# Snakamake Things

# Workflow logs

Barring a few exceptions, most of Harpy's options are Snakemake workflows. This means we are all at the mercy of how Snakemake operates, which includes the .snakemake/ folder in your project directory. That folder contains all sorts of things necessary for Snakemake to do its magic. However, as a convenience, Harpy workflows will also create a copy of the Snakemake workflow log (all those things that print on screen when snakemake is running) in a workflow's output directory. These logs are found in OUTDIR/logs/snakemake and are named workflow.X.DATE.log, where workflow is the harpy workflow (qc, sv_naibr, etc.), X is the attempt number (given by X, e.g. 4), and DATE is given as MONTH_DAY_YEAR. As an example, using the default settings of harpy qc, you will find your workflow's log in QC/logs/snakemake/qc.1.5_13_2024.log.gz.

Design choices

Snakemake prints a lot of text during runtime. For some workflows, the resulting log files could occupy >1 gigabyte of hard drive space 😱. For this reason, Harpy gzip compresses the log file and deletes the original Snakemake logfile that you would expect to find in .snakemake/logs/.

# Adding Snakemake Parameters

Harpy relies on Snakemake under the hood to handle file and job dependencies. Most of these details have been abstracted away from the end-user, but most Harpy modules have an optional flag --snakemake that you can use to augment the Snakemake workflow if necessary. The full list of Snakemake command line options can be found here. Whenever you use this flag, your argument must be enclosed in quotation marks, for example:

harpy qc --snakemake "--dry-run" rawseq

This means you can add several Snakemake arguments at once, as long as the entire thing is enclosed in quotes:

harpy qc --snakemake "--dry-run --debug --shadow-prefix /scratch" rawseq

# Common use cases

You likely wont need to invoke --snakemake very often, if ever. However, here examples of some possible use cases for this parameter.

Sometimes you want to generate a specific intermediate file (or files) rather than running the entire module to completion. For example, you want the beadtag report Harpy makes from the output of EMA count. To do this, just list the file/files (relative to your working directory) without any flags. Example for the beadtag report:

harpy align bwa -g genome.fasta -t 4 --snakemake "Align/ema/reports/bxstats.html" QC/

This of course necessitates knowing the names of the files ahead of time. See the individual workflows for a breakdown of expected outputs.

# `--shadow-prefix <dirname>`

If running Harpy on an HPC, your system administrator may enforce a policy that all data needs to be moved to a particular network-attached storage for execution. On some systems this is called a SCRATCH/ drive (or something similar) and files are automatically deleted from that drive after the completion of an HPC-scheduled job. Rather than manually adding and removing your files into that workspace storage to make sure you keep the output of your work while not running afoul of your HPC's policy, you may use --shadow-prefix <dirname> where <dirname> is the path to the mandatory directory you need to work out of. By configuring this "shadow directory" setting, Snakemake will automatically move the files in/out of that directory for you: