gardnersnake 0.1.1

Utilities for writing concise snakemake workflows

gardnersnake

Utilities for writing concise snakemake workflows

Table of Contents

1. Introduction
2. Class Objects
3. Command Line Tools

Introduction

Snakemake is an incredibly powerful workflow manager that enables computational biologists to produce clear, reproducible, and modular analysis pipelines using a familiar Python-based grammar. Unfortunately, the bioinformatics tools that we'd like to utilize inside of our Snakemake workflows are often a bit less well-behaved. Gardnersnake is a small package built on the python standard library (Python 3.6) that aims to make handling this wide variety of tools easier and more compact, especially when working on cluster-based systems.

Class Objects

gardnersnake.ConfigurationHelper()

The foundational object defined in gardnersnake is the ConfigurationHelper Class. At instantiation ConfigurationHelper takes a single argument, cfg_dict, which should be the config snakemake variable capturing the passed workflow configuration file.

Command Line Tools

check_directory

Many bioinformatics tools produce directories of various structure with large numbers of output files. Rather than require Snakemake to keep track of these outputs as global outputs, the check_directory command-line utility validates output directories against a known set of files, and returns a small file containing a return code (0) if the directory of interest was successfully validated. check_directory throws an error and does not return the return code file if it is unable to validate the contents according to the given requirements.

The options and requirements are specified in the usage message and can be retrieved using the -h or --help flags.

check_directory --help
usage: check_directory [-h] [--strict] [-o OUT] FILES [FILES ...] DIR

validates dynamic directory contents against expectations

positional arguments:
  FILES                 set of filepaths to check against dir contents
  DIR                   filepath of directory to verify

optional arguments:
  -h, --help            show this help message and exit
  --strict              directory should contain only the passed files
  -o OUT, --output OUT  name of return code output filei

Positional Options \

• FILES [required] a list of whitespace separated files to search for in the passed directory. these file names should be specified without their path extensions (i.e. a file whose full path is /home/user/analysis/myoutputs/output1.txt should be passed as output1.txt if the DIR is indicated to be /home/user/analysis/myoutputs/)
• DIR [required] is the full path of the directory to verify. ~/ conventions are acceptable but shell variable syntax such as $WORKDIR are not supported. Relative path functionality remains in active development but is not guaranteed to work as of the current version (0.1.0)

Flagged Options \

• --ouput -o [required] specifies the name of the file generated (containing the return code) when the passed directory is successfully validated.
• --strict [optional] indicates that the passed directory should only contain the files listed in the FILES positional argument, and no other files or subdirectories. the default setting, nonstrict will validate directories containing extra files so long as the required ones are present. This gives the user the ability to be more or less permissive with their checks. Typical usage may look like:

check_dir -o rc.out --strict output1.txt output2.txt ~/myanalysis/outputs/

which should return a file called rc.out if the folder ~/myanalysis/outputs/ has exactly two files --> output1.txt and output2.txt