Improving the Reliability and Quality of Nextflow Pipelines with nf-test
Forer, L.; Schoenherr, S.
Show abstract
The workflow management system Nextflow builds together with the nf-core community an essential ecosystem in Bioinformatics. However, ensuring the correctness and reliability of large and complex pipelines is challenging, since a unified and automated unit-style testing framework specific to Nextflow is still missing. To provide this crucial component to the community, we developed the testing framework nf-test. It introduces a modular approach that enables pipeline developers to test individual process blocks, workflow patterns and entire pipelines in insolation. nf-test is based on a similar syntax as Nextflow DSL 2 and provides unique features such as snapshot testing and smart testing to save resources by testing only changed modules. We show on different pipelines that these improvements minimize development time, reduce test execution time by up to 80% and enhance software quality by identifying bugs and issues early. Already adopted by dozens of pipelines, nf-test improves the robustness and reliability in pipeline development.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.