whole-tale / wt-prov-model

Experiments, design documents, and prototypes supporting a provenance model for Tales and runs.
MIT License
0 stars 1 forks source link

Lineage queries and visualizations #14

Open tmcphillips opened 4 years ago

tmcphillips commented 4 years ago

Using the process-data graph extracted from the system-level trace we can enable users to investigate how particular data products were computed during the execution of a Tale. For example, we can provide customizable reports that answer (with visualizations and/or with tables of results) questions of the following kind:

For a particular output (file, or set of files) produced by a Tale, what were the inputs (files, or sets of files) input to the Tale that the output depended upon?

What were the computational steps and flows that led from inputs of a Tale to a particular output?

For a particular input of a Tale, what outputs of the Tale did that input contribute to (directly or indirectly) or affect in some other way?

For a particular output, what were the software packages that the steps in the lineage of that output depended upon (which may be a subset of the dependencies of the Tale as a whole)?

What were the outputs of a Tale that depended on computational steps that employed a particular software package (or version of that package)?