ODIn/spark-instrumented-optimizer
1
0
Fork 0
Code Issues Pull requests Projects Releases Wiki Activity

refactor and fix bugs

Browse source
This commit is contained in:
Jianping J Wang 2013-12-31 16:59:05 +08:00
parent 600421d8bc
commit 779c66ae4e
1 changed files with 29 additions and 58 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

87
graph/src/main/scala/org/apache/spark/graph/algorithms/Svdpp.scala
View file

@ -5,18 +5,6 @@ import org.apache.spark.graph._
import scala.util.Random import scala.util.Random
import org.apache.commons.math.linear._ import org.apache.commons.math.linear._
class VT( // vertex type
var v1: RealVector, // v1: p for user node, q for item node
var v2: RealVector, // v2: pu + |N(u)|^(-0.5)*sum(y) for user node, y for item node
var bias: Double,
var norm: Double // |N(u)|^(-0.5) for user node
) extends Serializable
class Msg( // message
var v1: RealVector,
var v2: RealVector,
var bias: Double) extends Serializable
class SvdppConf( // Svdpp parameters class SvdppConf( // Svdpp parameters
var rank: Int, var rank: Int,
var maxIters: Int, var maxIters: Int,
@ -40,92 +28,75 @@ object Svdpp {
* @return a graph with vertex attributes containing the trained model * @return a graph with vertex attributes containing the trained model
*/ */
def run(edges: RDD[Edge[Double]], conf: SvdppConf): Graph[VT, Double] = { def run(edges: RDD[Edge[Double]], conf: SvdppConf): Graph[(RealVector, RealVector, Double, Double), Double] = {
// generate default vertex attribute // generate default vertex attribute
def defaultF(rank: Int) = { def defaultF(rank: Int): (RealVector, RealVector, Double, Double) = {
val v1 = new ArrayRealVector(rank) val v1 = new ArrayRealVector(rank)
val v2 = new ArrayRealVector(rank) val v2 = new ArrayRealVector(rank)
for (i <- 0 until rank) { for (i <- 0 until rank) {
v1.setEntry(i, Random.nextDouble) v1.setEntry(i, Random.nextDouble)
v2.setEntry(i, Random.nextDouble) v2.setEntry(i, Random.nextDouble)
} }
var vd = new VT(v1, v2, 0.0, 0.0) (v1, v2, 0.0, 0.0)
vd
} }
// calculate global rating mean // calculate global rating mean
val (rs, rc) = edges.map(e => (e.attr, 1L)).reduce((a, b) => (a._1 + b._1, a._2 + b._2)) val (rs, rc) = edges.map(e => (e.attr, 1L)).reduce((a, b) => (a._1 + b._1, a._2 + b._2))
val u = rs / rc // global rating mean val u = rs / rc
// construct graph // construct graph
var g = Graph.fromEdges(edges, defaultF(conf.rank)).cache() var g = Graph.fromEdges(edges, defaultF(conf.rank)).cache()
// calculate initial bias and norm // calculate initial bias and norm
var t0: VertexRDD[(Long, Double)] = g.mapReduceTriplets(et => var t0 = g.mapReduceTriplets(et =>
Iterator((et.srcId, (1L, et.attr)), (et.dstId, (1L, et.attr))), Iterator((et.srcId, (1L, et.attr)), (et.dstId, (1L, et.attr))), (g1: (Long, Double), g2: (Long, Double)) => (g1._1 + g2._1, g1._2 + g2._2))
(g1: (Long, Double), g2: (Long, Double)) => g = g.outerJoinVertices(t0) { (vid: Vid, vd: (RealVector, RealVector, Double, Double), msg: Option[(Long, Double)]) =>
(g1._1 + g2._1, g1._2 + g2._2)) (vd._1, vd._2, msg.get._2 / msg.get._1, 1.0 / scala.math.sqrt(msg.get._1))
g = g.outerJoinVertices(t0) {
(vid: Vid, vd: VT, msg: Option[(Long, Double)]) =>
vd.bias = msg.get._2 / msg.get._1; vd.norm = 1.0 / scala.math.sqrt(msg.get._1)
vd
} }
def mapTrainF(conf: SvdppConf, u: Double)(et: EdgeTriplet[VT, Double]): Iterator[(Vid, Msg)] = { def mapTrainF(conf: SvdppConf, u: Double)(et: EdgeTriplet[(RealVector, RealVector, Double, Double), Double])
assert(et.srcAttr != null && et.dstAttr != null) : Iterator[(Vid, (RealVector, RealVector, Double))] = {
val (usr, itm) = (et.srcAttr, et.dstAttr) val (usr, itm) = (et.srcAttr, et.dstAttr)
val (p, q) = (usr.v1, itm.v1) val (p, q) = (usr._1, itm._1)
var pred = u + usr.bias + itm.bias + q.dotProduct(usr.v2) var pred = u + usr._3 + itm._3 + q.dotProduct(usr._2)
pred = math.max(pred, conf.minVal) pred = math.max(pred, conf.minVal)
pred = math.min(pred, conf.maxVal) pred = math.min(pred, conf.maxVal)
val err = et.attr - pred val err = et.attr - pred
val updateP = ((q.mapMultiply(err)).subtract(p.mapMultiply(conf.gamma7))).mapMultiply(conf.gamma2) val updateP = ((q.mapMultiply(err)).subtract(p.mapMultiply(conf.gamma7))).mapMultiply(conf.gamma2)
val updateQ = ((usr.v2.mapMultiply(err)).subtract(q.mapMultiply(conf.gamma7))).mapMultiply(conf.gamma2) val updateQ = ((usr._2.mapMultiply(err)).subtract(q.mapMultiply(conf.gamma7))).mapMultiply(conf.gamma2)
val updateY = ((q.mapMultiply(err * usr.norm)).subtract((itm.v2).mapMultiply(conf.gamma7))).mapMultiply(conf.gamma2) val updateY = ((q.mapMultiply(err * usr._4)).subtract((itm._2).mapMultiply(conf.gamma7))).mapMultiply(conf.gamma2)
Iterator((et.srcId, new Msg(updateP, updateY, (err - conf.gamma6 * usr.bias) * conf.gamma1)), Iterator((et.srcId, (updateP, updateY, (err - conf.gamma6 * usr._3) * conf.gamma1)),
(et.dstId, new Msg(updateQ, updateY, (err - conf.gamma6 * itm.bias) * conf.gamma1))) (et.dstId, (updateQ, updateY, (err - conf.gamma6 * itm._3) * conf.gamma1)))
} }
for (i <- 0 until conf.maxIters) { for (i <- 0 until conf.maxIters) {
// phase 1, calculate v2 for user nodes // phase 1, calculate v2 for user nodes
var t1 = g.mapReduceTriplets(et => var t1 = g.mapReduceTriplets(et => Iterator((et.srcId, et.dstAttr._2)), (g1: RealVector, g2: RealVector) => g1.add(g2))
Iterator((et.srcId, et.dstAttr.v2)), g = g.outerJoinVertices(t1) { (vid: Vid, vd: (RealVector, RealVector, Double, Double), msg: Option[RealVector]) =>
(g1: RealVector, g2: RealVector) => g1.add(g2)) if (msg.isDefined) (vd._1, vd._1.add(msg.get.mapMultiply(vd._4)), vd._3, vd._4) else vd
g = g.outerJoinVertices(t1) { (vid: Vid, vd: VT, msg: Option[RealVector]) =>
if (msg.isDefined) vd.v2 = vd.v1.add(msg.get.mapMultiply(vd.norm))
vd
} }
// phase 2, update p for user nodes and q, y for item nodes // phase 2, update p for user nodes and q, y for item nodes
val t2: VertexRDD[Msg] = g.mapReduceTriplets(mapTrainF(conf, u), (g1: Msg, g2: Msg) => { val t2 = g.mapReduceTriplets(mapTrainF(conf, u), (g1: (RealVector, RealVector, Double), g2: (RealVector, RealVector, Double)) =>
g1.v1 = g1.v1.add(g2.v1) (g1._1.add(g2._1), g1._2.add(g2._2), g1._3 + g2._3))
g1.v2 = g1.v2.add(g2.v2) g = g.outerJoinVertices(t2) { (vid: Vid, vd: (RealVector, RealVector, Double, Double), msg: Option[(RealVector, RealVector, Double)]) =>
g1.bias += g2.bias (vd._1.add(msg.get._1), vd._2.add(msg.get._2), vd._3 + msg.get._3, vd._4)
g1
})
g = g.outerJoinVertices(t2) { (vid: Vid, vd: VT, msg: Option[Msg]) =>
vd.v1 = vd.v1.add(msg.get.v1)
if (vid % 2 == 1) vd.v2 = vd.v2.add(msg.get.v2)
vd.bias += msg.get.bias
vd
} }
} }
// calculate error on training set // calculate error on training set
def mapTestF(conf: SvdppConf, u: Double)(et: EdgeTriplet[VT, Double]): Iterator[(Vid, Double)] = { def mapTestF(conf: SvdppConf, u: Double)(et: EdgeTriplet[(RealVector, RealVector, Double, Double), Double]): Iterator[(Vid, Double)] = {
assert(et.srcAttr != null && et.dstAttr != null)
val (usr, itm) = (et.srcAttr, et.dstAttr) val (usr, itm) = (et.srcAttr, et.dstAttr)
val (p, q) = (usr.v1, itm.v1) val (p, q) = (usr._1, itm._1)
var pred = u + usr.bias + itm.bias + q.dotProduct(usr.v2) var pred = u + usr._3 + itm._3 + q.dotProduct(usr._2)
pred = math.max(pred, conf.minVal) pred = math.max(pred, conf.minVal)
pred = math.min(pred, conf.maxVal) pred = math.min(pred, conf.maxVal)
val err = (et.attr - pred) * (et.attr - pred) val err = (et.attr - pred) * (et.attr - pred)
Iterator((et.dstId, err)) Iterator((et.dstId, err))
} }
val t3: VertexRDD[Double] = g.mapReduceTriplets(mapTestF(conf, u), _ + _) val t3 = g.mapReduceTriplets(mapTestF(conf, u), (g1: Double, g2: Double) => g1 + g2)
g.outerJoinVertices(t3) { (vid: Vid, vd: VT, msg: Option[Double]) => g.outerJoinVertices(t3) { (vid: Vid, vd: (RealVector, RealVector, Double, Double), msg: Option[Double]) =>
if (msg.isDefined && vid % 2 == 1) vd.norm = msg.get // item nodes sum up the errors if (msg.isDefined && vid % 2 == 1) (vd._1, vd._2, vd._3, msg.get) else vd
vd
} }
g g
} }