Adds persis_in, removes alloc_specs['in'], various small refactors

[jlnav]

No description provided.

[coveralls]

Pull Request Test Coverage Report for Build 1283944417

• 16 of 16 (100.0%) changed or added relevant lines in 7 files are covered.
• 176 unchanged lines in 6 files lost coverage.
• Overall coverage decreased (-1.8%) to 91.257%

---

Files with Coverage Reduction	New Missed Lines	%
libensemble/gen_funcs/persistent_deap_nsga2.py	1	98.35%
libensemble/gen_funcs/persistent_surmise_calib.py	1	97.59%
libensemble/tools/consensus_subroutines.py	4	95.5%
libensemble/tools/gen_support.py	21	0%
libensemble/gen_funcs/persistent_gp.py	61	0%
libensemble/executors/balsam_executor.py	88	34.36%

Totals
Change from base Build 1270398337:	-1.8%
Covered Lines:	6113
Relevant Lines:	6615

---

💛 - Coveralls

[shuds13]

@jmlarson1 @jlnav Do we want to construct fields to give back like this - or expect the user to do it in calling script when creating gen_specs['persis_in'] ? An alloc could support a default, if the user does not set persis_in maybe, but seems to me potentially confusing to mix user supplied and add to it in alloc.

[jlnav]

Yeah, I'd lean towards expecting the user to define their fields in their calling script. I think I was more explicit with all the other cases except the new functions from the student.

Any major objections to refactoring the above conditional fields to be in their corresponding calling scripts instead?

[jmlarson1]

I also think that we should expect the user to define their fields in their calling script.

[shuds13]

Sounds good to me.