GRIT project directory

Contents

• ./activate:

  • Script to activate the environment for this project.
  • Usage: source ./activate

• data/

  • Any input data not created or modified by this project.

• docs/

  • Documentation and descriptions of the project results.

• results/

  • Any result/data created or modified by this project organised in subdirectories representing result components.

• src/

  • Code developed to create the results.

• bin/

  • Any installed software, e.g. with --prefix=$PROJECT_ROOT/bin

• see Wilson et al. 2017 (https://doi.org/10.1371/journal.pcbi.1005510) for details

Version control

• All code and config files created by humans are committed to git.
• Documentation (also in binary formats) is also committed to git.
• To clone the repository with submodules use: git clone --recurse-submodules <url>
• Changes to result files should be mainly represented by the source code, config and build files that created them.

Python and R environments

• This project uses Anaconda as a python and R package manager and python 3.6+.
• Create the environment using (last directory of prefix is name):

  conda env create -f src/conda_environment_base.yml -p bin/conda/ef-base
  

Workflows

• All results are created by snakemake workflows defined in the respective Snakefile.
• Executing a workflow locally on 10 CPUs:

  snakemake -j 10 <rule or file>
  

• Install snakemake slurm interface:

  pip install cookiecutter
  cookiecutter -f --no-input -o $PROJECT_ROOT/src/snakemake_slurm_profiles https://github.com/Snakemake-Profiles/slurm.git
  

Config

• Paths: in Snakefiles
• Constants
• Research assumptions/decisions
• variable names/labels/units
• plot styles/colours