TumblrCrawler.scala

import java.net.URL
import java.awt.image.BufferedImage
import javax.imageio.ImageIO
import java.io._
import scala.xml._
import scala.io.Source

object TumblrCrawler {

  def main( args:Array[String] ):Unit= {
    val tumblrUrl = "http://%s.tumblr.com/api/read".format( args.first )
    val r = """http\:\/\/.+/([^\/]+)*$""".r
    val extR = """.+\.(.+)$""".r

    def crawlingTumblrImages( cnt:Int ):Unit = {
      val url = tumblrUrl + "?type=photo&start=%d&num=50".format( cnt )
      println( url )
      val source = Source.fromURL( url )

      val xml = XML.loadString( source.getLines.mkString )
      val photos = xml \\ "post"

      photos size match {
        // 取れなくなったら終了
        case 0  => None
        // 画像をファイルに書き出して再帰
        case _ =>
          // もっとも解像度の高いURLを取り出す
          for( photo <- photos ) {
            val ps = photo \ "photo-url"
            val imageUrl = ( ps.first /: ps ){ (p1:Node, p2:Node) => {
              def getSize( node:Node ) = ( node \ "@max-width" text ).toInt
                if( getSize( p1 ) > getSize( p2 ) ) p1 else p2
            }}.text
            saveImage( imageUrl )
          }
          crawlingTumblrImages( cnt + 50 )
      }
    }

    def saveImage( url:String ) = {
      val r(fname) = url
      val (ext,file)  = fname match {
        case extR(e) => (e, new File( fname ) )
        case _ => ( "png", new File( fname + ".png" ) )
      }

      val img = ImageIO.read( new URL( url ) )
      ImageIO.write( img , ext , file )

      Thread.sleep( 1000  )

      println( "Download:" + fname )
    }

    crawlingTumblrImages( 0 )
  }
}