You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
vcztools view needs to produce the appropriate header in all situations, and having an input header as an option is not really very helpful. We need to subset the samples and fields in all sorts of ways, and parsing the source header here is not the right approach. Header parsing should be done in bio2zarr.
Note that do this fully, we will need to update the VCF Zarr specification and vcf2zarr in order to store sufficient information about the header, rather than the header itself. This will require some coordinated updates across repos.
I think the simplest approach here is to just delete the option for proving a complete header here, and see what we're missing.
See the linked issues in sgkit-dev/vcf-zarr-spec#15 (comment) for notes on how we might avoid storing the full header, and retain lossless encoding of the information in it.
vcztools view needs to produce the appropriate header in all situations, and having an input header as an option is not really very helpful. We need to subset the samples and fields in all sorts of ways, and parsing the source header here is not the right approach. Header parsing should be done in bio2zarr.
Note that do this fully, we will need to update the VCF Zarr specification and vcf2zarr in order to store sufficient information about the header, rather than the header itself. This will require some coordinated updates across repos.
I think the simplest approach here is to just delete the option for proving a complete header here, and see what we're missing.
See:
vcf_header
attribute and/or refine? vcf-zarr-spec#15The text was updated successfully, but these errors were encountered: