[SPARK-11938][PySpark] Expose numFeatures in all ML PredictionModel for PySpark

[vectorijk]

What changes were proposed in this pull request?

JIRA: https://issues.apache.org/jira/browse/SPARK-11938

SPARK-9715 provided support for numFeatures in all ML PredictionModels, we should also expose it at PySpark.

How was this patch tested?

Unit Test for PySpark

[vectorijk]

cc @jkbradley @yanboliang @Lewuathe

[SparkQA]

Test build #61295 has finished for PR 13922 at commit 872d384.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
  • class LogisticRegressionModel(JavaModel, JavaMLWritable, JavaMLReadable):
  • class DecisionTreeClassificationModel(DecisionTreeModel, JavaMLWritable, JavaMLReadable):
  • class RandomForestClassificationModel(TreeEnsembleModels, JavaMLWritable, JavaMLReadable):
  • class GBTClassificationModel(TreeEnsembleModels, JavaMLWritable, JavaMLReadable):
  • class NaiveBayesModel(JavaModel, JavaMLWritable, JavaMLReadable):
  • class MultilayerPerceptronClassificationModel(JavaModel, JavaMLWritable, JavaMLReadable):
  • class LinearRegressionModel(JavaModel, JavaMLWritable, JavaMLReadable):
  • class DecisionTreeRegressionModel(DecisionTreeModel, JavaMLWritable, JavaMLReadable):
  • class RandomForestRegressionModel(TreeEnsembleModels, JavaMLWritable, JavaMLReadable):
  • class GBTRegressionModel(TreeEnsembleModels, JavaMLWritable, JavaMLReadable):

[lins05]

Seems the numFeatures property is repeated in all these classification and regression classes. What about adding a HasNumFeatures mixin class and move the numFeatures property implementation there?

[MLnick]

@vectorijk this is covered in #12889 by @holdenk

[vectorijk]

@MLnick Thanks! I will close this PR.