seq2science
seq2science copied to clipboard
BUG: inconsequential CLI changes prompt reruns
Describe the bug Changing the number of cores should not prompt reruns.
I'll add other options to the list if I find them.
-r -k --snakemakeOptions untill=[] are fine.
--snakemakeOptions omit_from=[] prompts a rerun.
Ah good catch! Does it re-run everything or a small-ish selection/branch only (e.g. qc report)? I can imagine params like this causes the qc stuff to be re-run. Seems like we need a config parsing function for this :see_no_evil:
https://github.com/vanheeringen-lab/seq2science/blob/master/seq2science/rules/qc.smk#L502
I would expect both snakemakeoptions to cause the same behavior
these rules use threads in their params:
- bam2cram
- fastp_SE
- fastp_PE
- samtools_presort
- samtools_sort
these rules seem to get params changes listed:
- multiqc_explain
- multiqc_header_info
- multiqc_samplesconfig
Finally, the fastp and samtools_presort rules seem to get marked all the time, even when threads are removed from the params. Could this be related to the temp()?
I think you've solved this right?