snakemake

[sje30]

How does this compare to snakemake (which I dont [yet] use)?

[shouldsee]

Hi Stephen, a careful answer is not yet available. A short motivation was written long ago http://newflaw.com/blog/posts/0044_review-snakemake/ . But from my current thoughts, here are the pros and cons when compared with snakemake:

Pros:

1. LUCKFILE.py are pure python scripts.
   • pure python, no yaml. flat learning curve for python users
   • custom functions can be easily implemented
   • easy to install and portable
   • debuggable
   • snakemake uses smk file and is a mixture of python and yaml, and runs with snakemake binary as a pseudo-interpreter.
2. Easy to modularize by importing LUCKFILE.py as a python module
   • including sub-snakefile is confusing.
3. Easy to extend because there aren't any confilictng existing features.

Cons:

1. Offers less features than snakemake, meaning luckmake has:
   • no support for cluster scheduling or parallelization
   • no profiling yet.
   • no dashboard to view job status yet.
   • no daemon mode yet
2. documentation is scarce
3. not tested by real use cases thus may be buggy. snakemake has a bigger community and maintaining team.

Glad to discuss further!