improvement: restart from specific step in case the pipeline stopped prematurely

[lassancejm]

Hi, I have been running orthomcl-pipeline on a cluster, and it has happened that the job would timeout or run out of memory.

At present, correct me if I am wrong, any restart of the pipeline causes all results from the previous run (including blast, which takes the longest) to be lost, as one would typically use the --yes option to answer yes to all questions.

It would be nice to have the possibility to restart at the step that failed.

Thanks

JML

[apetkau]

Hello,

Thank you for the suggesting. I can see how it would be very useful if a job times out or fails (so you don't have to restart from scratch).

I do not have time for any additions to this project right now (beyond bugfixes) but since all necessary files are stored in the output directory, it would probably be possible to get something working by modifying the pipeline script to simply ignore all steps up to a certain point and just assume the necessary files exist (e.g., ignore all steps up to blast and just start up after blast).

Aaron