How to activate conda env in a docker images?

[liubingdong]

Many thanks for your excellent job on the nextflow.
I plan to build a workflow and create a docker image including qiime2, fastqc, and picrust2, like this:

I have checked these software in the conda env works well.
However, when I run the workflow, the pipeline cant finds the software like this:

Here is my question:
Could I activate the conda env in the nextflow based on a docker image?
This is my code:

nextflow.enable.dsl=2

/*
========================================================================================
    IMPORT LOCAL MODULES/SUBWORKFLOWS
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*/
params.read = 'data/real/*.fastq'
params.outdir = "test_results"

read_ch = Channel.fromFilePairs(params.read,checkIfExists: true)


include {FASTQC;MULTIQC} from './mouldes/qc.nf'



workflow qc_workflow {
    take:
      read_ch
    main:
      fastqc_ch=FASTQC(read_ch)
      MULTIQC(fastqc_ch.collect())
}

workflow {
    qc_workflow(read_ch)
}

// configuration_process-labels.config
container = 'bingdong_qiime2_picrust_qc:1.0'
docker.enabled = true


params {
    reads = 'data/real/*_{R1,R2}.fastq'
    outdir = 'test_results'
}

process {
    withName: FASTQC {
        cpus = 10
        memory = {2.GB * task.cpus }
        publishDir = [ path: "${params.outdir}/each", mode: 'copy' ]
    }
    withName: MULTIQC {
        publishDir = [ path: params.outdir, mode: 'copy' ]
    }
    withLabel: quality_control {
        conda =  "/opt/conda/envs/qc_moudle"
    }
}

process FASTQC {

    label "quality_control"  
    tag "fastqc on $sample_id"
    
    input:
    tuple val(sample_id), path(reads)

    output:
    path "fastqc_${sample_id}_logs/*"

    script:
    //flagstat simple stats on bam file
    """
    mkdir fastqc_${sample_id}_logs
    fastqc -o fastqc_${sample_id}_logs -f fastq -q ${reads} -t ${task.cpus}
    """
}

process MULTIQC {

    label "quality_control"  

    input:
    path '*'

    output:
    path 'multiqc_report.html'

    script:
    """
    multiqc .
    """
}

[pditommaso]

You can use Wave to build containers on demand starting from the conda declaration. See here https://www.nextflow.io/docs/latest/wave.html#build-conda-based-containers

[liubingdong]

Thanks for your quick response.
I found a solution to address the problem. and it worked. Just add "conda run " in the script, like this

process FASTQC {

    label "quality_control"  
    tag "fastqc on $sample_id"
    
    input:
    tuple val(sample_id), path(reads)

    output:
    path "fastqc_${sample_id}_logs/*"

    script:
    //flagstat simple stats on bam file
    """
    mkdir fastqc_${sample_id}_logs
    conda run -n qc_moudle fastqc -o fastqc_${sample_id}_logs -f fastq -q ${reads} -t ${task.cpus}
    """
}

I still wonder if we could use the withName function in the nextflow.config to activate the conda env like above
Any advice is welcome.