Skip to content

Add seven two-package combinations for the BRC pangenome tool set - #4131

Open
richard-burhans wants to merge 1 commit into
BioContainers:masterfrom
richard-burhans:add-brc-tools-combinations
Open

Add seven two-package combinations for the BRC pangenome tool set#4131
richard-burhans wants to merge 1 commit into
BioContainers:masterfrom
richard-burhans:add-brc-tools-combinations

Conversation

@richard-burhans

Copy link
Copy Markdown

Seven two-package combinations, all for steps in the BRC pangenome tool set.

Each wraps a command that needs binaries from two conda packages. Galaxy's User-Defined Tools give a job exactly one container, so with no published multi-package image each of these steps has to be split into two tools that pass files between them — an extra workflow node, an extra staging round trip, and a second place for the logic to drift.

combination why
gffread=0.12.7,python=3.12 gffread parses the GFF3, a Python helper reshapes it
blast=2.17.0,python=3.12 dustmasker / windowmasker pipe into a Python interval classifier
tantan=51,python=3.12 same shape, tantan instead of BLAST
fastan=0.8,fastga=1.5.20260729 FasTAN scans, FastGA builds the GDB and reads the .1ano
python=3.12,pyfaidx=0.9.0.4 random access to assemblies from a Python triage step
multiz=11.2,python=3.12 multiz folds the MAFs, Python drives the fold order
python=3.12,ucsc-bedtobigbed=469 a Python hub builder that shells out to bedToBigBed

Checked before opening:

  • every version pinned here already has a published single-package image in the Galaxy depot, so each pair should resolve;
  • none of the seven mulled images exists yet — each name was derived with mulled-hash and looked up against the depot's 3,472 published mulled-v2 images;
  • lastz=1.0.4,python=3.9 is already in hash.tsv, so this is the established shape rather than a new one.

Happy to split this into separate PRs if one combination is contentious.

Each of these wraps a pipeline step whose command needs binaries from two conda packages. Galaxy's
User-Defined Tools give a job exactly ONE container, so without a published multi-package image each
step has to be split into two tools that pass files between them -- an extra workflow node, an extra
staging round trip, and a second place for the logic to drift.

The names are derivable (`mulled-hash`), and every one of these was checked against the Galaxy
depot's 3,472 published `mulled-v2` images before being added here; none exists yet.

  gffread=0.12.7,python=3.12          gffread parses the GFF3, a Python helper reshapes it
  blast=2.17.0,python=3.12            dustmasker/windowmasker pipe into a Python interval classifier
  tantan=51,python=3.12               same shape, tantan instead of BLAST
  fastan=0.8,fastga=1.5.20260729      FasTAN scans, FastGA builds the GDB and reads the .1ano
  python=3.12,pyfaidx=0.9.0.4         random access to assemblies from a Python triage step
  multiz=11.2,python=3.12             multiz folds the MAFs, Python drives the fold order
  python=3.12,ucsc-bedtobigbed=469    a Python hub builder that shells out to bedToBigBed

Every version pinned above already has a published single-package image, so each pair should
resolve. `lastz=1.0.4,python=3.9` is the existing precedent for this shape.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant