Major conda updates. Non-trivial complications with the installation of updated cyvcf2/vep/r-base/vcf2maf conda packages in a single environment meant that one of those would get downgraded substantially to a 2-4 year version, which isn't desirable (e.g. cyvcf2 would go from 0.30 to 0.10, vep would go from 104 to 89, r-base would go back to 3.* etc.).
Instead, I propose splitting the R package into its own conda environment, and calling the R command with a 'hack' that basically exports the R env in the Python call and uses that to generate the HTML report (we use this 'hack' substantially in our production workflows, so it's not as scary as it sounds).
Still testing so will leave this PR as a draft for now.
Major conda updates. Non-trivial complications with the installation of updated cyvcf2/vep/r-base/vcf2maf conda packages in a single environment meant that one of those would get downgraded substantially to a 2-4 year version, which isn't desirable (e.g. cyvcf2 would go from 0.30 to 0.10, vep would go from 104 to 89, r-base would go back to 3.* etc.).
Instead, I propose splitting the R package into its own conda environment, and calling the R command with a 'hack' that basically exports the R env in the Python call and uses that to generate the HTML report (we use this 'hack' substantially in our production workflows, so it's not as scary as it sounds).
Still testing so will leave this PR as a draft for now.