Stop workflow (without error) if an output file is empty

[davidecrs]

Hi,

is it possible to stop the snakemake workflow if an output file is empty? I've tried with the checkpoints but couldn't properly create the rules and make it work.

For example, in the following rule, I'm computing the assembly with metaspades and filtering (with seqkit seq)the produced contigs.fasta to retrieve only the contigs with 1000 min-len (output: "filtered_contigs.fasta"). However, After this last step, the filtering removes all the contigs producing an empty file.

Since many other rules depend on the file "filtered_contigs.fasta", is it possible to stop the execution of snakemake (without errors) ?

run_metaspades:
    input:
        r1=expand("{wdir}/{sample}/{pre_processing}/R1.fastq.gz", pre_processing=config["MODULES"]["pre_processing"], allow_missing=True),
        r2=expand("{wdir}/{sample}/{pre_processing}/R2.fastq.gz", pre_processing=config["MODULES"]["pre_processing"], allow_missing=True),
    output:
        folder = directory("{wdir}/{sample}/metaspades"),
        contigs_fasta = "{wdir}/{sample}/metaspades/contigs.fasta",
        filtered_fasta = "{wdir}/{sample}/metaspades/filtered_contigs.fasta"
    threads: config["threads"]
    conda: config["ENVS"]["metaspades"] 
    shell:
        """
        mkdir -p {output.folder}

        spades.py --meta --only-assembler -t {threads} {params.user_params} -1 {input.r1} -2 {input.r2} -o {output.folder}
        seqkit seq --min-len 1000 {output.contigs_fasta} -o {output.filtered_fasta}
        """

Best regards

[nate-d-olson]

I believe you want to use filtered_fasta = ensure("{wdir}/{sample}/metaspades/filtered_contigs.fasta", non_empty = True) in your output directive.

https://snakemake.readthedocs.io/en/stable/snakefiles/rules.html#ensuring-output-file-properties-like-non-emptyness-or-checksum-compliance