phylo: Standard metadata id field#108
Conversation
Update to our new standard of defining inputs with `id_field` and expecting all downstream Augur commands to use `id` from <nextstrain/zika#110> Part of <nextstrain/public#44> Co-authored-by: Victor Lin <vlin@fredhutch.org>
There was a problem hiding this comment.
Thoughts on moving this code to a Snakemake module for reusability?
module merge_inputs:
snakefile: "../../shared/vendored/snakemake/merge_inputs.smk"
config: config
use rule merge_metadata from merge_inputs with:
output:
metadata = "results/metadata.tsv"
use rule merge_sequences from merge_inputs with:
output:
sequences = "results/sequences.fasta"There was a problem hiding this comment.
It's been a while since I've looked at modules. I think this could work now that we only need the merge rule without caring about the number of inputs...
My initial question is how to handle the _gather_inputs function since that may differ per pathogen (e.g. avian-flu handles segment sequences). Reading through the module docs again, I'm remembering that it's not possible to override a function within a module. I think we would need to pass in the input_sources variable via the config to the Snakemake module.
There was a problem hiding this comment.
Even if this gets weird for some repos like avian-flu, those repos can continue to keep their own version of merge_inputs.smk.
But the current code should work for most repos, and I wonder if the idea can be extended to other rules as well...
There was a problem hiding this comment.
Even if this gets weird for some repos like avian-flu, those repos can continue to keep their own version of merge_inputs.smk.
Ah, that's true! It's definitely an opt-in
But the current code should work for most repos, and I wonder if the idea can be extended to other rules as well...
I think it'd be nice to do for completely standardized steps like the initial merge of inputs. Other rules might be more up for debate (based on older discussions of modules)...I imagine pivoting to config based augur commands could also bring us to modules since the augur rules would essentially be identical across pathogens.
There was a problem hiding this comment.
Yeah. It can be considered separately from this PR though!
There was a problem hiding this comment.
2 participants
Description of proposed changes
Update to our new standard of defining inputs with
id_fieldand expecting all downstream Augur commands to useidfrom nextstrain/zika#110.Related issue(s)
Part of nextstrain/public#44
Checklist