SinkException if the sink is closed or becomes full, allowing event queues to support thresholding and backpressure. The interface design is heavily influenced by Matt Welsh's SandStorm server, his demonstration of the SEDA architecture. We have deviated where we felt the design differences where better.
@author Avalon Development TeamA sink is connected to a Channel and consumes its contents, sending them to a configured destination that may vary according to the sink type.
Sinks can be grouped together for various behaviors using SinkGroup and SinkProcessor. They are polled periodically by a SinkRunner via the processor
Sinks are associated with unique names that can be used for separating configuration and working namespaces.
While the {@link Sink#process()} call is guaranteed to only be accessedby a single thread, other calls may be concurrently accessed and should thus be protected.
@see org.apache.flume.Channel @see org.apache.flume.SinkProcessor @see org.apache.flume.SinkRunnerDoxia allows you to transform any supported input document format (ie for which a Parser exists) into any supported output document format (ie for which a Sink exists).
A parser is responsible for reading an input document and emitting a sequence of Doxia events which can then be consumed by a Doxia Sink. Thus, you can parse any front- end format (eg APT, FML, Xdoc, ...) and have them all contribute to a final XHTML version of a web site. All documents being parsed result in a stream of Doxia events (eg paragraph, bold, italic, text,...), which are then fed into a XHTML Sink to produce a set of XHTML pages.
A Sink is ultimately responsible for the final format and structure of the output document. For example, you can take a collection of APT documents, let a Parser emit a series of Doxia events and have that be fed into a Sink to produce a single PDF, a book, a site, or a Word document. The Sink is fully responsible for the final output.
You can easily integrate any custom (XML, Wiki,...) format by creating a Doxia Parser which reads your input document and produces a proper sequence of Doxia events. Those can then be fed into an arbitrary Sink to produce any desired final output.
Note: All implemented sink should use UTF-8 as encoding.
@since 1.0-alpha-6 @author Jason van Zyl @author Vincent Siveton @author ltheussl @version $Id: Sink.java 905077 2010-01-31 17:24:47Z hboutemy $An output is something that can receive graph events. The output will send or transform the graph events in another form: a file, a network stream, a visualization, an algorithm, a metric, etc.
The output can filter the stream of attribute events using {@link AttributePredicate}s.
This listener is in fact the grouping of two specialized listeners. The first one listens only at structural changes in the graph (node and edge addition and removal). It is the {@link org.graphstream.stream.ElementSink}. The second one listens only at attributes values changes on elements of the graph (attribute addition, removal and change of values). It is the {@link org.graphstream.stream.AttributeSink}.
It is possible to listen only at elements or attributes changes with these two interfaces. Registering a graph listener will allow to listen at the both elements and attributes at the same time.
@see Source @see Pipe @see AttributeSink @see ElementSink | |
| |
| |