-
Notifications
You must be signed in to change notification settings - Fork 630
How To Create A Ripper for HTML websites
This guide explains how to rip from an unsupported website using RipMe.
If you like to learn by example, check out the simple ImgboxRipper.java
.
- Some knowledge of the Java programming language
- Some knowledge of the build tool
Maven
- This is for dependency resolution & so you don't have to download a ton of .jar files
- Some knowledge of the HTML DOM, CSS selectors, and the like.
Step 0: Fork this repo
Create the file within / src / main / java / com / rarchives / ripme / ripper / rippers
File should follow the naming scheme <Site>Ripper.java
Step 2: Extend the AbstractHTMLRipper
class
public class YoursiteRipper extends AbstractHTMLRipper {
By extending AbstractHTMLRipper
, you have access to the this.url
object containing the URL to be ripped.
We need to let the superclass know what URL we're working with.
Change the constructor's class name to your ripper's class name.
public YoursiteRipper(URL url) throws IOException {
super(url);
}
The methods below are defined in AbstractHTMLRipper
and must be overridden in your .java file.
Returns: The name of the website (no need for .com
).
This String is used in naming the save directory.
@Override
public String getHost() {
return "imgur";
}
Returns: The domain of the website.
This String is used in the canRip()
method to determine if a URL can be ripped.
@Override
public String getDomain() {
return "imgur.com";
}
Returns: A unique identifier for the album (Gallery ID or GID).
Note: The URL to every album on the website should return a different GID.
This is because the save directory will be named in the scheme HOST_GID
Most rippers use regex
to strip out the GID.
Example: imgur.com/a/abc123
could return abc123
@Override
public String getGID(URL url) throws MalformedURLException {
Pattern p = Pattern.compile("^https?://imgur\\.com/a/([a-zA-Z0-9]+).*$");
Matcher m = p.matcher(url.toExternalForm());
if (m.matches()) {
// Return the text contained between () in the regex
return m.group(1);
}
throw new MalformedURLException("Expected imgur.com URL format: " +
"imgur.com/a/albumid - got " + url + " instead");
}
Returns: A Jsoup Document
object containing the contents of the first page.
Tip: Use the Http
class for easy methods of retrieving the page.
Most rippers just need to get the page, and do so with:
@Override
public Document getFirstPage() throws IOException {
// "url" is an instance field of the superclass
return Http.url(url).get();
}
This works for the majority of websites (most sites don't require cookies, referrers, etc).
Input: Jsoup Document
retrieved in the getFirstPage()
method.
Returns: The next page to retrieve images from.
Throws: IOException
if no next page can be retrieved.
Note: By default, this method throws an IOException
within AbstractHTMLRipper
, meaning it assumes there is no next page. If you need to rip multiple pages, override this method & retrieve the next page. See ImagebamRipper.java
for an example of how this is used.
Returns: The url that will be either written to or checked against the url history
This function normalizes a url so that non-static urls can be used with the url history file
Here is an example removing the time stamp ID from instagram links
@Override
public String normalizeUrl(String url) {
// Remove the date sig from the url
return url.replaceAll("/[A-Z0-9]{8}/", "/");
}
Input: Jsoup Document
retrieved in the getFirstPage()
method (and optionally the getNextPage()
method).
Returns: List of URLs to be downloaded or retrieved.
This is where the URLs are extracted from the page Document.
Some rippers return a list of subpages to be ripped in separate threads (e.g. ImagevenueRipper.java
)
This is when CSS-Selectors come in handy. Say you wanted to grab every image that appears on the page:
@Override
public List<String> getURLsFromPage(Document doc) {
List<String> result = new ArrayList<String>();
for (Element el : doc.select("img")) {
result.add(el.attr("src"));
}
return result
}
This would return the source to all images on the page (although they will likely be thumbnails).
The URLs returned are passed into the next method...
Input: URL
: One of the URLs returned by getURLsFromPage()
Input: index
: The number for this URL (whether it's the 1st image, 2nd image, etc).
This is where your ripper downloads the image/file.
Most rippers simply use the AlbumRipper
's method addURLToDownload()
:
@Override
public void downloadURL(URL url, int index) {
addURLToDownload(url, getPrefix(index));
}
The above will download the URL to the appropriate save directory, guessing the filename to save based on the url
and a given prefix (index
).
The addURLToDownload()
method is heavily overloaded with lots of options.
Variants of this method allow you to:
- Define the exact file name to save as,
- The subdirectory to save to,
- HTTP headers (such as cookies or referrers) that should be used while downloading the file,
- Saving the file with the MIME type as it's extension
Other rippers, as mentioned before, start a separate Thread to retrieve the full-size image from the provided URL. Your implementation may vary.
If this function returns true then ripme does not call addURLToDownload and does not expect getURLsFromPage to return any urls. This is usful if you want to download images as you find them instead of waiting for getURLsFromPage to finish
For an example of a ripper which rips like this see the 8muses ripper
Note: This has nothing to do with support for ripmes download Queue. This has to do with a ripper which adds new albums to ripmes download Queue. Rippers without this override can still have links for site they support added to the download Queue by the user.
This denotes whether or not a ripper has support for QuickQueuing (Adding new albums to ripmes download Queue as if a user had entered it into the rip textbox and then pressed rip)
If this function returns true then ripme expects nothing on the paged to be ripped, but instead to add links from the page to ripmes download queue. If you over ride this you must also override hasQueueSupport and getAlbumsToQueue
Input: The Document returned by getFirstPage() Returns: A list of urls that will be added to the download Queue to be ripped
This is very useful when you want your ripper to be able to support subalbums.
For an example of a ripper which uses these functions see the Myhentaicomics ripper
RipMe automatically detects new rippers without any other code changes required.
- Execute the ripper.
- Paste in a URL to the site you're trying to rip.
- Click
Rip
- Look at the output for errors, warnings, Exceptions, etc.
- Fix any bugs.
- Repeat.