apply a snakemake rule on a list of values in parallel

Question

I'm currently exploring using snakemake as a workflow tool.

In my specific use case I don't start from a list of files but rather from a list of values that should result in the creation of a list of files.

In my example, I create the files with a small python snippet which works fine but when I want to use those files in parallel in a second rule, they are concatenated into one parameter:

rule all:
    input:
        expand('{file}.bar', file=data)

rule foo:
    output:
        expand('{file}.foo', file=data)
    run:
        for item in data:
            with open(f'{item}.foo', 'w') as fout:
                fout.write('foo')

rule bar:
    input:
        file=expand('{file}.foo', file=data)
    output:
        outfile=expand('{file}.bar', file=data)
    shell:
        """echo {output.outfile};echo bar > {output.outfile} """

the example prints

"one.bar two.bar three.bar"

at once, so the rule is applied only once,

and then raises an error because the expected output files are not created.

Answer 1

Probably you want rule bar without expand's:

rule bar:
    input:
        file='{file}.foo',
    output:
        outfile='{file}.bar',
    shell:
        """echo {output.outfile};echo bar > {output.outfile} """

this is because you apply this rule to each {file} individually not to all of them at once. You would use expand if, for example, you wanted to concatenate all {file}'s in a single one.

By the same reasoning, you could also change rule foo to:

rule foo:
    output:
        '{file}.foo',
    run:
        item=wildcards.file
        with open(f'{item}.foo', 'w') as fout:
            fout.write('foo')

(not tested)