Refactor hybrid query pagination logic

[vibrantvarun]

Refactor hybrid query pagination logic to be consistent.

1. Think of combining version check and objects.nonnull method in doXContent to define the existence of the field in the query builder.
2. Add IndexUtil check suggested by @ryanbogan in Update pagination_depth datatype from int to Integer #1094 (comment)

[vibrantvarun]

@martin-gaievski I tried to fit all the minclustercheck and isnull check of doXContent method under one but it breaks bwc in the edge cases.

[vibrantvarun]

I did extensive testing and concluded that it is better now compared to one method for all conditions. Regarding your concern of people making the change unknowingly, we can add comments and also we are following the same pattern of put checks in StreamInput, StreamOutput and fromXContent. It won't be an issue.

[martin-gaievski]

can you be more specific, what is the edge case that you couldn't handle properly in your local change? I don't think such code is not possible, it can be not the best from implementation or code organization point of view. You can do draft PR, this will make discussion more productive

[heemin32]

@vibrantvarun Could you elaborate more? Especially for Think of combining version check and objects.nonnull method in doXContent to define the existence of the field in the query builder..

Also, what is not consistent with pagination? Refactor hybrid query pagination logic to be consistent.

[vibrantvarun]

I think we are good now and don't need to do anything more here.