fix: Wrapper crashing if dirs not clean

[xiamaz]

Closes #443

Not elegant but will always avoid the crash

[coveralls]

coverage: 85.866%. remained the same
when pulling df94bb2 on 443-contig-ploidy-fix
into 4874074 on main.

[Nicolai-vKuegelgen]

I have also started work on a fix for this (see branch 443-gcnv_contig_ploidy-wrapper-script-crahses-because-gatk-determinegermlinecontigploidy-output-is-not-cleaned-but-always-fully-read).

I see some issues with your solution:

1. Skipping unknown samples silently is not very future proof, since we won't get any clear error message if somehow the sample names output of gcnv doesn't match the samplesheet anymore.
2. If the additional sample folders are not removed it is possible that later GCNV steps will pick them up again (i.e. call_cnvs which reads in the output dir of the contig_ploidy step).
3. Applying this fix with the same expectations for the case mode and the corohort mode for gcnv is maybe not correct, since the 'usual' problem of re-running snappy in the same project with a smaller set of (new) samples should only apply to case-mode.

[xiamaz]

Can your fix be merged? I'm impartial as long as behavior for germline gcnv pipelines does not require one to clean the output directories.

[Nicolai-vKuegelgen]

My fix could be merged right now (I've tested it & it works) - will make a pull request.
The only reliable & clean solution I could come up with is to have the affected wrapper scripts remove the directories they will write into. Leaving old files there is problematic since GCNV often just requires a base path if i.e. the contigs and will then read whatever is there.

[Nicolai-vKuegelgen]

Should be fixed by #468 now, so I'm closing this