[Scala][Web] 伊莉文章的Parser

gistfile1.scala

package rainy.maid.server.seeker
import java.util.ArrayList
import rainy.maid.server.domain.EneyPost
import scala.xml.NodeSeq
import scala.xml.{ XML, Node }
import de.hars.scalaxml._
import java.text.SimpleDateFormat
class eneySeeker(root: String, url: String) {

    var nextPage = url
    val normalThreadR = "normalthread_([0-9]{6,7})".r
    val dateFormat = new SimpleDateFormat("yyyy-MM-dd")
    //回傳資料型態圍List[EneyPost] 分別放著ID,文章名稱,文章超連結,最後修改日
    def next() = {
        val result = new TagSoupFactoryAdapter loadString (scala.io.Source.fromURL(nextPage, "big5").mkString)
        val tbodys = result \\ "tbody"
        val subjects = new ArrayList[EneyPost]
        //只取出id為normalthread的文章，因為這些不是置頂文
        for (tbody ← tbodys if (((tbody \ "@id").text.contains("normalthread")))) 
        {
            subjects add (getSubject(tbody))
        }
        nextPage = root + getNextPage(result)
        subjects
    }
    def getSubject(tbody: Node) =
        {
            //id即為此篇文章在論壇的ID
            //normalthread底下的第一個span存放著每篇文章的標題與超連結
            val titleEl = (tbody \\ "span").head
            val normalThreadR(id) = (tbody \ "@id").text
            val newPost = new EneyPost
            newPost setId (id.toLong)
            newPost setTitleName (titleEl.text)
            newPost setURL (root + (titleEl \ "a" \ "@href"))
            newPost setLastModifyTime (dateFormat.parse(getLastModify(tbody)))
            newPost
        }
    def getLastModify(tbody: Node) =
        {
            //取出Tbody底下的td Tag，並且此td Tag的Class屬性為author
            (RichNodeSeq(tbody) \\ "td[@class==author]" \\ "em").text
        }

    def getNextPage(body: Node) =
        {
            //找出所有Div Tag屬性Class為Pages的節點，而此Div底下的所有超連結，我們只看class為Next那個
            (RichNodeSeq(body) \\ "div[@class==pages]" \\ "a[@class==next]" \\ "@href").text
        }
}

test.scala

package test.rainy.maid.server.maidwork
import org.junit.After
import org.junit.Before
import org.junit.Test
import com.google.appengine.tools.development.testing.LocalDatastoreServiceTestConfig
import com.google.appengine.tools.development.testing.LocalServiceTestHelper
import org.scalatest.junit.JUnitSuite
import scala.collection.mutable.ListBuffer
import rainy.maid.server.domain._
import rainy.maid.server.maidwork._
import scala.collection.JavaConversions._

class testEneySearchNewPost extends JUnitSuite{
        val helper:LocalServiceTestHelper = new LocalServiceTestHelper(new LocalDatastoreServiceTestConfig());

        @Before
        def setUp()=
        {
                helper.setUp
        }
        
        @After
        def tearDown()=
        {
                helper.tearDown
        }
        
        @Test
        def testSearchNewPost()=
        {
                val workLog = new EneyWorkLog
                workLog setRootPath "http://www.eyny.com/"
                workLog setFullPath "http://www.eyny.com/forumdisplay.php?fid=22&filter=&orderby=dateline"
                workLog setLastPostID 5506759
                
                val toDoList = new MaidToDoList with EneySearchNewPost
                toDoList doWork(workLog)
                
                val posts = EneyPostManager findPosts(false)
                posts foreach(post=>println(post.getId,post.getTitleName))
        }
}