Question

I’m writing a crawler with wombat. And somehow i’m using CSS selectors, not XPATH. And i have very difficult selection here – that i can’t achieve using css.

I have div elements that i want to grab from a page:

<div class="do_cat_ads_box"> ... </div>
<div class="do_cat_ads_box2"> ... </div>
<div class="do_cat_ads_box" style=".."> ...</div>
<div class="do_cat_ads_box2" style=".."> ... </div>

But elements with ‘style’ attribute – are garbage (ads) that i don’t need.

So my question is, can I grab all div elements with classes ‘do_cat_ads_box’ and ‘do_cat_ads_box2’, but avoid div elements that have ‘style’ attribute?

I ended up with something like this and it is not working:

application 'css=div.do_cat_ads_box2, div.do_cat_ads_box,  div.do_cat_ads_box:not(@style)', :iterator do
  href 'css=div.do_cat_ads_image  a @href'
  name 'css=div.do_cat_ads_detail a'
end

if it’s not double with css selectors, then there is always xpath way. But i’m very interested in css-selectors approach.

Answer 1

Attribute selectors in CSS use [attr] notation. The @attr notation pertains to attribute locators (as well as XPath).

Assuming Wombat supports the CSS syntax for attribute selectors, try changing :not(@style) to :not([style]) and rewriting your class selectors to the following:

application 'css=div.do_cat_ads_box:not([style]), div.do_cat_ads_box2:not([style])', :iterator do
  href 'css=div.do_cat_ads_image  a @href'
  name 'css=div.do_cat_ads_detail a'
end