[SPARK-23256][ML][PYTHON] Add columnSchema method to PySpark image reader

[HyukjinKwon]

What changes were proposed in this pull request?

This PR proposes to add columnSchema in Python side too.

>>> from pyspark.ml.image import ImageSchema
>>> ImageSchema.columnSchema.simpleString()
'struct<origin:string,height:int,width:int,nChannels:int,mode:int,data:binary>'

How was this patch tested?

Manually tested and unittest was added in python/pyspark/ml/tests.py.

[HyukjinKwon]

@MrBago, @BryanCutler, @imatiach-msft, and @MLnick, could you take a look please?

[HyukjinKwon]

I am fine with 2.4.0. Let me know.

[BryanCutler]

Since this came out of the 2.3 ml QA and it's mostly an improvement to the python API, I think maybe 2.4 is best. But it is a new API do maybe it's ok to include in 2.3..

[HyukjinKwon]

Yea, let me go with 2.4.0.

[SparkQA]

Test build #86932 has finished for PR 20475 at commit e180ade.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[imatiach-msft]

@HyukjinKwon looks like a great change to me, thank you for exposing the method in pyspark

[imatiach-msft]

nice method!

[dongjoon-hyun]

+1, LGTM.

[BryanCutler]

LGTM, thanks @HyukjinKwon!

[BryanCutler]

Since this came out of the 2.3 ml QA and it's mostly an improvement to the python API, I think maybe 2.4 is best. But it is a new API do maybe it's ok to include in 2.3..

[SparkQA]

Test build #86996 has finished for PR 20475 at commit ffa1b48.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[HyukjinKwon]

Thank you @imatiach-msft, @dongjoon-hyun, @felixcheung and @BryanCutler.

Merged to master only.