I'm trying to parse an html page and output it into a list of links with there x,y corrdinates. I'm already using getLinks from the DOM Object in PHP, as described here: http://www.phpro.org/examples/Get-Links-With-DOM.html ... it works wonders, I trim the list and only return the Text Description of the link.

From everything that I can find and everything I've tried, I can get the x,y coordinates using javascript on the client but this won't be running on the client so that's no good... and I really don't like javascript.

Does anyone know how I could go about this to grab the X,Y's coordinates of these links maybe in Perl? Any help would be appreciated, btw: Again I'm trying to keep this server side.

Thanks in advance for any help.

This is only possible with JavaScript, because JavaScript "knows" the actual state of the rendered HTML/CSS and the actual window size and things like that you need to calculate XY coordinates within a webpage.

There is the possiblity to run a virtual browser from within Perl though, but AFAIK JavaScript is still necessary.

Forgot how the project was called - but it was made to measure user's click'n'drag behavior in a browser.

Ha! Here it is: http://seleniumhq.org/

And short reminder: On the server side there are no XY coordinates _yet_. They exist only after the browser rendered the page - and they all depend on window size, font size, typ of layout of the webpage - not to mention the possibility that user X filters the links...

If you just need the links - at least 10 perl modules will deliver them to you without every calculating one coordinate.

Check CPAN for HTML and parsers.

Just using

Perl parse DOM html

in yahoo.com yields http://search.cpan.org/~sprout/HTML-...ib/HTML/DOM.pm .

The module is capable of returning links - if I understand correctly. Is it what you need ?

Only the browser, and its built-in javascript interpreter, knows how the HTML is rendered, and browsers will render the page differently based on many things. Perhaps you can tell us what you really want to do, or why you think you want to do this, and alternative solutions can be suggested.
--- rod.

Actually I don't need the client browser to generate the page view, I'm using CutyCapt "http://cutycapt.sourceforge.net/" which emulates a safarie browser environment through the use of webkit. It requires that the server have X running on it so in essence it's viewing the page being requested. Additionally It enables me to set a fixed browser size such as 1024X768.

Perhaps I can get the javascript to do it's thing within the webkit environment...

What exactly are you trying to accomplish?

There is no such thing as a fixed XY-coordinate of some link within a website.

This entirely depends on how the browser renders the page - and even if a page's got a fixed layout with fixed pixel widths and heights, I still can change the font size in my browser which will move any text - including links - to different coordinates.

Not to mention that many people just don't use 1024x768 as a resolution (me for example), don't open their browser in full screen mode (me for example) and don't use webkit-based browsers (all Firefox, Opera and IE users, for example.)

You would have to grab all this values from within a user's browser environment (which I would consider _highly_ intrusive into my privacy), it would require that a user has _actually_ JavaScript enabled, doesn't filter stuff or uses NoScript...

And I haven't even started yet with userContent.css manipulations...

If you just want to grab links, use _anything_ able handling DOM - there're at least 10 Perl modules to do that on CPAN, as Sergei suggested.

Even some Regexp would do.

Okay, I think I can read between the lines. You want to use CutyCapt to grab web pages, save them to image files, and then turn the links into image-maps associated with the bitmap images. My recommendation is to dig into the source for CutyCap, and add your requirement as a feature. It sounds like something a lot of users could make use of, and that is precisely the definitive place to acquire the information. Since the tool can store web pages as SVG, and SVG is a XML formatted text file, you may be able to extract XY coordinate information from that format.
--- rod.