NewEBook.scala

package com.josephpconley.books

import com.gargoylesoftware.htmlunit.html.{HtmlAnchor, HtmlImage, HtmlDivision, HtmlPage}
import com.gargoylesoftware.htmlunit.WebClient
import org.apache.http.client.methods.HttpGet
import org.apache.http.HttpStatus
import org.apache.http.impl.client.HttpClients
import org.apache.http.protocol.BasicHttpContext
import org.apache.http.util.EntityUtils
import com.josephpconley.rss.{Feed, Item}
import scala.xml.Elem
import scalax.io.Resource
import java.io.{File, PrintWriter}

/**
 * User: joe
 * Date: 10/21/13
 */

class NewEBookFeed(val name: String, val title: String, val description: String, val link: String, baseUrl: String) extends Feed {
  val atomLink: String = link
  lazy val webClient = new WebClient()
  lazy val opts = webClient.getOptions
  opts.setCssEnabled(false)
  opts.setJavaScriptEnabled(false)

  def htmlDescription(title: String, author: String, url: String, imgSrc: String) =
    <div>
      <a href={url}>
        <h3>{title}</h3>
        <h4>by {author}</h4>
      </a>
      <img src={imgSrc}/>
    </div>

  def items: Seq[Item] = {
    //go to homepage
    val page: HtmlPage = webClient.getPage(baseUrl)

    //grab the new ebooks page
    val newBookPage: HtmlPage = page.getByXPath("//a[starts-with(text(),'New eBooks')]").toArray().apply(1).asInstanceOf[HtmlAnchor].click()
    val titles = newBookPage.getByXPath("//div[@class='trunc-title-line']").toArray
    val authors = newBookPage.getByXPath("//div[@class='trunc-author-line']").toArray
    val images = newBookPage.getByXPath("//img[@class='lrgImg']").toArray

    //Should see a title of New ebooks with the 20 latest ebooks
    println(newBookPage.getTitleText)
    println(newBookPage.getPage.getUrl.getPath)
    println(titles.length)

    titles.indices.map{ i =>
      val titleAnchor = titles(i).asInstanceOf[HtmlDivision].getFirstChild
      val authorDiv = authors(i).asInstanceOf[HtmlDivision]

      val title = titleAnchor.getAttributes.getNamedItem("title").getNodeValue
      val author = authorDiv.getAttribute("title")
      val imgSrc = images(i).asInstanceOf[HtmlImage].getSrcAttribute
      val bookUrl = baseUrl + "/" + titleAnchor.getAttributes.getNamedItem("href").getNodeValue

      Item(title + " by " + author, htmlDescription(title, author, bookUrl, imgSrc).toString, bookUrl, bookUrl)
    }
  }
}