I don't think this is relevant to the current failing workflows, but we have a large number of pointless metadata-update callables. This process of updating an analysis dict with a called method in a later job is... probably not a good mechanic.
I would personally use a combination of analysis entry type and meta.stage to find entries, and AFAIK none of the fields removed here are used in scripts/other stages. The gvcf_qc meta update should have already been removed in https://github.com/populationgenomics/production-pipelines/pull/718
In code these callable methods are annotated with TODO: Replace this once dynamic analysis types land in metamist. - we have that now, so we should probably update the analysis types we generate
illusional
commented
1 month ago
I don't think this is relevant to the current failing workflows, but we have a large number of pointless metadata-update callables. This process of updating an analysis dict with a called method in a later job is... probably not a good mechanic.
I would personally use a combination of analysis entry type and meta.stage to find entries, and AFAIK none of the fields removed here are used in scripts/other stages. The gvcf_qc meta update should have already been removed in https://github.com/populationgenomics/production-pipelines/pull/718
In code these callable methods are annotated with
TODO: Replace this once dynamic analysis types land in metamist.- we have that now, so we should probably update the analysis types we generate