Github user mengxr commented on a diff in the pull request:
https://github.com/apache/spark/pull/4047#discussion_r23980385
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/impl/PeriodicGraphCheckpointer.scala
---
@@ -0,0 +1,168 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.mllib.impl
+
+import scala.collection.mutable
+
+import org.apache.hadoop.fs.{Path, FileSystem}
+
+import org.apache.spark.Logging
+import org.apache.spark.graphx.Graph
+import org.apache.spark.storage.StorageLevel
+
+
+/**
+ * This class helps with persisting and checkpointing Graphs.
+ *
+ * This class maintains a FIFO queue of Graphs, each of which is persisted
and some of which are
+ * checkpointed. Once one Graph has been checkpointed, then previous RDDs
are unpersisted and their
+ * checkpoint files are removed.
+ *
+ * Users should call [[PeriodicGraphCheckpointer.updateGraph()]] when a
new graph has been created,
+ * before the graph has been materialized. When called, this does the
following:
+ * - Persist new graph (if not yet persisted), and put in queue of
persisted graphs.
+ * - Unpersist graphs from queue until there are at most 3 persisted
graphs.
+ * - If using checkpointing and the checkpoint interval has been reached,
+ * - Checkpoint the new graph, and put in a queue of checkpointed
graphs.
+ * - Remove older checkpoints.
+ *
+ * WARNINGS:
+ * - This class should NOT be copied (since copies may conflict on which
Graphs should be
+ * checkpointed).
+ * - This class removes checkpoint files once later graphs have been
checkpointed.
+ * However, references to the older graphs will still return
isCheckpointed = true.
+ *
+ * Example usage:
+ * val (graph1, graph2, graph3, ...) = ...
+ * val cp = new PeriodicGraphCheckpointer(graph, dir, 2)
+ * // persisted: graph1
+ * cp.updateGraph(graph2)
+ * // persisted: graph1, graph2
+ * // checkpointed: graph2
+ * cp.updateGraph(graph3)
+ * // persisted: graph1, graph2, graph3
+ * // checkpointed: graph2
+ * cp.updateGraph(graph4)
+ * // persisted: graph2, graph3, graph4
+ * // checkpointed: graph4
+ * cp.updateGraph(graph5)
+ * // persisted: graph3, graph4, graph5
+ * // checkpointed: graph4
+ *
+ * @param currentGraph Initial graph
+ * @param checkpointDir The directory for storing checkpoint files
+ * @param checkpointInterval Graphs will be checkpointed at this interval
+ * @tparam VD Vertex descriptor type
+ * @tparam ED Edge descriptor type
+ */
+private[mllib] class PeriodicGraphCheckpointer[VD, ED](
--- End diff --
This should work on normal RDDs as well. Let's put a TODO here. Otherwise,
this seems to be a little strange to live under `mllib`.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
