1.0.53 - 8Muses ripper can rip subalbums/comics

Should close #52 since it also grabs various image types that aren't
consistent across all albums
This commit is contained in:
4pr0n 2014-06-04 20:23:21 -07:00
parent 2e7e0ec7bb
commit c5eb83a338
3 changed files with 135 additions and 27 deletions

View File

@ -4,7 +4,7 @@
<groupId>com.rarchives.ripme</groupId> <groupId>com.rarchives.ripme</groupId>
<artifactId>ripme</artifactId> <artifactId>ripme</artifactId>
<packaging>jar</packaging> <packaging>jar</packaging>
<version>1.0.52</version> <version>1.0.53</version>
<name>ripme</name> <name>ripme</name>
<url>http://rip.rarchives.com</url> <url>http://rip.rarchives.com</url>
<properties> <properties>

View File

@ -1,5 +1,6 @@
package com.rarchives.ripme.ripper.rippers; package com.rarchives.ripme.ripper.rippers;
import java.io.File;
import java.io.IOException; import java.io.IOException;
import java.net.MalformedURLException; import java.net.MalformedURLException;
import java.net.URL; import java.net.URL;
@ -7,12 +8,14 @@ import java.util.regex.Matcher;
import java.util.regex.Pattern; import java.util.regex.Pattern;
import org.apache.log4j.Logger; import org.apache.log4j.Logger;
import org.jsoup.Connection.Method;
import org.jsoup.Connection.Response; import org.jsoup.Connection.Response;
import org.jsoup.Jsoup; import org.jsoup.Jsoup;
import org.jsoup.nodes.Document; import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element; import org.jsoup.nodes.Element;
import com.rarchives.ripme.ripper.AlbumRipper; import com.rarchives.ripme.ripper.AlbumRipper;
import com.rarchives.ripme.ui.RipStatusMessage.STATUS;
import com.rarchives.ripme.utils.Utils; import com.rarchives.ripme.utils.Utils;
public class EightmusesRipper extends AlbumRipper { public class EightmusesRipper extends AlbumRipper {
@ -21,6 +24,8 @@ public class EightmusesRipper extends AlbumRipper {
HOST = "8muses"; HOST = "8muses";
private static final Logger logger = Logger.getLogger(EightmusesRipper.class); private static final Logger logger = Logger.getLogger(EightmusesRipper.class);
private Document albumDoc = null;
public EightmusesRipper(URL url) throws IOException { public EightmusesRipper(URL url) throws IOException {
super(url); super(url);
} }
@ -34,36 +39,139 @@ public class EightmusesRipper extends AlbumRipper {
public URL sanitizeURL(URL url) throws MalformedURLException { public URL sanitizeURL(URL url) throws MalformedURLException {
return url; return url;
} }
@Override
public String getAlbumTitle(URL url) throws MalformedURLException {
try {
// Attempt to use album title as GID
if (albumDoc == null) {
albumDoc = Jsoup.connect(url.toExternalForm())
.userAgent(USER_AGENT)
.method(Method.GET)
.timeout(Utils.getConfigInteger("download.timeout", 5000))
.execute()
.parse();
}
Element titleElement = albumDoc.select("meta[name=description]").first();
String title = titleElement.attr("content");
title = title.substring(title.lastIndexOf('/') + 1);
return HOST + "_" + title.trim();
} catch (IOException e) {
// Fall back to default album naming convention
logger.info("Unable to find title at " + url);
}
return super.getAlbumTitle(url);
}
@Override @Override
public void rip() throws IOException { public void rip() throws IOException {
logger.info(" Retrieving " + this.url); ripAlbum(this.url.toExternalForm(), this.workingDir);
Response resp = Jsoup.connect(this.url.toExternalForm())
.userAgent(USER_AGENT)
.execute();
Document doc = resp.parse();
int index = 0;
for (Element thumb : doc.select("img")) {
if (!thumb.hasAttr("data-cfsrc")) {
continue;
}
String image = thumb.attr("data-cfsrc");
if (image.contains("-cu_")) {
image = image.replaceAll("-cu_[^.]+", "-me");
}
if (image.startsWith("//")) {
image = "http:" + image;
}
index += 1;
URL imageURL = new URL(image);
String prefix = "";
if (Utils.getConfigBoolean("download.save_order", true)) {
prefix = String.format("%03d_", index);
}
addURLToDownload(imageURL, prefix);
}
waitForThreads(); waitForThreads();
} }
private void ripAlbum(String url, File subdir) throws IOException {
logger.info(" Retrieving " + url);
sendUpdate(STATUS.LOADING_RESOURCE, url);
if (albumDoc == null) {
Response resp = Jsoup.connect(url)
.userAgent(USER_AGENT)
.timeout(Utils.getConfigInteger("download.timeout", 5000))
.execute();
albumDoc = resp.parse();
}
int index = 0; // Both album index and image index
if (albumDoc.select(".preview > span").size() > 0) {
// Page contains subalbums (not images)
for (Element subalbum : albumDoc.select("a.preview")) {
ripSubalbumFromPreview(subalbum, subdir, ++index);
}
}
else {
// Page contains images
for (Element thumb : albumDoc.select("img")) {
downloadImage(thumb, subdir, ++index);
}
}
}
/**
* @param subalbum Anchor element of a subalbum
* @throws IOException
*/
private void ripSubalbumFromPreview(Element subalbum, File subdir, int index) throws IOException {
// Find + sanitize URL from Element
String subUrl = subalbum.attr("href");
subUrl = subUrl.replaceAll("\\.\\./", "");
if (subUrl.startsWith("//")) {
subUrl = "http:";
}
else if (!subUrl.startsWith("http://")) {
subUrl = "http://www.8muses.com/" + subUrl;
}
// Prepend image index if enabled
// Get album title
String subTitle = subalbum.attr("alt");
if (subTitle.equals("")) {
subTitle = getGID(new URL(subUrl));
}
subTitle = Utils.filesystemSafe(subTitle);
// Create path to subdirectory
File subDir = new File(subdir.getAbsolutePath() + File.separator + subTitle);
if (!subDir.exists()) {
subDir.mkdirs();
}
albumDoc = null;
ripAlbum(subUrl, subDir);
try {
Thread.sleep(2000);
} catch (InterruptedException e) {
logger.warn("Interrupted whiel waiting to load next album");
}
}
private void downloadImage(Element thumb, File subdir, int index) {
// Find thumbnail image source
String image = null;
if (thumb.hasAttr("data-cfsrc")) {
image = thumb.attr("data-cfsrc");
}
else if (thumb.hasAttr("src")) {
image = thumb.attr("src");
}
else {
logger.warn("Thumb does not havedata-cfsrc or src: " + thumb);
return;
}
// Remove relative directory path naming
image = image.replaceAll("\\.\\./", "");
if (image.startsWith("//")) {
image = "http:" + image;
}
// Convert from thumb URL to full-size
if (image.contains("-cu_")) {
image = image.replaceAll("-cu_[^.]+", "-me");
}
// Set download path
try {
URL imageURL = new URL(image);
String saveAs = subdir.getAbsolutePath() + File.separator;
if (Utils.getConfigBoolean("download.save_order", true)) {
// Append image index
saveAs += String.format("%03d_", index);
}
// Append image title
saveAs += Utils.filesystemSafe(thumb.attr("title"));
// Append extension
saveAs += image.substring(image.lastIndexOf('.'));
File saveFile = new File(saveAs);
// Download
addURLToDownload(imageURL, saveFile, thumb.baseUri(), null);
} catch (IOException e) {
logger.error("Failed to download image at " + image, e);
sendUpdate(STATUS.DOWNLOAD_ERRORED, "Failed to download image at " + image);
}
}
@Override @Override
public String getHost() { public String getHost() {

View File

@ -21,7 +21,7 @@ import com.rarchives.ripme.utils.Utils;
public class UpdateUtils { public class UpdateUtils {
private static final Logger logger = Logger.getLogger(UpdateUtils.class); private static final Logger logger = Logger.getLogger(UpdateUtils.class);
private static final String DEFAULT_VERSION = "1.0.52"; private static final String DEFAULT_VERSION = "1.0.53";
private static final String updateJsonURL = "http://rarchives.com/ripme.json"; private static final String updateJsonURL = "http://rarchives.com/ripme.json";
private static final String updateJarURL = "http://rarchives.com/ripme.jar"; private static final String updateJarURL = "http://rarchives.com/ripme.jar";
private static final String mainFileName = "ripme.jar"; private static final String mainFileName = "ripme.jar";