How to reference input variable in output section of rule of Snakemake rule

Question

[Dhatri Badri]

I have a function def get_ref_name and a rule called bioawk on my snakemake script.

def get_ref_name(ref_genome_path):
       file_name = os.path.basename(ref_genome_path)
       ref_name = file_name.split('.')[0]
       return ref_name

rule all:
    input: 
        bioawk_ref_size_file = expand("results/{prefix}/bioawk/{ref_name}.size", prefix=PREFIX, ref_name=get_ref_name(config["reference_genome"]))            

rule bioawk:
    input:
        ref_name=get_ref_name(config["reference_genome"]),
        ref_genome=config["reference_genome"]
    output:
        reference_size_file=f"results/{{prefix}}/bioawk/{{ref_name}}.size"
    conda:
        "envs/bioawk.yaml"
    shell:
        "bioawk -c fastx '{{ print $name, length($seq) }}' < {input.ref_genome} > {output.reference_size_file}"

I am trying to get the name of the reference genome from a path specified on my config file. So for example, if the path leads to a KPNIH1.fasta file the output of the function is supposed to be KPNIH1. I am trying to call the output from the get_ref_name function in the input section as ref_name and calling it again in the output section {{ref_name}}. However, I am getting this error:

MissingInputException in line 320 of /.../snpkit-snakemake-test/snpkit_v3.smk:
Missing input files for rule bioawk:
KPNIH1

I assume there may be an issue with the way i am calling the function in the input section but not sure how to fix it. Thanks in advance for your help!

[Ram]

If both your "inputs" are from a config file, are they really inputs? You could place them in params, because inputs are expected to be files.

Answer 1

[Jeremy Leipzig]

a lookup function helps Snakemake identify the input it needs to find to generate an output using a wildcard term, when the lookup is complex. In your case it is not complex, it's just the output has a .fasta and the input doesn't.

also you are confusing python f-strings with snakemake wildcards

input:
   os.path.join(config['reference_genome'],"{ref_name_stem}")
output:
   "results/{prefix}/bioawk/{ref_name_stem}.fasta"

[Dhatri Badri]

The output has a .size file and I would like to have the reference genome name (ref_name) be part of the file name in the output (ref_name.size). The config['reference_genome'] points to a filepath as a string (/.../variant_calling_bin/reference/KPNIH1/KPNIH1.fasta) so I would like to extract just the KPNIH1 from KPNIH1.fasta and have it as

output:
   "results/{prefix}/bioawk/{ref_name}.size" # KPNIH1.size

[Jeremy Leipzig]

you don't need a lookup function if there is a simple relationship between your input and output filenames