gistfile1.hs

{-# LANGUAGE OverloadedStrings #-}

import Control.Monad
import Control.Monad.IO.Class
import Data.Char
import Data.Conduit
import Data.Conduit.Binary (sinkFile)
import Network.HTTP.Conduit
import Text.HTML.DOM
import Text.Printf
import Text.XML.Cursor
import qualified Data.ByteString as B
import qualified Data.ByteString.Lazy as L
import qualified Data.Text as T

extractLink cursor
  | null links = Nothing
  | otherwise  = Just . T.unpack . head $ links
  where
    contains k = check $ T.isInfixOf k . T.concat . content
    links      = cursor
      $// element "a"
      &/  element "strong"
      &/  contains "Download to Spreadsheet"
      >=> parent
      >=> parent
      >=> attribute "href"

getLink manager = do
  msym <- await
  case msym of
    Nothing     -> return ()
    Just sym -> do
      let sym' = concatMap (printf "%%%02x" . ord) $ sym
      let url  = "http://finance.yahoo.com/q/hp?s=" ++ sym'
      req <- parseUrl url
      res <- http req manager
      xml <- responseBody res $$+- sinkDoc
      yield (sym, extractLink . fromDocument $ xml)

downloadContent manager = do
  mx <- await
  case mx of
    Nothing               -> return ()
    Just (sym, Nothing)   -> return ()
    Just (sym, Just link) -> do
      req <- parseUrl link
      res <- http req manager
      responseBody res $$+- sinkFile sym

sourceSyms = do
  yield "VTI"

main = do
  runResourceT $ do
    withManager $ \manager -> do
      sourceSyms $= getLink manager $$ downloadContent manager
  return ()