Unused args in VCF merge

[davidbenjamin]

@lbergelson The merge vcfs code is notoriously hard to understand. Removing a bunch of arguments that the GATK never uses seems like a start.

[codecov-io]

Codecov Report

Merging #5745 into master will increase coverage by 0.028%.
The diff coverage is 100%.

@@               Coverage Diff               @@
##              master     #5745       +/-   ##
===============================================
+ Coverage     87.079%   87.107%   +0.028%     
- Complexity     31918     31998       +80     
===============================================
  Files           1942      1943        +1     
  Lines         146909    147126      +217     
  Branches       16242     16268       +26     
===============================================
+ Hits          127927    128157      +230     
+ Misses         13062     13050       -12     
+ Partials        5920      5919        -1

Impacted Files	Coverage Δ	Complexity Δ
...utils/variant/GATKVariantContextUtilsUnitTest.java	85.745% <ø> (-0.14%)	159 <0> (-4)
...walkers/genotyper/GenotypingGivenAllelesUtils.java	75% <ø> (ø)	5 <0> (ø)	⬇️
...kers/haplotypecaller/AssemblyBasedCallerUtils.java	80.303% <ø> (ø)	41 <0> (ø)	⬇️
...lbender/utils/variant/GATKVariantContextUtils.java	87.615% <100%> (+2.723%)	246 <0> (-10)	⬇️
...ellbender/tools/funcotator/FilterFuncotations.java	92.857% <0%> (-0.476%)	30% <0%> (+12%)
...ools/funcotator/filtrationRules/ClinVarFilter.java	100% <0%> (ø)	13% <0%> (+6%)	⬆️
.../funcotator/FilterFuncotationsIntegrationTest.java	100% <0%> (ø)	7% <0%> (+2%)	⬆️
...er/tools/funcotator/filtrationRules/LofFilter.java	100% <0%> (ø)	10% <0%> (+5%)	⬆️
...r/filtrationRules/FilterFuncotationsExacUtils.java	81.818% <0%> (ø)	12% <0%> (?)
...ls/walkers/mutect/CreateSomaticPanelOfNormals.java	94.898% <0%> (+0.868%)	27% <0%> (+5%)	⬆️
... and 3 more

[davidbenjamin]

@jamesemery Would you mind taking a look at this bit of code maintenance?

[jamesemery]

Nice finding these unused paths. Looking back it appears that as far back as gatk3 most if this functionally appears to be unused (even in testing). While I would say that some of this functionality seems like it may be useful, the more advanced use cases are already managed by referenceConfidenceVariantContextMerger. Some day we should unify code between the two paths, but this step if anything will help illuminate what is important in such a merge.

[davidbenjamin]

Thanks @jamesemery!