source code study

[mewu3]

Hello,

I am studying the source code.

rule run_checkm_lineage_wf:
    input:
        touched_output="logs/checkm_init.txt",
        bins="{sample}/binning/{binner}/bins",  # actualy path to fastas
    output:
        "{sample}/binning/{binner}/checkm/completeness.tsv",
        "{sample}/binning/{binner}/checkm/storage/tree/concatenated.fasta",
    params:
        output_dir=lambda wc, output: os.path.dirname(output[0]),
    conda:
        "%s/checkm.yaml" % CONDAENV
    threads: config.get("threads", 1)
    log:
        "{sample}/logs/binning/{binner}/checkm.log",
    resources:
        time=config["runtime"]["long"],
        mem=config["large_mem"],
    benchmark:
        "logs/benchmarks/checkm_lineage_wf/{sample}_{binner}.tsv"
    shell:
        """
        rm -r {params.output_dir}
        checkm lineage_wf \
            --file {params.output_dir}/completeness.tsv \
            --tab_table \
            --quiet \
            --extension fasta \
            --threads {threads} \
            {input.bins} \
            {params.output_dir} &> {log}
        """

why output_dir=lambda wc, output: os.path.dirname(output[0]) has to be in a lambda function, and what wc means ? cause as I understand here the only thing matters would be os.path.dirname(output[0]) which give the path.

mewu3

[SilasK]

First I made a lambda function to be more flexible. If I change the output it automatically adapts. The wc stands for wildcards.

The function follows the way snakemake requires it. (at least at the time I wrote the code). Snakemake allows for function that have wildcards, output or input as argument. But the wildcard are always required

[mewu3]

thank you for the answer !