spark Graph 的PregelAPI 理解和使用
图本质上是一种递归的数据结构,可以使用Spark GraphX 的PregelAPI接口对图数据进行批量计算,
之前一直不怎么理解Pregel计算模型,因此花点时间整理一下,该api的理解以及使用方法等。
1、Pregel的计算模型
Pregel接口的官方定义:
/**
* Execute a Pregel-like iterative vertex-parallel abstraction. The
* user-defined vertex-program `vprog` is executed in parallel on
* each vertex receiving any inbound messages and computing a new
* value for the vertex. The `sendMsg` function is then invoked on
* all out-edges and is used to compute an optional message to the
* destination vertex. The `mergeMsg` function is a commutative
* associative function used to combine messages destined to the
* same vertex.
*
* On the first iteration all vertices receive the `initialMsg` and
* on subsequent iterations if a vertex does not receive a message
* then the vertex-program is not invoked.
*
* This function iterates until there are no remaining messages, or
* for `maxIterations` iterations.
*
* @param A the Pregel message type
*
* @param initialMsg the message each vertex will receive at the on
* the first iteration
*
* @param maxIterations the maximum number of iterations to run for
*
* @param activeDirection the direction of edges incident to a vertex that received a message in
* the previous round on which to run `sendMsg`. For example, if this is `EdgeDirection.Out`, only
* out-edges of vertices that received a message in the previous round will run.
*
* @param vprog the user-defined vertex program which runs on each
* vertex and receives the inbound message and computes a new vertex
* value. On the first iteration the vertex program is invoked on
* all vertices and is passed the default message. On subsequent
* iterations the vertex program is only invoked on those vertices
* that receive messages.
*
* @param sendMsg a user supplied function that is applied to out
* edges of vertices that received messages in the current
* iteration
*
* @param mergeMsg a user supplied function that takes two incoming
* messages of type A and merges them into a single message of type
* A. ''This function must be commutative and associative and
* ideally the size of A should not increase.''
*
* @return the resulting graph at the end of the computation
*
*/
def pregel[A: ClassTag](
initialMsg: A,
maxIterations: Int = Int.MaxValue,
activeDirection: EdgeDirection = EdgeDirection.Either)(
vprog: (VertexId, VD, A) => VD,
sendMsg: EdgeTriplet[VD, ED] => Iterator[(VertexId, A)],
mergeMsg: (A, A) => A)
: Graph[VD, ED] = {
Pregel(graph, initialMsg, maxIterations, activeDirection)(vprog, sendMsg, mergeMsg)
}
方法的注释根据自己的实验理解如下:
执行类似Pregel的迭代顶点并行抽象。
在一次迭代计算中,图的各个顶点收到默认消息或者上一轮迭代发送的消息后;
首先调用mergeMsg函数将具有相同目的地的消息合并成一个消息;
然后调用vprog
顶点函数计算出新的顶点属性值;
然后再调用sendMsg
函数向出边顶点发送下一轮迭代的消息;
迭代计算直到没有消息剩余或者达到最大迭代次数退出。
在首轮迭代的时候,所有的顶点都会接收到initialMsg消息,在次轮迭代的时候,如果顶点没有接收到消息,verteProgram则不会被调用。
这些函数迭代会一直持续到没有剩余消息或者达到最大迭代次数maxIterations
。
VD : 顶点的属性的数据类型。
ED : 边的属性的数据类型
VertexId : 顶点ID的类型
A : Pregel message的类型。
graph:计算的输入的图
initialMsg : 图的每个顶点在首轮迭代时收到的初始化消息
maxIterations:最大迭代的次数
vprog:
vprog是用户定义的顶点程序,会运行在每一个顶点上,该vprog函数的功能是负责接收入站的message,
并计算出的顶点的新属性值。
在首轮迭代时,在所有的顶点上都会调用程序vprog函数,传人默认的defaultMessage;在次轮迭代时,只有接收到message消息的顶点才会调用vprog函数。
vprog: (VertexId, VD, A) => VD
输入参数: 顶点ID ,该顶点对应的顶点属性值,本轮迭代收到的message
输出结果: 新的顶点属性值
sendMsg:
用户提供的函数,应用于以当前迭代计算收到消息的顶点为源顶点的边edges;sendMsg函数的功能
是发送消息,消息的发送方向默认是沿着出边反向(向边的目的顶点发送消息)。
sendMsg: EdgeTriplet[VD, ED] => Iterator[(VertexId, A)],
输入参数是 EdgeTriplet :当前迭代计算收到消息的顶点为源顶点的边edges的EdgeTriplet对象。
输出结果: 下一迭代的消息。
mergeMsg:
用户提供定义的函数,将具有相同目的地的消息合并成一个;如果一个顶点,收到两个以上的A类型的消息message,该函数将他们合并成一个A类型消息。 这个函数必须是可交换的和关联的。理想情况下,A类型的message的size大小不应增加。
mergeMsg: (A, A) => A)
输入参数:当前迭代中,一个顶点收到的2个A类型的message。
输出结果:A类型的消息
下面的例子是使用Pregel计算单源最短路径,在图中节点间查找最短的路径是非常常见的图算法,所谓“单源最短路径”,就是指给定初始节点StartV,
计算图中其他任意节点到该节点的最短距离。我简化了官方的示例,使我们可以更简单的理解pregel计算模型。
package graphxTest
import org.apache.spark.rdd.RDD
import org.apache.spark.sql.SparkSession
import org.apache.spark.graphx.{Edge, Graph, VertexId}
/**
* Created by Mtime on 2018/1/25.
*/
object GraphxPregelTest {
val spark = SparkSession
.builder
.appName(s"${this.getClass.getSimpleName}").master("local[2]")
.getOrCreate()
val sc = spark.sparkContext
/**
* 计算最短路径
**/
def shortestPath(): Unit = {
//生成一个图对象
val graph: Graph[Long, Double] = genGraph
//打印出图的值
graph.triplets.foreach(t => {
println(s"t.srcId=${t.srcId} t.dstId=${t.dstId} t.srcAttr=${t.srcAttr} t.dstAttr=${t.dstAttr}")
})
val sourceId: VertexId = 1 // 计算顶点1到图各个顶点的最短路径
// Initialize the graph such that all vertices except the root have distance infinity.
val initialGraph = graph.mapVertices((id, att) =>
if (id == sourceId) 0.0 else Double.PositiveInfinity)
println("------------------------------")
//打印出图的值
initialGraph.triplets.foreach(t => {
println(s"t.srcId=${t.srcId} t.dstId=${t.dstId} t.srcAttr=${t.srcAttr} t.dstAttr=${t.dstAttr}")
})
val sssp:Graph[Double,Double] = initialGraph.pregel(Double.PositiveInfinity)(
(vid, vidAttr, message) => math.min(vidAttr, message), // Vertex Program
triplet => {
// Send Message
if (triplet.srcAttr + triplet.attr < triplet.dstAttr) {
Iterator((triplet.dstId, triplet.srcAttr + triplet.attr))
} else {
Iterator.empty
}
},
(message_a, message_b) => math.min(message_a, message_b) // Merge Message
)
println("------------------------------")
//打印出计算结果
println(sssp.vertices.collect.mkString("
"))
}
/**
* 初始化图对象
*
* @return
*/
private def genGraph(): Graph[Long, Double] = {
val vertices: RDD[(VertexId, Long)] =
sc.parallelize(Array(
(1L, 0L),
(2L, 0L),
(3L, 0L),
(4L, 0L),
(5L, 0L),
(6L, 0L))
)
// Create an RDD for edges
val edges: RDD[Edge[Double]] =
sc.parallelize(Array(
Edge(1L, 2L, 1.0),
Edge(1L, 4L, 1.0),
Edge(1L, 5L, 1.0),
Edge(2L, 3L, 1.0),
Edge(4L, 3L, 1.0),
Edge(5L, 4L, 1.0),
Edge(3L, 6L, 1.0)
)
)
val graph: Graph[Long, Double] = Graph(vertices, edges, 0)
graph
}
def main(args: Array[String]) {
shortestPath
}
}