Fix aligned PathSeq input getting filtered by WellformedReadFilter

[mwalker174]

PathSeqFilterSpark and PathSeqPipelineSpark clear all the sequences from the input header file, as the Bwa step only accepts unaligned reads. However, the header sequences were being cleared before the reads were loaded, causing WellformedReadFilter to remove any mapped reads (by failing to find the corresponding sequence name in the header).

This PR fixes this bug by creating a deep copy of the header. It also refactors this code, which is used in both the Filter and Pipeline tools, into a utility function checkAndClearHeaderSequences() in PSUtils. Tests have also been added/updated accordingly.

[codecov-io]

Codecov Report

Merging #3453 into master will increase coverage by 0.008%.
The diff coverage is 75%.

@@              Coverage Diff               @@
##              master    #3453       +/-   ##
==============================================
+ Coverage     80.522%   80.53%   +0.008%     
- Complexity     17612    17613        +1     
==============================================
  Files           1176     1176               
  Lines          63617    63617               
  Branches        9903     9901        -2     
==============================================
+ Hits           51226    51231        +5     
+ Misses          8435     8432        -3     
+ Partials        3956     3954        -2

Impacted Files	Coverage Δ	Complexity Δ
...bender/tools/spark/pathseq/PathSeqFilterSpark.java	80.769% <100%> (+9.801%)	5 <0> (-2)	⬇️
...nder/tools/spark/pathseq/PathSeqPipelineSpark.java	89.474% <100%> (+5.603%)	7 <0> (ø)	⬇️
...titute/hellbender/tools/spark/pathseq/PSUtils.java	75% <70%> (-2.778%)	15 <3> (+3)

[SHuang-Broad]

Looks good to me. 👍