Repository files navigation

Reusable and maintained Luigi tasks to incorporate in bioinformatics pipelines

Provides Luigi tasks for tools fromsamtools,bcftools,STAR,RSEM,vcfanno,GATK,Ensembl VEP and much more!

Reuses as much as possible theExternalProgramTask interface from theexternal_program contrib moduleand extends its feature to make it work on modern scheduler such asSlurm.

Provides basic resource management for a local scheduler: all tasks are annotated with reasonable defaultcpus andmemory parameters that can be tuned and constrained via the[resources] configuration. In the case of externally scheduled tasks, the resource management is deferred.

Provides a command-line interface for interacting more conveniently with Luigi scheduler.

bioluigi list [--status STATUS] [--user USER] [--detailed] TASK_GLOBbioluigi show TASK_ID

Here's a list of supported tools:

• sratoolkit withprefetch andfastq-dump
• bcftools
• FastQC
• MultiQC

• local
• Slurm

The most convenient way of using the pre-defined tasks is to yield them dynamically in the body of therun function. It's also possible to require them since they inherit fromluigi.Task.

importluigifrombioluigi.tasksimportbcftoolsdefMyTask(luigi.Task)definput(self):returnluigi.LocalTarget('source.vcf.gz')defrun(self):yieldbcftools.Annotate(self.input().path,annotations_file,self.output().path,                                ...,scheduler='slurm',cpus=8)defoutput(self):returnluigi.LocalTarget('annotated.vcf.gz')

You can define your own scheduled task by implementing theScheduledExternalProgramTask class. Note that the default scheduler islocal and will use Luigi's[resources] allocation mechanism.

importdatetimefrombioluigi.scheduled_external_programimportScheduledExternalProgramTaskclassMyScheduledTask(ScheduledExternalProgramTask):scheduler='slurm'walltime=datetime.timedelta(seconds=10)cpus=1memory=1defprogram_args(self):return ['sleep','10']