You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The way we currently present reproducibility is with three things: code, environment and workflow. The way we present these through the tools is, however, a little fragmented. We go from code (Git) to environment (Conda) to workflow (Snakemake/Nextflow) to code and/or workflow depending on how you view it (Quarto/Jupyter) to environment (Docker/Apptainer). There was some discussions during the course to re-order this to be more consistent/linear in terms of what thing we are working with. I generally thing this is a good idea, depending on if we can get a better order going and if that order works with the logistics of making a five-day schedule work without too many breakups between days for each session.
Going by which tutorials mention the others, the simplest solution schedule-wise might actually be just switch the container day (fourth day) with the workflow day (second day). Most other things are mentioned by the workflow managers (as they utilise e.g. containers and Conda environments). The Jupyter tutorial mentions containers in its extra material as well. The container tutorial mentions Jupyter twice (both linking to Jupyter base images, which could probably be moved or removed) and Quarto once (as having used Quarto to write the supplementary material for the case study). The biggest problem here would be the mentions of Snakemake and Nextflow throughout the container tutorial. If we want to do this change we would have to re-write sections of the container tutorial, but the question is if we want to do this regardless, see #289.
The text was updated successfully, but these errors were encountered:
I think it would be good if it went
Data management -> Git -> Conda/Pixi -> Containers -> Notebooks -> Workflows
The container part could then hopefully then be reduced to only describe what containers are, how to find them (Biocontainers/Seqera Containers), how run/reuse them ( in bash ), and for the advanced people (i.e extra material), how to write and build a Dockerfile.
The way we currently present reproducibility is with three things: code, environment and workflow. The way we present these through the tools is, however, a little fragmented. We go from code (Git) to environment (Conda) to workflow (Snakemake/Nextflow) to code and/or workflow depending on how you view it (Quarto/Jupyter) to environment (Docker/Apptainer). There was some discussions during the course to re-order this to be more consistent/linear in terms of what thing we are working with. I generally thing this is a good idea, depending on if we can get a better order going and if that order works with the logistics of making a five-day schedule work without too many breakups between days for each session.
Going by which tutorials mention the others, the simplest solution schedule-wise might actually be just switch the container day (fourth day) with the workflow day (second day). Most other things are mentioned by the workflow managers (as they utilise e.g. containers and Conda environments). The Jupyter tutorial mentions containers in its extra material as well. The container tutorial mentions Jupyter twice (both linking to Jupyter base images, which could probably be moved or removed) and Quarto once (as having used Quarto to write the supplementary material for the case study). The biggest problem here would be the mentions of Snakemake and Nextflow throughout the container tutorial. If we want to do this change we would have to re-write sections of the container tutorial, but the question is if we want to do this regardless, see #289.
The text was updated successfully, but these errors were encountered: