A call for a bio-cwl-tools team

[pvanheus]

At the moment I have a number of PRs open against this repository, and I have opened a discussion on moving this repository forward. I would like to see a few things addressed:

1. Adding a PR template
2. Increasing the pool of people who are contributors to the repository
3. Developing the expectations for tool testing and testing infrastructure further
4. Developing best practice for using bio-cwl-tools in your own workflows (also see this) discussion

Please comment below if you think this is a reasonable programme of action and if you would like to contribute.

[michael-kotliar]

:thumbsup:

[hmenager]

:+1: one thing I'd like to include is cross-links to bio.tools entries, if possible. It would eventually make it possible for us at bio.tools to link from bio.tools to bio-cwl-tools.

[pvanheus]

@hmenager cool - how would we do this? can you provide some examples via PR and then we can PR this as best-practice into the CONTRIBUTING.md doc?

[hmenager]

Here is an existing example for bio.tools annotation I found in the repository: https://github.com/common-workflow-library/bio-cwl-tools/blob/release/qualimap/qualimap_rnaseq.cwl#L21

We could advise people to link this way, couldn't we?

[pvanheus]

@hmenager - yes, that looks like a useful thing to add to the CONTRIBUTING guide.

One more question on best practice: should tools be run through cwl-format before being added to this repository?

[nsheff]

I'm interested in helping with this. A few thoughts:

• is there an approval process required to have a tool appear in here?

• should individual tool files be version controlled? or, at least, should we version control the repository with a proper development/release cycle?

perhaps a more flexible approach would be to have a "bio-cwl-tools" organization, with a repository for each tool, and then there can be branches/tags for different versions of a tool. This way, you can check out the cwl tool file for the version of each tool that you need. The disadvantage of putting everything here into folders in one repository is that you may not be using exactly the set of versions for all the tools listed here. But maybe that's overkill? An alternative would be to follow the common approach used in environment modules systems, where you do something like:

refgenie/
  1.2.3.cwl
  1.2.4.cwl
  1.3.0.cwl

In this framework, different versions live as parallel files within a subfolder. In this case we would need to standardize the naming of the tool files and versions.

[swzCuroverse]

Hi there - I would like to contribute or figure out if we can donate some resources from Curii to help out. This is something we could fund via some of our CZI grant funds.

[inutano]

Hi @pvanheus I'd like to help! The CWL community JP has some tools which we would like to deposit to this repo. I'd also love to discuss how we test this! cc. @tom-tan @suecharo @manabuishii

[hmenager]

@hmenager - yes, that looks like a useful thing to add to the CONTRIBUTING guide.

One more question on best practice: should tools be run through cwl-format before being added to this repository?

@pvanheus I would be in favor of such a move, to promote some homogeneity in style for the contributed CWL files, but reading the formatting rules of cwl-format, these seem to be very opinionated, so it might require some formal agreement from most contributors/potential contributors.