chore(deployment,website,tests): remove dummy pipeline and organisms#5942
chore(deployment,website,tests): remove dummy pipeline and organisms#5942theosanderson wants to merge 8 commits intomainfrom
Conversation
Added troubleshooting section for k3d setup and updated deployment instructions.
chaoran-chen
left a comment
There was a problem hiding this comment.
Very nice, thanks @theosanderson! Before reviewing the code in detail, I have a few general comments.
There was a problem hiding this comment.
There're (at least?) two mentions of the dummy pipeline in the docs that should be removed
There was a problem hiding this comment.
Thanks for the review! Will address these
There was a problem hiding this comment.
It would be great to have a short migration guide just as part of the PR description about how to migrate from the dummy pipeline to the Nextclade one. As the dummy pipeline was the only one that supported files, I've used it whenever showing someone how to set up an instance with files, so it would be good to have a link to send to people for migrating.
There was a problem hiding this comment.
The dummy pipeline is explicitly described as just for running tests and demonstrations, rather than for end-users, and I think that has been our consensus understanding of it as developers (hence tolerating various bugs), so I don't really consider this supported behaviour. At the time of implementation I did suggest implementing the functionality directly in the Nextclade pipeline. As below, I've no problem with us keeping the dummy pipeline around for weeks or a month, and of course I have no problem with anyone writing such a guide for when we remove it but atm I wouldn't personally plan to.
There was a problem hiding this comment.
Similar to the point above, as we've so far been using the dummy pipeline whenever we needed files, including in our live wastewater instance, it would be great if we can first test that the Nextclade pipeline is working well before the dummy pipeline gets dropped. Could we first merge the new Nextclade pipeline feature and changes to the Loculus previews but leave the old pipeline code and image for two more weeks after merging this PR before removing?
There was a problem hiding this comment.
Yes, there's no rush to remove the old version!
| dataIsLoaded={!!metadataFile && (!enableConsensusSequences || !!sequenceFile)} | ||
| /> | ||
| )} | ||
| {(organism.startsWith('not-aligned-organism') || organism === 'ebola-sudan') && action === 'submit' && ( |
There was a problem hiding this comment.
This is hardcoded into Loculus so would appear on PPX for ebola sudan! Must fix :)
Remove Dummy Organisms and Dummy Preprocessing Pipeline
Summary
This PR removes the
dummy-organism,dummy-organism-with-files, and the entirepreprocessing/dummydirectory while retaining all test coverage by migrating to use real organisms with nextclade preprocessing.Changes
Kubernetes/Helm Configuration
dummy-organismfromvalues.yaml- tests now useebola-sudaninsteaddummy-organism-with-filestotest-organism-files- this organism is used for file upload testing and now uses nextclade preprocessing instead of the dummy preprocessing pipelinetest-organism-filesorganism:consensusSequences: false(metadata-only, no sequences required)raw_readscategory)preprocessing-nextcladewith no genes configured (for metadata-only processing)Preprocessing Changes
Added file preservation to the nextclade preprocessing pipeline so that files attached to submissions are carried through to the processed output:
datatypes.py: Addedfilesfield toUnprocessedDataclassbackend.py: Extract files from backend JSON response when parsing unprocessed dataprepro.py: Pass files through toProcessedDatain the no-alignment code pathIntegration Tests
Updated tests to use real organisms instead of dummy organisms:
file-sharing.spec.ts: ChangedORGANISM_URL_NAMEfromdummy-organism-with-filestotest-organism-filescolumn-visibility.dependent.spec.ts: Changed from "Test Dummy Organism" to "Ebola Sudan"landing-page.spec.ts: Changed featured organism to "Ebola Sudan"navigation.spec.ts: Changed organism to "Ebola Sudan"authentication.spec.ts: Changed URL from/dummy-organism/...to/ebola-sudan/...submission.page.ts: RenamedfillSubmissionFormDummyOrganismtofillSubmissionFormFilesOrganismWebsite
Updated the dev example data generator in
SequenceEntryUploadComponent.tsx:ebola-sudanpages instead ofdummy-organismpagessubmissionId,sampleCollectionDate,geoLocCountry,authorAffiliations)Deleted Files
preprocessing/dummy/(entire directory includingmain.py,Dockerfile,mock-sequences.json, etc.).github/workflows/preprocessing-dummy-image.ymlCI/Workflow Updates
trigger-dummy-preprocessingjob from.github/workflows/build-arm-images.yaml🚀 Preview: https://dummy2.loculus.org