Snakemake: Target rules may not contain wildcards - snakemake

Question

I am trying to supply a bunch of files as input to snakemake and wildcards are not working for some reason:
rule cluster:
input:
script = '/Users/nikitavlasenko/python_scripts/python/dbscan.py',
path = '/Users/nikitavlasenko/python_scripts/data_files/umap/{sample}.csv'
output:
path = '/Users/nikitavlasenko/python_scripts/output/{sample}'
shell:
"python {input.script} -data {input.path} -eps '0.3' -min_samples '10' -path {output.path}"
I want snakemake to read files in from the umap directory, get their names, and then use them to pass to the python script, so that each result would get a unique name. How this task can be achieved without such an error that I am getting right now:
Building DAG of jobs...
WorkflowError:
Target rules may not contain wildcards. Please specify concrete files or
a rule without wildcards.
Update
I found that most probably the rule all is required at the top:
https://bioinformatics.stackexchange.com/questions/2761/how-to-resolve-in-snakemake-error-target-rules-may-not-contain-wildcards
So I added it like that:
samples='SCID_WT_CCA'
rule all:
input:
expand('/Users/nikitavlasenko/python_scripts/data_files/umap/
{sample}_umap.csv', sample=samples.split(' '))
However, I am getting the following weird message:
Building DAG of jobs...
Nothing to be done.
So, it is not running.
Update
I thought that it could be related to the fact that I had just one sample name at the top, so I changed it to:
samples='SCID_WT_CCA WT SCID plus_1 minus_1'
And added the respective files, of course, but it did not fix this error.
Actually if I run snakemake cluster I get the same error as at the very top, but if I just run snakemake, then there is nothing to be done error. I tried to substitute absolute paths for the relative ones, but it did not help:
samples='SCID_WT_CCA WT SCID plus_1 minus_1'
rule all:
input:
expand('data_files/umap/{sample}_umap.csv', sample=samples.split(' '))
rule cluster:
input:
script = 'python/dbscan.py',
path = 'data_files/umap/{sample}_umap.csv'
output:
path = 'output/{sample}'
shell:
"python {input.script} -data {input.path} -eps '0.3' -min_samples '10' -path {output.path}"

Answer 1

The "all" rule should have as input the list of files you want the other rule(s) to generate as output. Here, you seem to be using the list of your starting files instead.
Try the following:
samples = 'SCID_WT_CCA WT SCID plus_1 minus_1'
rule all:
input:
expand('output/{sample}', sample=samples.split(' '))
rule cluster:
input:
script = 'python/dbscan.py',
path = 'data_files/umap/{sample}_umap.csv'
output:
path = 'output/{sample}'
shell:
"python {input.script} -data {input.path} -eps '0.3' -min_samples '10' -path {output.path}"

Answer 2

Following bli's answer, I was able to solve the issue. However, one additional modification was needed. I passed output/{sample} to the python script and it generated two files from this path. Seems like that should not be done because I got another error when snakemake wrote that it could not see output/file_name. Obviously it will be able to see them only if I set all the paths manually right away without python modifying it on the fly, so I did that and here is the final Snakefile that worked well:
samples='SCID_WT_CCA WT SCID plus_1 minus_1'
rule all:
input:
expand('output/{sample}_umap.png', sample=samples.split(' ')),
expand('output/{sample}_clusters.csv', sample=samples.split(' '))
rule cluster:
input:
script = 'python/dbscan.py',
path = 'data_files/umap/{sample}_umap.csv'
output:
path_to_umap = 'output/{sample}_umap.png',
path_to_clusters = 'output/{sample}_clusters.csv'
shell:
"python {input.script} -data {input.path} -eps '0.3' -min_samples '10' -path_to_umap {output.path_to_umap} -path_to_clusters {output.path_to_clusters}"