add jvm import org.apache.spark.sql.api.python.*

[jinhaichen]

What changes were proposed in this pull request?

jvm import org.apache.spark.sql.api.python.*. Or else pandas dataframe to spark dataframe will fail.

How was this patch tested?

import pandas as pd
from pyspark.sql import SparkSession

pandas_df = pd.DataFrame({
    'A': [1, 2, 3],
    'B': ['a', 'b', 'c']
})

spark = SparkSession.builder.getOrCreate()
spark_df = spark.createDataFrame(pandas_df)
spark_df.show()

this failed before, and now it's OK.

[jinhaichen]

any one check here?