Exposed maximum chunk size in CNV panel of normals.

[samuelklee]

This enables PoNs to be built with a number of intervals greater than 16777216, which lets us push down to bin sizes <175bp if really necessary. At some point, it is better to scatter and build a separate PoN for each contig so we don't run into memory issues. However, the official WDL is not set up to do the scatter, so one would have to write a custom WDL.

@mwalker174 Do you mind taking a look?

Closes #4365.

[codecov-io]

Codecov Report

Merging #4528 into master will increase coverage by 0.003%.
The diff coverage is 85.714%.

@@               Coverage Diff               @@
##              master     #4528       +/-   ##
===============================================
+ Coverage     79.119%   79.122%   +0.003%     
  Complexity     16677     16677               
===============================================
  Files           1051      1051               
  Lines          60285     60289        +4     
  Branches        9875      9875               
===============================================
+ Hits           47697     47702        +5     
+ Misses          8760      8759        -1     
  Partials        3828      3828

Impacted Files	Coverage Δ	Complexity Δ
...mber/denoising/HDF5SVDReadCountPanelOfNormals.java	87.218% <100%> (ø)	37 <1> (ø)	⬇️
...ools/copynumber/CreateReadCountPanelOfNormals.java	88.889% <75%> (-1.111%)	9 <0> (ø)
...e/hellbender/tools/copynumber/utils/HDF5Utils.java	78.947% <83.333%> (+1.053%)	8 <0> (ø)	⬇️
...utils/smithwaterman/SmithWatermanIntelAligner.java	90% <0%> (+10%)	3% <0%> (ø)	⬇️

[samuelklee]

@mwalker174 Can you take a quick look?

[samuelklee]

@mwalker174 Let me know if I should assign this to someone else. Thanks!

[mwalker174]

Looks good. Somehow I didn't see these mentions, sorry! This has motivated me to finally set up github filters in my email.