Louis Le Nézet
07/02/2025, 9:58 AMopt["dataset"] = "${task.ext.dataset ?: meta.dataset}"
The issue I see with such pattern is that the user needs to pass the value either by the ext arguments or by the meta map.
I have recommended to use a value channel instead to allow a description of the input in the yml file.
This is similar to @Mahesh Binzer-Panchal issue with ext arguments, what are the different key of the meta map that should be allowed ?
What should I answer to the contributor ?
The dataset is always expected to be provided, and it has already been included in the meta.yml via thehttps://github.com/nf-core/modules/pull/8698map.meta
Simon Pearce
07/02/2025, 10:02 AMLouis Le Nézet
07/02/2025, 10:04 AMdataset_id = opt["dataset"]
processed = input_processing(data, dataset_id, genome)
## main.nf.test
input[0] = [
[ id:'test', dataset: 'HCC1395' ],
file(params.modules_testdata_base_path + 'delete_me/sigprofiler/test_GRCh38/HCC1395_WXS_HCC1395_joint_table.tsv', checkIfExists: true),
]Simon Pearce
07/02/2025, 10:11 AMSimon Pearce
07/02/2025, 10:11 AMSimon Pearce
07/02/2025, 10:13 AMprefixLouis Le Nézet
07/02/2025, 10:15 AMnew_columns = {'Project': dataset_id, 'Genome': genome, 'Type': "SOMATIC", 'mut_type': "SNP"}
df = data.assign(**new_columns)
It's also use as input to compute some kind of matrix
generate_count_matrix = (
f"SigProfilerMatrixGenerator matrix_generator "
f"{dataset_id} {genome} {input_path} --volume {opt["volume"]}"
)Louis Le Nézet
07/02/2025, 10:16 AMFelix Kummer
07/02/2025, 10:31 AMmeta.dataset it is not a standard field as far as I know.
Not sure about prefix . The parameter seems to be used for things other than naming output files.Igor Trujnara
07/02/2025, 10:33 AMtask.ext feels hacky. If it's a mandatory input, then I see no reason not to do it by the book and make it a channel.Louis Le Nézet
07/02/2025, 10:34 AMIgor Trujnara
07/02/2025, 10:52 AMMahesh Binzer-Panchal
07/02/2025, 10:55 AMtask.ext maps are intended for users to override pipeline developer settings.