remove useless GridPartitioner in package ml.optim
Usage
Use mvn package command, will generate two jars:
spark-vlbfgs-1.0-SNAPSHOT.jarspark-vlbfgs-1.0-SNAPSHOT-tests.jar
Running the test suites, use the following command:
bin/spark-submit
--driver-memory 3G
--class org.scalatest.run #use scalatest.run to trigger the test suite running
--jars ... #(add path to these jars: spark-core_2.11-2.1.0-SNAPSHOT-tests.jar, spark-mllib_2.11-2.1.0-SNAPSHOT-tests.jar, spark-mllib-local_2.11-2.1.0-SNAPSHOT-tests.jar, scalatest_2.11-2.2.6.jar, spark-vlbfgs-1.0-SNAPSHOT.jar)
PATH/TO/spark-vlbfgs-1.0-SNAPSHOT-tests.jar
org.apache.spark.ml.classification.VLogisticRegressionSuite #the test suite class need to run
What changes were proposed in this pull request?
<goal>test-jar</goal>
so thatmvn package
will generate tests-jar.spark-graphx_2.11
MLlibTestSparkContext
,TestingUtils
andTempDirectory
which copied from spark test source.SparkFunSuite
subclass, instead ofFunSuite
. So that it will keep consistent with spark project.VLogisticRegressionSuite
illegaltoLocal
usage.GridPartitioner
in packageml.optim
Usage
Use
mvn package
command, will generate two jars:spark-vlbfgs-1.0-SNAPSHOT.jar
spark-vlbfgs-1.0-SNAPSHOT-tests.jar
Running the test suites, use the following command: