extract links (URLs), with nokogiri in ruby, from a href html tags?

You can do it like this:

doc = Nokogiri::HTML.parse(<<-HTML_END)
<div class="heat">
   <a href="http://example.org/site/1/">site 1</a>
   <a href="http://example.org/site/2/">site 2</a>
   <a href="http://example.org/site/3/">site 3</a>
</div>
<div class="wave">
   <a href="http://example.org/site/4/">site 4</a>
   <a href="http://example.org/site/5/">site 5</a>
   <a href="http://example.org/site/6/">site 6</a>
</div>
HTML_END

l = doc.css('div.heat a').map { |link| link['href'] }

This solution finds all anchor elements using a css selector and collects their href attributes.