Fix phylo input - #40
Conversation
I missed this in review of #38. The phylogenetic workflow was still pulling from old S3 URLs and not the ingest workflow output data. This commit corrects the S3 URL to the ingest output files and updates the `strain_id_field` config param to use the appropriate ID column from the ingest output.
c73dde5 to
a42192d
Compare
|
Triggered trial phylogenetic workflow run with |
|
Thank you for working on this! I'll be curious how you solve the problem of adding an "Accession URL". The accession is important for investigating outliers in the tree or getting to the full GenBank entry. 
I'd been using a post-processing script in zika: https://github.com/nextstrain/zika/blob/9892009b95ea802f288142734614e29f00212922/phylogenetic/scripts/set_final_strain_name.py#L6-L15 But I think you mentioned there may be a more straightforward method: nextstrain/zika#32 |
a42192d to
8acbbca
Compare
Thanks for bringing this up @j23414, I totally forgot about this issue. |
8acbbca to
71f52b9
Compare
|
Hmm, CI failing is a separate issue due to installation of the Conda runtime failing. |
|
I'll wait to merge this on Monday so I can monitor for anything unexpected. I'll probably drop 5c85067 before merging since I'm working on a better way to split out ingest/phylogenetic GH Action workflows in https://github.com/nextstrain/zika/tree/reusable-workflows |
Match the ingest outputs to use `genbnak_accession` column.
Include the `genbank_accession` as a separate field in the exported metadata so that it can be searched in the Auspice filter input and so Auspice automatically generates a URL for the GenBank accession.
71f52b9 to
9b9d1b2
Compare
|
Manually deleted cache and triggered a new run of ingest-to-phylogenetic so that the live builds are actually using the ingested data. |
Updates the links of the FASTA and metadata in the phylogenetic README. This probably should have been part of the previous PR: #40
2 participants
Description of proposed changes
I missed this in review of #38.
The phylogenetic workflow was still pulling from old S3 URLs
and not the ingest workflow output data.
Checklist