613-Support missing aggregation functions in aggregation kernels

[inikokali]

Issue
var is not supported for any of AggAll, AggRow, AggCol and stddev is not supported for full aggregation (AggAll) and row-wise aggregation (AggRow).

Implementation
This pull request proposes implementation for the above aggregation functions following the implementation of MEAN aggregation function in each instance.

Changes

• Changes in files AggAll.h, AggRow.h, AggCol.h related to the implementation of stddev and var.
• Updated file AggOpCode.h to add var opCode.
• Updated file kernels.json to add stddev and var opCodes where needed.
• Include testing scenarios to ensure the validity of my implementation. There are tests for both stddev and var functions for all three kernels (AggAll, AggRow, AggCol). The tests can be found in the related test cases ( AggAllTest.cpp, AggRowTest.cpp, AggCoTest.cpp ).
• Updated test cases matrix_agg.daphne and matrix_agg.py.

[pdamme]

Thanks for this contribution, @inikokali. With that, DAPHNE finally supports the aggregation functions SUM, MIN, MAX, MEAN, VAR, and STDDEV consistently for full, row-wise, and column-wise aggregation. Great!

Your code looks good to me and the tests pass. I only did some minor formatting and, more importantly, adapted the unit test cases to check for approximate equality of floating-point numbers. So far, the tiny example data in these test cases was chosen such that all results are "nice" numbers, without a lot of digits after the point. That's, of course, more difficult to construct for VAR and STDDEV. To avoid future test failures from insignificant floating-point deviations, we do approximate comparisons now.