11 KiB
Implement a service
Services or better service connectors are the parts of NewPipe which communicate with an actual service like YouTube. This Page will describe how you can implement and add your own to the extractor. Please make sure you read and understand the Concept of Extractors and the Concept of LinkHandler before doing this.
Required and optional parts
Your service does not have to implement everything. Some parts are optional. This is because not all services support every feature other services support. For example, it might be that a certain service does not support channels. If so, you can leave out the implementation of channels, and make the corresponding factory method of the your StreamingService implementation return null. The frontend will handle the lack of having channels then.
However if you start to implement one of the optional parts of the list below, you have to implement all parts/classes of it. NewPipe will crash if you only implement the extractor for the list item of a channel, but not the channel extractor itself.
The parts of a service:
Allowed Libraries
The NewPipe Extractor already comes along with a lot of usable tools and external libraries that should make extracting easy. For some specific (tiny) tasks regex is allowed. Here you can take a look at the Parser, which will give you a little help with that. Use Regex with care!!! Avoid it as often as possible. It's better to ask us to introduce a new library than start using regex to often.
- Html/XML Parsing: jsoup
- JSON Parsing: nanojson
- JavaScript Parsing/Execution: Mozilla Rhino
- Link detection in strings: AutoLink
If you need to introduce new libraries please tell us before you do it.
Head of Service
First of all if you want to create a new service you should create a new package below org.schabi.newpipe.services
, with the name of your service as package name.
Parts required to be implemented:
StreamingService is a factory class that will return objects of all important parts of your service. Every extractor, handler, and info type you add, and which should be part of your implementation, must be instantiated using an instance of this class. You can see it as a factory for all objects of your implementation.
ServiceInfo
will return some meta information about your service. Such as the name, the capabilities, and your name as well as your
email address for further notice and maintenance issues. Remember, after extending this class you need to return an
instance of it by through your implementation of
StreamingService.getServiceInfo().
When these two classes are extended by you, you need to add your StreamingService to the
ServiceList
of NewPipe. This way your service will become an official part of the NewPipe Extractor.
Every service has an ID, which will be set when this list gets created. You need to set this ID by entering it in the constructor.
So when adding your service just give it the ID of the previously last service in the list incremented by one.
Stream
Streams are considered single entities of video or audio. They come along with metainformation like a title, a description, next/related videos, a thumbnail and comments. For getting the URL to the actual stream data as well as this metainformation StreamExtractor is used. The LinkHandlerFactory will represent a link to such a stream. StreamInfoItemExtractor will extract one item in a list of items representing such Streams, like a search result or a playlist. Since every Streaming service (obviously) provides streams this is required to implement. Otherwise your service was pretty useless :)
Parts required to be implemented:
Search
The SearchExtractor is also required to be implemented. It will take a search query represented as SearchQueryHandler and return a list of search results. Since many services support a suggestion popup while you type you will also want to implement a SuggestionExtractor. This will make it possible for the frontend to as well display a suggestion while typing.
Parts required to be implemented:
Playlist
Playlists are lists of streams provided by the service (you may not have to take care about locally saved playlists. This will be handled by the frontend). A playlist may only contains StreamInfoItems, but no other InfoItem types.
Parts required to be implemented:
Channel
A Channel is mostly a Playlist, the only difference is that it does not represent a simple list of streams, but a user, a channel, or any entity that could be represented as a user. This is why the metadata supported by the channel extractor differs form the one of a playlist.
Parts required to be implemented:
Kiosk
A kiosk is a list of InfoItems which will be displayed on the main page of NewPipe. A kiosk is mostly similar to the content displayed on the main page of a video platform. A kiosk could be something like "Top 20", "Charts", "News", "Creators selection" etc. Kiosk are controversial, many people may not like them. If you don't like them as well please don't refuse to add them thought. Your service would look pretty empty if you select it and no video is being displayed. Also you should not override the preference of the user, since users of NewPipe can decide by the settings weather they want to see the kiosk page or not.
Multiple Kiosks
Most services will implement more than one Kiosk, so a service might have a "Top 20" for different categories like "Country Music", "Techno", etc. This is why the extractor will let you implement multiple KioskExtractors. Since different kiosk pages might also differ with their HTML structure every page you want to support has to be implemented as its own KioskExtractor. However if the pages are similar you can use the same Implementation, but set the page type when you instantiate your KioskExtractor through the KioskList.KioskExtractorFactory.
Every Kiosk you implement needs to be added to your KioskList which you return with your StreamingService implementation.
It is also important to set the default kiosk. This will be the kiosk that will be shown by the first start of your service.
An example implementation of the getKioskList() could look like this:
@Override
public KioskList getKioskList() throws ExtractionException {
KioskList list = new KioskList(getServiceId());
list.addKioskEntry(new KioskList.KioskExtractorFactory() {
@Override
public KioskExtractor createNewKiosk(StreamingService streamingService,
String url,
String id,
Localization local)
throws ExtractionException {
return new YoutubeTrendingExtractor(YoutubeService.this,
new YoutubeTrendingLinkHandlerFactory().fromUrl(url), id, local);
}
}, new YoutubeTrendingLinkHandlerFactory(), "Trending");
list.setDefaultKiosk("Trending");
return list;
}
Parts required to be implemented: