CRISPResso2 as a python package

[cdustinr]

Hi,

Thank you so much for your continued work on this fantastic resource! We use this on a nearly weekly basis upon hundreds of samples across many projects. CRISPResso is an (invaluable) component of our analysis/automation pipeline.

As far as I can tell, the CRISPResso resource isn't configured as a proper python package that can be imported and called from another python script. Pardon if I'm wrong on this, but I think arguments can only be taken from the command line (argparse). Is there any interest in a bit of restructuring so that it serves as a package? Given that our overarching pipeline is python-driven, it would really help if CRISPResso was an importable package rather than running as a command line from within python.

Thanks for any thoughts!
Dustin

[kclem]

Thanks for using CRISPResso2! I'm glad you've found it useful.

Once installed, you can import CRISPResso2 for example from CRISPResso2 import CRISPRessoPlot to access those functions, but you're right -- because of argparse we don't really have an easy way to run an entire CRISPResso run. Are there any other specific functions you think would be useful to expose?

[cdustinr]

First of all, I can get by with invoking command line from within python. Or the more hacky version I'm currently using:

my_sys_argv = sys.argv       #store current sys.argv 
sys.argv = kwargs            #rewrite sys.argv with CRISPResso kwargs: amplicon sequence, gRNA sequence, file names
crispressocore.main()
sys.argv = my_sys_argv       #restore sys args

But I'm primarily thinking of the main CRISPResso functionality (aligning reads, calling variants, producing reports). From my perspective, my pipeline includes scraping info from our databases to populate CRISPResso parameters, preprocessing reads, calling CRISPResso, and some re-organization & analysis of CRISPResso output specific to our needs. So I guess the functions that would be useful to expose would be the primary modules: CRISPRessoCORE, CRISPRessoPooledCore, CRISPRessoBatchCORE, etc. Thought I'm not sure how that's possible without some heavy restructuring.