using urllib on a more complex site

[Adam W.]

I'm trying to write a simple script to scrape http://www.vudu.com/movies/#tag/99centOfTheDay/99c%20Rental%20of%20the%20day

in order to send myself an email every day of the 99c movie of the day.

However, using a simple command like (in Python 3.0):
urllib.request.urlopen('http://www.vudu.com/movies/#tag/99centOfTheDay/99c%20Rental%20of%20the%20day').read()

I don't get the all the source I need, its just the navigation buttons. Now I assume they are using some CSS/javascript witchcraft to load all the useful data later, so my question is how do I make urllib "wait" and grab that data as well?

[Chris Rebert]

urllib isn't a web browser. It just requests the single (in this case, HTML) file from the given URL. It does not parse the HTML (indeed, it doesn't care what kind of file you're dealing with); therefore, it obviously does not retrieve the other resources linked within the document (CSS, JS, images, etc.) nor does it run any JavaScript. So, there's nothing to "wait" for; urllib is already doing everything it was designed to do.

Your best bet is to open the page in a web browser yourself and use the developer tools/inspectors to watch what XHR requests the page's scripts are making, find the one(s) that have the data you care about, and then make those requests instead via urllib (or the `requests` 3rd-party lib, or whatever). If the URL(s) vary, reverse-engineering the scheme used to generate them will also be required.

Alternatively, you could use something like Selenium, which let's you drive an actual full web browser (e.g. Firefox) from Python.

Cheers,

Chris

--
Cheers,
Chris
--
http://rebertia.com

[Dave Angel]

The CSS and the jpegs, and many other aspects of a web "page" are loaded
explicitly, by the browser, when parsing the tags of the page you
downloaded. There is no sooner or later. The website won't send the
other files until you request them.

For example, that site at the moment has one image (prob. jpeg)
highlighted,

<img class="gwt-Image" src="http://images2.vudu.com/poster2/179186-m"
alt="Sex and the City: The Movie (Theatrical)">

if you want to look at that jpeg, you need to download the file url
specified by the src attribute of that img element.

Or perhaps you can just look at the 'alt' attribute, which is mainly
there for browsers who don't happen to do graphics, for example, the
ones for the blind.

Naturally, there may be dozens of images on the page, and there's no
guarantee that the website author is trying to make it easy for you.
Why not check if there's a defined api for extracting the information
you want? Check the site, or send a message to the webmaster.

No guarantee that tomorrow, the information won't be buried in some
javascript fragment. Again, if you want to see that, you might need to
write a javascript interpreter. it could use any algorithm at all to
build webpage information, and the encoding could change day by day, or
hour by hour.

--
DaveA

[Adam W.]

The problem is, the image url you found is not returned in the data urllib grabs. To be clear, I was aware of what urllib is supposed to do (ie not download image data when loading a page), I've used it before many times, just never had to jump through hoops to get at the content I needed.

I'll look into figuring out how to find XHR requests in Chrome, I didn't know what they called that after the fact loading, so now my searching will be more productive.

[Adam W.]

On Sunday, February 24, 2013 7:30:00 PM UTC-5, Dave Angel wrote:

[Adam W.]

Huzzah! Found it: http://apicache.vudu.com/api2/claimedAppId/myvudu/format/application*2Fjson/callback/DirectorSequentialCallback/_type/contentSearch/count/30/dimensionality/any/followup/ratingsSummaries/followup/totalCount/offset/0/tag/99centOfTheDay/type/program/type/season/type/episode/type/bundle

Thanks for the tip about XHR's

[Adam W.]

On Sunday, February 24, 2013 7:27:54 PM UTC-5, Chris Rebert wrote: