search

chrisquince / STRONG

Strain Resolution ON Graphs
MIT License
44 stars 9 forks source link

Refactorising and restructuring workflow #48

Closed chrisquince closed 5 years ago

chrisquince commented 5 years ago

I am going to try and implement some of the changes we discussed but never made progress on, refactoring scripts, and reorganising output to make things cleaner. Therefore don't make any updates over the next few days whilst I do this.

chrisquince commented 5 years ago

Seb, regarding the workflow structure, why for the tree construction did you decide to control execution by creating a fake file as a flag rather than having a separate MAGAnalysis.snake that is called only if argument is set as you did for Desman? The latter seems cleaner and easier? In general, the use of fake files as workflow flags seems suboptimal but maybe I am wrong about that?

chrisquince commented 5 years ago

OK well I changed to a flag in start.py and am testing that now

Sebastien-Raguideau commented 5 years ago

Hi Chris, late answer, there was no particular rational behind that. After thinking about it, I agree with you : if using a fake file, the user can't, at a later date, run the maganalysis if he choose not to run it at first.

chrisquince commented 5 years ago

Hi Seb, yes and I have changed bayespaths and desman outputs to be generated in their own dirs too that facilitates rerunning. I am happier with the output structure now but some of the snake structure seems a bit over complex though. Sergey may look at that next week.

snurk commented 5 years ago

Me and @Sebastien-Raguideau discussed that it makes sense to extract everything related to bayespaths into its own snakefile.