links_from_hn.hs

import Network.Curl
import Text.XML.HXT.Core
import Control.Monad
import Data.List (isPrefixOf)

getHtmlRespOnly :: (CurlCode, String) -> String
getHtmlRespOnly (_, s) = s

main :: IO ()
main = do
    getHN <- curlGetString "https://news.ycombinator.com" []
    let contents = getHN
        html = getHtmlRespOnly contents
        doc = readString [withParseHTML yes, withWarnings no] html
        isExternalLink = isPrefixOf "http://"
        isYCombinator = isPrefixOf "http://ycombinator"
    links <- runX $ doc //> hasName "a" >>> getAttrValue "href"
    let links' = ([l | l <- links, (isExternalLink l) == True])
    mapM_ putStrLn ([l | l <- links', (isYCombinator l) == False])

-- test output
{-|
(northpole)kenny@hbox:~/locker/haskell$ runhaskell web.hs
http://www.s3for.me/index.html
http://www.nytimes.com/2013/09/06/health/gut-bacteria-from-thin-humans-can-slim-mice-down.html?src=me&ref=general&_r=0
http://vombat.tumblr.com/post/60530544401/date-parsing-performance-on-ios-nsdateformatter-vs
http://www.joystiq.com/2013/09/06/ouya-free-the-games-fund-project-suspended-by-kickstarter/
http://www.linuxjournal.com/content/linux-vs-bullshit
http://www.sciencebasedmedicine.org/i-refute-it-thus/
http://www.popehat.com/2013/09/06/nsa-codebreaking-i-am-the-other/
http://productforums.google.com/forum/#!topic/maps/toaewYdz0kg
http://blog.agilebits.com/2013/09/06/1password-and-the-crypto-wars/
http://www.fastcolabs.com/3016958/hey-japan-whats-up-with-your-startup-culture
http://www.usatoday.com/story/tech/2013/09/07/nasa-robotic-explorer-moon-va/2779137/
http://www.indiegogo.com/projects/support-0-a-d-an-open-source-strategy-game/
http://www.washingtonpost.com/business/technology/google-encrypts-data-amid-backlash-against-nsa-spying/2013/09/06/9acc3c20-1722-11e3-a2ec-b47e45e6f8ef_story.html?hpid=z1
http://www.washingtonpost.com/blogs/the-switch/wp/2013/09/06/the-feds-pays-for-60-percent-of-tors-development-can-users-trust-it/
http://learn.adafruit.com/introducing-trinket/introduction
http://www.wired.com/threatlevel/2013/09/tech-industry-tainted/
http://www.wired.com/wiredenterprise/2013/09/nginx/
http://blog.ircmaxell.com/2013/09/rambling-on-internals.html#comment-form
http://www.autocar.co.uk/car-video/video-tesla-model-s-vs-aston-martin-rapide-s
http://www.BeeLineReader.com
http://crystal-lang.org/2013/09/04/happy-birthday-crystal.html
|-}