Find all <a>s to who's href points to an image with XPath 1.0

Question

First, I can assume that all urls that end with jpeg, jpg, bmp, png or gif are images, and others aren't.

I thought of, and tried two solutions:

• Matching the regular expression .(jpe?g|bmp|png|gif)$
• Using ends-with to check each separately

But, it appears that neither of these exist in XPath 1.0, or at least, they don't exist in Firefox (I am writing a greasemonkey script, so it is only important for the path to work in Firefox).

Sorry about the title, SO didn't think "Find all links to images in XPath 1.0" was unique enough

Answer 1

It's going to be difficult because XPath does not have an ends-with() function, which you would need to use to check the end of the HREF attribute for your relevant file extensions. You will have to resort to using substring-after().

It's not going to be pretty, something like...

fn:substring-after(a[@href], '.') = 'jpg'

EDIT: It appears that substring-after is not a part of XPath 1.0, so you'll have to use the even uglier substring and string-length

Answer 2

You can use a combination of substring and string-length (both of which are in XPath 1.0) to simulate ends-with. It's not pretty, but it works:

substring(@href, string-length(@href) - 3 + 1, 3) = 'jpg'

(the 3s here are the length of jpg; the 1 is to adjust for substring's 1-based indexing)

should have the same truth value as

ends-with(@href, 'jpg')

I assume from your question that you know how to check for each possible extension separately.

Answer 3

Although you're asking for an XPath solution, an alternative approach would be to use something like jQuery or Prototype, which uses CSS Selectors to select elements.

With jQuery, for instance, you could use:

$("a[href$='jpg'],a[href$='gif'],a[href$='png']").each(functionOfChoice);

Answer 4

There are no regular expressions in XPath 1.0, and there is neither a ends-with().

URLs cannot contain spaces without becoming invalid, so you can use them to achieve an end-of-string matching. You could do this:

//a[
  contains(concat(@href, ' '), '.jpg ' or
  contains(concat(@href, ' '), '.bmp ' or
  contains(concat(@href, ' '), '.png '
]

or this (as @AakashM suggests):

//a[
  substring(@href, string-length(@href) - 2, 3) = 'jpg' or
  substring(@href, string-length(@href) - 2, 3) = 'bmp' or
  substring(@href, string-length(@href) - 2, 3) = 'png'
]

I think the latter option would perform a bit better, while the former is shorter and somewhat more pleasing to look at. Depends if you desperately need performance here.