookieProcessor(cj))
  urllib2.install_opener(opener)

Next, we need to make a request to the login page using the username and password as login parameters.
  authentication_url = 'https://site/login/index.php'
  payload = {
      'username' : username,
      'password' : password
  }
  data = urllib.urlencode(payload)
  req = urllib2.Request(authentication_url, data)

Now, we can proceed with opening the webpages.
  resp = urllib2.urlopen(  'https://site/link' )

Doing it the Mechanize way 🙂
Mechanize is a very powerful engine and includes a browser itself. So users don’t need to worry about cookies as long as they use the same browser object. Here is a small snippet with inline comments to describe how to use it. Make a new browser object
  br = Browser()

Open a page
  page = br.open( 'https://example.com/examplepage' )

Mechanize has inbuilt functions for handling forms easily
  br.select_form( nr = 0 )   #Selecting a form using the form number 0 for first
  br.select_form( name = 'form1' ) #Select form using the name- can be easily found using the source of the page

Note that the browser object takes care of cookies automatically. Now the form is selected, you can fill in form fields
  br.form[ 'userLogin' ] = user_name
  br.form[ 'userPassword' ] = password

You can also select radio options this way. Here 2 is the option-ID, if not specified, radio buttons start with 0 as option-ID(Inspect the element to check if it is specified)
  br.form[ 'radio-name' ] = [ '2' ]

Using select drop-downs can be a bit tricky at times but here is the way
  select1 = br.form.find_control( 'dropdown-name' )
  for item in select1.items:
      if item.name == 'required':
          item.selected = True

Submit the form with all these parameters
  br.submit()

Directly go to the next page. Cookies taken care of by Mechanize browser
  resp = br.open( 'https://nextpage.html' )

Note: This is an error that I faced (in case the HTML in the site has no closing tags somewhere)


  mechanize._form.ParseError: nested FORMs



Here’s the fix. Analyze the response and then modify it to what mechanize would like to see ie fix the non matching tags to generate the response that should have been. Setting response for Page as the hardcoded response:
  resp.set_data( hardcoded_resp )
  br.set_response( resp )

I will end this post here to avoid making it too lengthy, I will add more ways of doing this in the second part of this post.