Github user mengxr commented on a diff in the pull request:
https://github.com/apache/spark/pull/1367#discussion_r14855841
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/stat/correlation/Correlation.scala
---
@@ -0,0 +1,121 @@
+/*
+ * 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.stat.correlation
+
+import org.apache.spark.mllib.linalg.{DenseVector, Matrix, Vector}
+import org.apache.spark.rdd.RDD
+
+/**
+ * New correlation algorithms should implement this trait
+ */
+trait Correlation {
+
+ /**
+ * Compute correlation for two datasets.
+ */
+ def computeCorrelation(x: RDD[Double], y: RDD[Double]): Double
+
+ /**
+ * Compute the correlation matrix S, for the input matrix, where S(i, j)
is the correlation
+ * between column i and j.
+ */
+ def computeCorrelationMatrix(X: RDD[Vector]): Matrix
+
+ /**
+ * Combine the two input RDD[Double]s into an RDD[Vector] and compute
the correlation using the
+ * correlation implementation for RDD[Vector]
+ */
+ def computeCorrelationWithMatrixImpl(x: RDD[Double], y: RDD[Double]):
Double = {
+ val mat: RDD[Vector] = x.zip(y).mapPartitions({ iter =>
+ iter.map {case(xi, yi) => new DenseVector(Array(xi, yi))}
+ }, preservesPartitioning = true)
+ computeCorrelationMatrix(mat)(0, 1)
+ }
+
+}
+
+/**
+ * Delegates computation to the specific correlation object based on the
input method name
+ *
+ * Currently supported correlations: pearson, spearman.
+ * After new correlation algorithms are added, please update the
documentation here and in
+ * Statistics.scala for the correlation APIs.
+ *
+ * Cases are ignored when doing method matching. We also allow initials,
e.g. "P" for "pearson", as
+ * long as initials are unique in the supported set of correlation
algorithms. In addition, a
--- End diff --
It is not just for low maintenance cost. A strict contract also helps users
understand exactly what to input and what to expect. For the fault tolerance,
there are many ways to break Spark by putting an extra "s" somewhere. Let's not
worry too much in this PR. If we find a mismatch at runtime, throw an error and
list available methods.
Btw, I like git's solution. You cannot use "git commits" for "git commit"
but git will suggest:
~~~
Did you mean this?
commit
~~~
Anyway, let's not spend time on this auto-suggestion feature either. Please
update the PR with strict string matching. Thanks!
---
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.
---
