JSON evaluator bug fixes

[joshuaphelpsms]

Purpose

This fixes a couple bugs:

1. When the ground truth had an array but the actual didn't have all values in the array, the total scores were calculated incorrectly
2. When the ground truth contained a key that didn't exist in the actual, the individual output for that field was not represented in the output data.
   Ex:

ground_truth = {"test1": "1", "test2": "2"}
actual = {"test1": "1"}
expected output = {"json_evaluator.CustomStringEvaluator.test1": 1, "json_evaluator.CustomStringEvaluator.test2": 0}
previous_output_before_fix =  {"json_evaluator.CustomStringEvaluator.test1": 1}

This also adds the ability to process a schema that is an array. It previously only handled JSON dicts as the initial input

Does this introduce a breaking change?

[ ] Yes
[x] No

Pull Request Type

What kind of change does this Pull Request introduce?

[x] Bugfix
[ ] Feature
[ ] Code style update (formatting, local variables)
[ ] Refactoring (no functional changes, no api changes)
[ ] Documentation content changes
[ ] Other... Please describe:

[kjeffc]

Thanks, was running into the bug where the missing fields in the actual data wasn't being taken into consideration when comparing against the ground truth, but this fixed that discrepancy I was seeing.