Skip to content

Commit

Permalink
Update scraper.rb
Browse files Browse the repository at this point in the history
  • Loading branch information
boundedrational committed Dec 13, 2014
1 parent e92806f commit e3616cc
Showing 1 changed file with 3 additions and 9 deletions.
12 changes: 3 additions & 9 deletions scraper.rb
Expand Up @@ -145,9 +145,7 @@
pageurl = helpurl+"&start="+j.to_s()
page = Nokogiri::HTML(open(URI::encode(pageurl)))
page.search("div[@itemtype='http://schema.org/JobPosting']").each do |node|

if node.count > 0

if node.count > 0
jobtitle=node.css("h2 a").inner_html
jobtitle = Sanitize.clean(jobtitle)
ident=node.css("h2")
Expand Down Expand Up @@ -183,8 +181,6 @@
"efccid"=>'',
"cmpid"=>'',
"geo"=>'',


}


Expand Down Expand Up @@ -213,12 +209,10 @@
puts "Skipping already saved record " + record['title']
end
end




end
end
end
end

j=j + resultsperpage
end
Expand Down

0 comments on commit e3616cc

Please sign in to comment.