Reproducible Science Template

More than 70% of researchers have tried and failed to reproduce another scientist's experiments, and more than half have failed to reproduce their own experiments. [1]

In the field of modeling and simulation, we can and should ensure reproducibility with a fully automated workflow [2] from the input scripts to a finished document. This allows for external verification, greater trust in your results and comparative ease in reconstructing details of your own work in the future.

We would like to establish a working group within BAM that develops a wiki/handbook of possible approaches to tackle these challenges. Everyone should be able to access these documents and will be able to use these as a basis for future projects. The competence center modeling and simulation will coordinate the efforts towards this goal, provide guidance and the infrastructure (e.g. a wiki, repositories). For completing the tasks itself, however, we mainly rely on the work of the interested participants.

Website with further information

Visit https://BAMresearch.github.io/Reproducible-Science/ for further content or build the sphinx-based website locally by running doit (see requirements).

Run the example

To run the example, you can first start our docker container

<host> cd <root of this git repository>
<host> sudo docker run -ti -v $(pwd):/repro fb77/reproducible

in an interactive terminal (-ti) and with the current directory ($(pwd) - use ${PWD} (curly brackets!) on windows powershell) available at /repro within the container. You are now in the fb77/reprodicble docker container. To start the pydoit script, run

<docker> cd repro/example
<docker> doit

Alternatively, you can combine all the steps above into one command

<host> sudo docker run -v "$(pwd)":/repro fb77/reproducible /bin/bash -c "cd repro/example && doit"