Adding apple-container as a new executor

[Joon-Klaps]

(Attemps to) Adress #7064

Claude* and I have tried to implement the container as a new executor and as a mini project to learn more about apple containers and nextflow.

Changes

• New apple container engine (apple { enabled = true }) and apple_container executor
• CLI flag -with-apple-container (analogous to -with-docker)
• Cross-architecture support: when process.arch = 'x86_64', emits --platform linux/amd64 --rosetta so x86_64 images run via Rosetta
• Wave compatibility: apple added to the DOCKER_LIKE engines list in WaveContainerResolver, since Apple uses the same OCI format as Docker

Testing

Tested on Apple Silicon (macOS) with the Apple container CLI installed.

nf-core/rnaseq with Wave (recommended):

NXF_PLUGINS_MODE=prod NXF_PLUGINS_DIR=$HOME/.nextflow/plugins \
  ./launch.sh run nf-core/rnaseq \
    -profile 'test',wave,arm64 \
    -with-apple-container \
    --outdir ./test-output

Pipeline ran end-to-end successfully (51+ processes, containerEngine: apple shown in run summary).

Simple pipeline without Wave:

./launch.sh run nextflow-io/rnaseq-nf -with-apple-container -c custom.config

Where custom.config contained process.arch = 'x86_64' as container assumes the container is linux/arm64.

*I mention claude as a declaration of the use of AI and that it has heavily assisted me in the development of the code & docs of this PR

[netlify]

✅ Deploy Preview for nextflow-docs-staging ready!

Name	Link
🔨 Latest commit	dcb34a0
🔍 Latest deploy log	https://app.netlify.com/projects/nextflow-docs-staging/deploys/69ea2a09e278ca0008a672d4
😎 Deploy Preview	https://deploy-preview-7066--nextflow-docs-staging.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.

[pditommaso]

Thanks for the contribution but nor sure we need this

[Joon-Klaps]

Hi @pditommaso, could you elaborate a bit more on this?

I would assume that reducing overhead would still be beneficial, even if small relative to task execution time. Is it because it doesn't translate into meaningful gains for large analyses (which would get run on HPC/cloud), or that macOS is mostly used for development, where such optimisations don’t matter as much?

[bentsherman]

I would try to implement this in a plugin instead of adding it to core. Since you just need a local executor that uses the apple container runtime, it should be possible.

You already have the executor, you just need to use the apple container runtime "under the hood" instead of exposing it is a user-facing container runtime.

https://docs.seqera.io/nextflow/plugins/developing-plugins

[pditommaso]

I mean, as developer, what's benefits of this approach considering docker is mainstream ?

[pditommaso]

I've thought a bit more about this, and I agree that could make sense supporting apple container, however as container engine not as an executor. Made an implementation here #7073. Thanks for opening this discussion.

Closing in favour of #7073