ImgScroll/src/main/java/com/rarchives/ripme/ripper/rippers/XhamsterRipper.java
2014-06-07 18:06:38 -07:00

88 lines
2.6 KiB
Java

package com.rarchives.ripme.ripper.rippers;
import java.io.IOException;
import java.net.MalformedURLException;
import java.net.URL;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import com.rarchives.ripme.ripper.AlbumRipper;
import com.rarchives.ripme.utils.Utils;
public class XhamsterRipper extends AlbumRipper {
private static final String DOMAIN = "xhamster.com",
HOST = "xhamster";
public XhamsterRipper(URL url) throws IOException {
super(url);
}
@Override
public boolean canRip(URL url) {
return url.getHost().endsWith(DOMAIN);
}
@Override
public URL sanitizeURL(URL url) throws MalformedURLException {
return url;
}
@Override
public void rip() throws IOException {
int index = 0;
String nextURL = this.url.toExternalForm();
while (nextURL != null) {
logger.info(" Retrieving " + nextURL);
Document doc = Jsoup.connect(nextURL).get();
for (Element thumb : doc.select("table.iListing div.img img")) {
if (!thumb.hasAttr("src")) {
continue;
}
String image = thumb.attr("src");
image = image.replaceAll(
"http://p[0-9]*\\.",
"http://up.");
image = image.replaceAll(
"_160\\.",
"_1000.");
index += 1;
String prefix = "";
if (Utils.getConfigBoolean("download.save_order", true)) {
prefix = String.format("%03d_", index);
}
addURLToDownload(new URL(image), prefix);
}
nextURL = null;
for (Element element : doc.select("a.last")) {
nextURL = element.attr("href");
break;
}
}
waitForThreads();
}
@Override
public String getHost() {
return HOST;
}
@Override
public String getGID(URL url) throws MalformedURLException {
Pattern p = Pattern.compile("^https?://([a-z0-9.]*?)xhamster\\.com/photos/gallery/([0-9]{1,})/.*\\.html");
Matcher m = p.matcher(url.toExternalForm());
if (m.matches()) {
return m.group(2);
}
throw new MalformedURLException(
"Expected xhamster.com gallery formats: "
+ "xhamster.com/photos/gallery/#####/xxxxx..html"
+ " Got: " + url);
}
}