This page is to discuss and explore the possible ideas for Cyberintegrator for BrownDog. From this discussion, we will derive the tangible tasks for BrownDog.
Discussion points:
- Architecture or mechanism (how to)
- Cost in terms of time, Difficulty level
Table of Contents
Running CI workflow as a extractor
The medici extraction service is able to execute CI workflow as a extractor. The CI workflow will be queued at extractor event bus.
Option 1: Medici extractor can be a CI RESTful service client
- Medici extractor uses CI RESTful service endpoints to do
- upload the dataset
- execute the workflow
- retrieve the results from the execution
- put those results to Medici metadata
Running Polyglot by using CI
The polyglot will be run by using CI workflow to have reproducibility.
Option 1: Spit out CI workflow from Polyglot conversion
Running CI workflow on Medici
This is on-demand execution of CI workflow different from extractor idea. The user will see the possible CI workflow to execute on the dataset on the dataset page.
Integration of Ci and Medici
Tight integration of CI and Medici. Ci will use the Medici as a metadata and dataset storage. It may requires to remove "Spring" framework and adopting other "injection" mechanism.
Running Other workflows (Kepler, Galaxy) as CI workflow
CI workflow will have a step that execute other workflow such as Kepler, or Galaxy.