io.kgraph.streaming.library

Class Spanner<K extends Comparable<K>,EV>

java.lang.Object

io.kgraph.streaming.SummaryAggregation<K,EV,S,T>

io.kgraph.streaming.SummaryBulkAggregation<K,EV,AdjacencyListGraph<K>,AdjacencyListGraph<K>>

io.kgraph.streaming.library.Spanner<K,EV>

Type Parameters:

K - the vertex ID type

EV - the edge value type

public class Spanner<K extends Comparable<K>,EV>
extends SummaryBulkAggregation<K,EV,AdjacencyListGraph<K>,AdjacencyListGraph<K>>

The Spanner library method continuously computes a k-Spanner of an insertion-only edge stream. The user-defined parameter k defines the distance estimation error, i.e. a k-spanner preserves all distances with a factor of up to k.

This is a single-pass implementation, which uses a SummaryBulkAggregation to periodically merge the partitioned state.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Spanner.CombineSpanners<K extends Comparable<K>> Merge the local spanners of each partition into the global spanner.
static class	Spanner.UpdateLocal<K extends Comparable<K>,EV> Decide to add or remove an edge to the local spanner in the current window.

Constructor Summary

Constructors

Constructor and Description
Spanner(long mergeWindowTime, int k) Creates a Spanner instance.

Method Summary

Methods inherited from class io.kgraph.streaming.SummaryBulkAggregation

run

Methods inherited from class io.kgraph.streaming.SummaryAggregation

aggregator, combineFun, initialValue, isTransientState, transform, updateFun

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Spanner

public Spanner(long mergeWindowTime,
               int k)

Creates a Spanner instance.

Parameters:

mergeWindowTime - Window time in millisec for the merger.

k - the distance error factor