From f7ae0e8d96c2e8f537b7138e6b674fa67a2bbae2 Mon Sep 17 00:00:00 2001 From: rephormat Date: Fri, 12 Jan 2018 19:24:40 -0600 Subject: [PATCH 1/3] Add yuvutu.com ripper. --- .../ripme/ripper/rippers/YuvutuRipper.java | 70 ++++++++++++++++ .../ripper/rippers/video/YuvutuRipper.java | 82 +++++++++++++++++++ 2 files changed, 152 insertions(+) create mode 100644 src/main/java/com/rarchives/ripme/ripper/rippers/YuvutuRipper.java create mode 100644 src/main/java/com/rarchives/ripme/ripper/rippers/video/YuvutuRipper.java diff --git a/src/main/java/com/rarchives/ripme/ripper/rippers/YuvutuRipper.java b/src/main/java/com/rarchives/ripme/ripper/rippers/YuvutuRipper.java new file mode 100644 index 00000000..97365aa8 --- /dev/null +++ b/src/main/java/com/rarchives/ripme/ripper/rippers/YuvutuRipper.java @@ -0,0 +1,70 @@ +package com.rarchives.ripme.ripper.rippers; + +import java.io.IOException; +import java.net.MalformedURLException; +import java.net.URL; +import java.util.ArrayList; +import java.util.List; +import java.util.regex.Matcher; +import java.util.regex.Pattern; + +import org.jsoup.nodes.Document; +import org.jsoup.nodes.Element; + +import com.rarchives.ripme.ripper.AbstractHTMLRipper; +import com.rarchives.ripme.utils.Http; + +public class YuvutuRipper extends AbstractHTMLRipper { + + private static final String DOMAIN = "yuvutu.com", + HOST = "yuvutu"; + + public YuvutuRipper(URL url) throws IOException { + super(url); + } + + @Override + public String getHost() { + return HOST; + } + @Override + public String getDomain() { + return DOMAIN; + } + + @Override + public boolean canRip(URL url) { + Pattern p = Pattern.compile("^http://www\\.yuvutu\\.com/modules\\.php\\?name=YuGallery&action=view&set_id=([0-9]+)$"); + Matcher m = p.matcher(url.toExternalForm()); + return m.matches(); + } + + @Override + public String getGID(URL url) throws MalformedURLException { + Pattern p = Pattern.compile("^http://www\\.yuvutu\\.com/modules\\.php\\?name=YuGallery&action=view&set_id=([0-9]+)$"); + Matcher m = p.matcher(url.toExternalForm()); + if (m.matches()) { + return m.group(1); + } + throw new MalformedURLException("Expected yuvutu.com URL format: " + + "yuvutu.com/modules.php?name=YuGallery&action=view&set_id=albumid - got " + url + "instead"); + } + + @Override + public Document getFirstPage() throws IOException { + return Http.url(url).get(); + } + @Override + public List getURLsFromPage(Document doc) { + List imageURLs = new ArrayList<>(); + for (Element thumb : doc.select("div#galleria > a > img")) { + String image = thumb.attr("src"); + imageURLs.add(image); + } + return imageURLs; + } + @Override + public void downloadURL(URL url, int index) { + addURLToDownload(url, getPrefix(index)); + } +} diff --git a/src/main/java/com/rarchives/ripme/ripper/rippers/video/YuvutuRipper.java b/src/main/java/com/rarchives/ripme/ripper/rippers/video/YuvutuRipper.java new file mode 100644 index 00000000..f455abe2 --- /dev/null +++ b/src/main/java/com/rarchives/ripme/ripper/rippers/video/YuvutuRipper.java @@ -0,0 +1,82 @@ +package com.rarchives.ripme.ripper.rippers.video; + +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.nodes.Document; +import org.jsoup.nodes.Element; +import org.jsoup.select.Elements; + + +import com.rarchives.ripme.ripper.VideoRipper; +import com.rarchives.ripme.utils.Http; + +public class YuvutuRipper extends VideoRipper { + + private static final String HOST = "yuvutu"; + + public YuvutuRipper(URL url) throws IOException { + super(url); + } + + @Override + public String getHost() { + return HOST; + } + + @Override + public boolean canRip(URL url) { + Pattern p = Pattern.compile("^http://www\\.yuvutu\\.com/video/[0-9]+/(.*)$"); + Matcher m = p.matcher(url.toExternalForm()); + return m.matches(); + } + + @Override + public URL sanitizeURL(URL url) throws MalformedURLException { + return url; + } + + @Override + public String getGID(URL url) throws MalformedURLException { + Pattern p = Pattern.compile("^http://www\\.yuvutu\\.com/video/[0-9]+/(.*)$"); + Matcher m = p.matcher(url.toExternalForm()); + if (m.matches()) { + return m.group(1); + } + + throw new MalformedURLException( + "Expected yuvutu format:" + + "yuvutu.com/video/####" + + " Got: " + url); + } + + @Override + public void rip() throws IOException { + logger.info("Retrieving " + this.url); + Document doc = Http.url(url).get(); + Element iframe = doc.select("iframe").first(); + String iframeSrc = iframe.attr("src"); + if (iframeSrc != null) { + doc = Http.url("http://www.yuvutu.com" + iframeSrc).get(); + } else { + throw new IOException("Could not find iframe code at " + url); + } + Elements script = doc.select("script"); + if (script.size() == 0) { + throw new IOException("Could not find script code at " + url); + } + Pattern p = Pattern.compile("file: \"(.*?)\""); + + for (Element element : script) { + Matcher m = p.matcher(element.data()); + if (m.find()){ + String vidUrl = m.group(1); + addURLToDownload(new URL(vidUrl), HOST + "_" + getGID(this.url)); + } + } + waitForThreads(); + } +} From 4464fa873aa9d7c6159a072f4bf99f0fdc02b43c Mon Sep 17 00:00:00 2001 From: rephormat Date: Sat, 13 Jan 2018 09:31:35 -0600 Subject: [PATCH 2/3] Add yuvutu.com ripper test. --- .../tst/ripper/rippers/YuvutuRipperTest.java | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) create mode 100644 src/test/java/com/rarchives/ripme/tst/ripper/rippers/YuvutuRipperTest.java diff --git a/src/test/java/com/rarchives/ripme/tst/ripper/rippers/YuvutuRipperTest.java b/src/test/java/com/rarchives/ripme/tst/ripper/rippers/YuvutuRipperTest.java new file mode 100644 index 00000000..a74b76e9 --- /dev/null +++ b/src/test/java/com/rarchives/ripme/tst/ripper/rippers/YuvutuRipperTest.java @@ -0,0 +1,19 @@ +package com.rarchives.ripme.tst.ripper.rippers; + +import java.io.IOException; +import java.net.URL; + +import com.rarchives.ripme.ripper.rippers.YuvutuRipper; + +public class YuvutuRipperTest extends RippersTest { + + public void testYuvutuAlbum1() throws IOException { + YuvutuRipper ripper = new YuvutuRipper(new URL("http://www.yuvutu.com/modules.php?name=YuGallery&action=view&set_id=127013")); + testRipper(ripper); + } + + public void testYuvutuAlbum2() throws IOException { + YuvutuRipper ripper = new YuvutuRipper(new URL("http://www.yuvutu.com/modules.php?name=YuGallery&action=view&set_id=420333")); + testRipper(ripper); + } +} From 38231c82de85b25e9418b0582f862b3cbe9ed25b Mon Sep 17 00:00:00 2001 From: reph Date: Sat, 13 Jan 2018 18:40:15 -0600 Subject: [PATCH 3/3] Add yuvutu.com video ripper test. --- .../ripme/tst/ripper/rippers/VideoRippersTest.java | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/src/test/java/com/rarchives/ripme/tst/ripper/rippers/VideoRippersTest.java b/src/test/java/com/rarchives/ripme/tst/ripper/rippers/VideoRippersTest.java index 73c818d0..80abe56e 100644 --- a/src/test/java/com/rarchives/ripme/tst/ripper/rippers/VideoRippersTest.java +++ b/src/test/java/com/rarchives/ripme/tst/ripper/rippers/VideoRippersTest.java @@ -10,6 +10,7 @@ import com.rarchives.ripme.ripper.rippers.video.PornhubRipper; import com.rarchives.ripme.ripper.rippers.video.VineRipper; import com.rarchives.ripme.ripper.rippers.video.XvideosRipper; import com.rarchives.ripme.ripper.rippers.video.YoupornRipper; +import com.rarchives.ripme.ripper.rippers.video.YuvutuRipper; public class VideoRippersTest extends RippersTest { @@ -77,5 +78,14 @@ public class VideoRippersTest extends RippersTest { videoTestHelper(ripper); } } + + public void testYuvutuRipper() throws IOException { + List contentURLs = new ArrayList<>(); + contentURLs.add(new URL("http://www.yuvutu.com/video/828499/female-reader-armpit-job/")); + for (URL url : contentURLs) { + YuvutuRipper ripper = new YuvutuRipper(url); + videoTestHelper(ripper); + } + } }