avallecam
commented
1 year ago Hi Hugo, thanks for opening this issue.
I want to relate this discussion with the "modular" approach to generating reports, as commented in #41. A feature like this could provide flexibility to the user to join any input-package-output as needed. Splitting the current pipeline will be a great opportunity to add output connections to new and current report pieces.
In parallel, the package (or platform) can provide pre-defined pipelines, given their popularity and usability, or reuse those made by others (like in moveapps shared by @pratikunterwegs ).
Bisaloo
Following #95, it would be good to provide a way to stitch / consolidate pipelines into a single report.
After #95, the "middle-task" pipelines, such as the transmissibility pipeline (Rt estimation) or the severity pipeline (CFR) estimation will start directly from incidence data.
But users might have linelist data instead and still want to go all the data until Rt estimation. In this situation, it would be helpful to provide a way to consolidate the "data preparation / descriptive epi" and the "transmissibility" pipeline into a single report, going all the way from linelist import to Rt estimation.
The way we achieve this has strong implications in terms of package interface and sustainability so the choice needs to be carefully thought through.