[SPARK-19270][FOLLOW-UP][ML] PySpark GLR model.summary should return a printable representation.

## What changes were proposed in this pull request?
PySpark GLR ```model.summary``` should return a printable representation by calling Scala ```toString```.

## How was this patch tested?
```
from pyspark.ml.regression import GeneralizedLinearRegression
dataset = spark.read.format("libsvm").load("data/mllib/sample_linear_regression_data.txt")
glr = GeneralizedLinearRegression(family="gaussian", link="identity", maxIter=10, regParam=0.3)
model = glr.fit(dataset)
model.summary
```
Before this PR:
![image](https://user-images.githubusercontent.com/1962026/29021059-e221633e-7b96-11e7-8d77-5d53f89c81a9.png)
After this PR:
![image](https://user-images.githubusercontent.com/1962026/29021097-fce80fa6-7b96-11e7-8ab4-7e113d447d5d.png)

Author: Yanbo Liang <ybliang8@gmail.com>

Closes #18870 from yanboliang/spark-19270.

python/pyspark/ml/regression.py

@@ -1745,6 +1745,9 @@ class GeneralizedLinearRegressionTrainingSummary(GeneralizedLinearRegressionSumm
         """
         return self._call_java("pValues")
 
+    def __repr__(self):
+        return self._call_java("toString")
+
 
 if __name__ == "__main__":
     import doctest