Regulatory annotation download link is hardcoded

[julia-luz]

Hello!

cyrcular-calling/workflow/rules/ref.smk

Line 74 in ccd556a

"""wget https://ftp.ensembl.org/pub/release-{params.release}/regulation/homo_sapiens/homo_sapiens.GRCh38.Regulatory_Build.regulatory_features.20220201.gff.gz --no-check-certificate -O {output} 2> {log}"""

The number between regulatory_features and .gff.gz in the url varies between releases. Attempting to wget https://ftp.ensembl.org/pub/release-{params.release}/regulation/homo_sapiens/homo_sapiens.GRCh38.Regulatory_Build.regulatory_features.20220201.gff.gz if the release version is set to 110 on the config file will cause this to fail, since the proper url for release 110 would be https://ftp.ensembl.org/pub/release-110/regulation/homo_sapiens/homo_sapiens.GRCh38.Regulatory_Build.regulatory_features.20221007.gff.gz

[tedil]

Do you happen to know of a programmatic way to get that date correct? Everything else can easily be derived/specified, but the date cannot be known in advance, unless you either check the ftp directory / there is an overview file with release <-> freeze date information.

[julia-luz]

I'm pretty new to programming in general, which is why my first instinct is probably to regex hammer it - wget does have an option for it, I could play around with that.