[SPARK-8392] RDDOperationGraph: getting cached nodes is slow

```def getAllNodes: Seq[RDDOperationNode] =
{ _childNodes ++ _childClusters.flatMap(_.childNodes) }```

when the ```_childClusters``` has so many nodes, the process will hang on. I think we can improve the efficiency here.

Author: xutingjun <[email protected]>

Closes #6839 from XuTingjun/DAGImprove and squashes the following commits:

53b03ea [xutingjun] change code to more concise and easier to read
f98728b [xutingjun] fix words: node -> nodes
f87c663 [xutingjun] put the filter inside
81f9fd2 [xutingjun] put the filter inside

Author: XuTingjun

core/src/main/scala/org/apache/spark/ui/UIUtils.scala

@@ -362,7 +362,7 @@ private[spark] object UIUtils extends Logging {
               { g.incomingEdges.map { e => <div class="incoming-edge">{e.fromId},{e.toId}</div> } }
               { g.outgoingEdges.map { e => <div class="outgoing-edge">{e.fromId},{e.toId}</div> } }
               {
-                g.rootCluster.getAllNodes.filter(_.cached).map { n =>
+                g.rootCluster.getCachedNodes.map { n =>
                   <div class="cached-rdd">{n.id}</div>
                 }
               }

core/src/main/scala/org/apache/spark/ui/scope/RDDOperationGraph.scala

@@ -66,9 +66,9 @@ private[ui] class RDDOperationCluster(val id: String, private var _name: String)
     _childClusters += childCluster
   }
 
-  /** Return all the nodes container in this cluster, including ones nested in other clusters. */
-  def getAllNodes: Seq[RDDOperationNode] = {
-    _childNodes ++ _childClusters.flatMap(_.childNodes)
+  /** Return all the nodes which are cached. */
+  def getCachedNodes: Seq[RDDOperationNode] = {
+    _childNodes.filter(_.cached) ++ _childClusters.flatMap(_.getCachedNodes)
   }
 }