Internal VCF sort during stvk standardize

[RCollins13]

Reposting the contents of our previous conversation on slack:

Based on what I can tell, stvk standardize flips CHR1/CHR2 for BND records in the VCF to have the numerically-sorted smaller chromosome assigned to the CHR field in the standardized record, but then doesn’t appear to re-sort the records afterwards. Internalizing the re-sorting of the vcf would be preferred, because otherwise it breaks tabix.

For example, I have a VCF with calls only on three chromosomes (chr8, 11, and 19), some of which are BNDs connected to other chromosomes (like, for example, a chr1-chr19 BND). I make sure to sort my VCF when I’m filtering it down to the three chromosomes of interest, but then when I run it through standardize the order becomes unsorted.

I've attached an example VCF before and after standardization here:
G94051_VIR_00001.manta.standardized.vcf.gz
G94051_VIR_00001.manta.vcf.gz

In particular, there are two offending records:

19      10966268        G94051_VIR_00001_29     N       <DEL>   .       PASS    SVTYPE=DEL;CHR2=19;STRANDS=+-;SVLEN=3854;ALGORITHMS=manta       GT:manta        0/1:1
17      8069464 G94051_VIR_00001_30     N       ]19:11060451]N  .       PASS    SVTYPE=BND;CHR2=19;STRANDS=-+;SVLEN=-1;ALGORITHMS=manta GT:manta        0/1:1
19      11294842        G94051_VIR_00001_31     N       <DEL>   .       PASS    SVTYPE=DEL;CHR2=19;STRANDS=+-;SVLEN=1663;ALGORITHMS=manta       GT:manta        0/1:1
1       38241383        G94051_VIR_00001_32     N       N[19:11422016[  .       PASS    SVTYPE=BND;CHR2=19;STRANDS=+-;SVLEN=-1;ALGORITHMS=manta GT:manta        0/1:1
19      11535419        G94051_VIR_00001_33     N       <DEL>   .       PASS    SVTYPE=DEL;CHR2=19;STRANDS=+-;SVLEN=153;ALGORITHMS=manta        GT:manta        0/1:1

Thanks,
Ryan