Move ZeppelinContext code to Scala

[doanduyhai]

This is the follow up of #80

Warning: this PR should be merged before https://issues.apache.org/jira/browse/ZEPPELIN-148

The source code of ZeppelinContext has been converted to Scala to make the integration easier.

Now it is possible to do:

z.show(rdd)

z.display(rdd,"Header1","Header2",...)

z.display(rdd, limit, "Header1","Header2",...)

z.html(...)

z.img(...)

z.img64(...)

Of course, all the implicits developed in the previous pull request are also back-ported

I also added unit tests to have a decent code coverage

Example of z.display() with limit manually set to 2:

Example of z.display() using default limit (zeppelin.spark.maxResult = 100):

Example of calling display() directly on RDD with limit set to 3:

Example of calling display() directly on a Scala traversable:

Example of calling html():

We can also call z.html() on the ZeppelinContext instance itself:

Example of calling img(url):

Of course, calling z.img(url) is also possible. The same remarks also apply to img64(...) and z.img64(...)

Last but not least, z.show(...) has been refactored to display SchemaRDD/DataFrame

[jongyoul]

Please fix this incorrect indentation.

[corneadoug]

@doanduyhai
I'm going around old PRs,
What should we do about that one?