This is awesome! How can I help?

[rpetit3]

Hi @Mxrcon and @abhi18av,

This is really awesome and really want to commend your progress over the last few months! I think Bactopia is to the point that DSL2 is necessary to make things easier going forward (e.g. assembly only workflow, workflow with PIRATE, etc..).

Given so much effort has already put in on your end, I'd like help here where I can. So! Let me know where you think I might be useful!

Again, super awesome stuff and greatly appreciated!
Robert

[abhi18av]

Hi @rpetit3 ,

Thanks for noticing this!

We started working on bactopia as a good foundation for helping @Mxrcon get started with NF and workflow design choices and he's been very dedicated in this goal :)

The result is this
https://github.com/bioinformatics-lab/bactopia/tree/dsl2

So, moving forward as you might have noticed, we have already isolated vairous processes into modules and with some more elbow grease made them individually testable. The new stub feature has also been implemented in the same vein.

We were planning to start a discussion within the upstream bactopia repo to ask you how would you like to move forward with this.

One of the salient features of bactopia is the use of templates, however with DSL2 till the point when nextflow-io/nextflow#1798 is resolved, for testing thing individually we had to copy/paste the same template within each of the modules, so that you can pick and choose the modules which you like.

[abhi18av]

The main design choices for this transition are influenced by what I've used extensively in other workflows and hopefully it leads to an easy to read code.

[Mxrcon]

@rpetit3 We're very happy on hear that!
My main focus was on the modules, if you have time, you can try to test them! I've just finished the process to make all of them test-able, and Abhinav will test them again later on!

I'm very excited to contribute with bactopia, as soon as we open the discussion on the upstream branch lets discuss our changes there! I'll be very happy on hear your thoughts about our design choices for dsl2.

Regards,
Davi

[rpetit3]

We started working on bactopia as a good foundation for helping @Mxrcon get started with NF and workflow design choices and he's been very dedicated in this goal :)

haha I've been periodically checking on it since the accidental PR, Davi you're doing awesome! It's been fun watching the progress.

One of the salient features of bactopia is the use of templates, however with DSL2 till the point when nextflow-io/nextflow#1798 is resolved

My initial use of templates was to avoid a super long main.nf, but with DSL2 and everything being a module, please feel free to drop the templates and move them into the module workflows. Especially if it makes things easier, unless of course you see a need keep them.

I'll be very happy on hear your thoughts about our design choices for dsl2.
Overall I think it works great! It definitely looks like something I can run with, especially when the DSL1 folders/files are removed.

I would probably change some of the folder names in in the modules folder. For example, shovill\assemble_genome, I would probably just change it to assemble_genome since it also include Unicylcer. Although, with DSL2 it could just split into a shovill module and a unicycler module (but still include the post processing e.g. assembly stats).

But at the end of the day, any changes I'd make would be very minor and nothing to the actual structure (because I like it!).

Again, great work! I'll start testing on my end

[abhi18av]

That's really heart warming to hear Robert!

We'd be opening up a PR very soon against the dsl2 branch on the main repo and then we can proceed with module/workflow level.

We've only focused on the module level work, however there are still parts like downloading genomes and channel processing which needs discussion and finishing touches. We didn't push for this since we'd wanted to hear your thoughts once we initiate the PR.

Anyways, let's get bactopia to DSL2 :)