nextflow inspect v25.04 still requiring input parameters

[kchaung-lilly]

Bug report

Expected behavior and actual behavior

I wanted to use the updated nextflow inspect CLI tool to grab all containers from a pipeline. According to the migration doc, the new command should "includes all processes that are included by the entry workflow (directly or indirectly)."

The expected behavior was observed with nf-core pipelines, but I was unable to re-create with our internal pipelines.

Steps to reproduce the problem

I tested the command on a minimal pipeline, where each process container is defined within the module file.

Example module:

process BED_INTERSECT {

    label 'process_small'

    container "${ workflow.containerEngine == 'singularity' ?
        "wave_containers-bedtools--2.31.1--7c4ce4cb07c09ee4.sif" :
        'wave/containers:bedtools--2.31.1--7c4ce4cb07c09ee4' }"

    conda "bioconda::bedtools"

    input:
    path(ref_bed)
    path(sample_bed)

    output:
    path(out_bed)

    script:
    out_bed = "intersected.${sample_bed}"
    """
    bedtools intersect -a ${ref_bed} -b ${sample_bed} > ${out_bed}
    """
}

Program output

Nextflow version:

$ nextflow -version

      N E X T F L O W
      version 25.04.2 build 5947
      created 13-05-2025 21:58 UTC (14:58 PDT)
      cite doi:10.1038/nbt.3820
      http://nextflow.io

Testing the command on a minimal pipeline, with no process inclusions

$ nextflow inspect main.nf 
ERROR ~ Validation of pipeline parameters failed!


The following invalid input values have been detected:

* Missing required parameter(s): samplesheet, input_dir, reference_bed, outdir

Testing the command on a minimal piipeline, with process inclusions

$ nextflow inspect main.nf -params-file examples/params.local.json 
{
    "processes": [
        {
            "name": "ANNOTATE",
            "container": "pandas-extras:0.1"
        },
        {
            "name": "FASTQC",
            "container": "biocontainers/fastqc:0.11.9--0"
        },
        {
            "name": "MULTIQC",
            "container": "biocontainers/multiqc:1.21--pyhdfd78af_0"
        },
        {
            "name": "AWS_LOOKUP",
            "container": "aws-cli:2.27.6"
        },
        {
            "name": "BED_INTERSECT",
            "container": "wave/containers:bedtools--2.31.1--7c4ce4cb07c09ee4"
        }
    ]
}

Testing with docker enabled:

nextflow inspect main.nf --docker.enabled true
ERROR ~ Validation of pipeline parameters failed!


The following invalid input values have been detected:

* Missing required parameter(s): samplesheet, input_dir, reference_bed, outdir

 -- Check script 'main.nf' at line: 25 or see 'null' file for more details

Environment

• Nextflow version: 25.04.2 build 5947
• Java version: openjdk version "17.0.10" 2024-01-16
• Operating system: macOS
• Bash version: GNU bash, version 3.2.57(1)-release (arm64-apple-darwin24)

Additional context

(Add any other context about the problem here)

[kchaung-lilly]

Update: Providing dummy values for the params worked, I guess that's sufficient to bypass the nf-schema plugin.
Is this expected behavior? Is there anyway to bypass the plugin requirements from the CLI call?

$ nextflow inspect main.nf --samplesheet tmp --input_dir tmp --reference_bed tmp --outdir tmp
{
    "processes": [
        {
            "name": "FASTQC",
            "container": "biocontainers/fastqc:0.11.9--0"
        },
        {
            "name": "ANNOTATE",
            "container": "pandas-extras:0.1"
        },
        {
            "name": "MULTIQC",
            "container": "biocontainers/multiqc:1.21--pyhdfd78af_0"
        },
        {
            "name": "AWS_LOOKUP",
            "container": "aws-cli:2.27.6"
        },
        {
            "name": "BED_INTERSECT",
            "container": "wave/containers:bedtools--2.31.1--7c4ce4cb07c09ee4"
        }
    ]
}

[bentsherman]

Make sure all of your validation logic is contained within a workflow definition, such as the entry workflow. This is why the nf-core pipelines work without params:

https://github.com/nf-core/rnaseq/blob/b96a75361a4f1d49aa969a2b1c68e3e607de06e8/main.nf#L139-L173

nextflow lint will report such problematic code for you, since statements are not allowed to be mixed with script declarations in the strict syntax.