Skip to main content
Fig. 3 | Journal of Cheminformatics

Fig. 3

From: Towards agile large-scale predictive modelling in drug discovery with flow-based programming design principles

Fig. 3

Solving the problem of duplicated parameters. This figure shows: a How parameters are defined and forwarded in vanilla Luigi; b How parameters are defined and forwarded in SciLuigi, using a wrapping workflow task; and c How SciLuigi also defines dependencies between the data on in- and out-ports rather than only directly between tasks. Commenting in more detail we see in a how every parameter definition has to be repeated for every downstream task, from where it was first defined. In b, we see that when the workflow is wrapped in a task, parameters only need to be duplicated once (in the workflow task). In c, we see how—in SciLuigi—dependencies are defined between in-ports and out-ports instead of directly between tasks. This means that task-specific information about names or positions of out-ports of upstream tasks are not contaminating downstream tasks

Back to article page