D101 Visual workflow editor¶

Task leader: ICM. Participants: ICM, CNR, ARC.

• Aim: Make Information Inference Service (IIS) more flexible.
• Description:
  • Implement tools for enhanced resource management within IIS.
  • Implement “training workflows” to automate calibration of machine learning algorithms in order to accommodate to Information Space data changing in time.
  • Implement IIS-internal “visual workflow editor” for an easier way to manage data flow between modules (i.e., functionalities) of IIS.
  • Implement tools to provide an easier feedback from OpenAIRE portal on modules of IIS.

Task Timeline (Including Deliverables & Milestones)¶

• Aug 2015 (deliverable D10.1): “Visual workflow editor”: report on design, implementation, rationale, and main functionalities of the solution.
• Dec 2015 (milestone M10.1): “Visual workflow editor”: production version of the solution integrated with IIS.
• Aug 2017 (deliverable D10.4): “Training workflows”: report on architecture and implementation
• Dec 2017 (milestone M10.4): “Training workflows”: production version of the solution integrated with the rest of the system.

Areas of priority (where to concentrate first)

• Finish activities (integrations etc.) that were started in OpenAIRE+.
• Migrate code to GitHub.
• Deliver deliverables explicitly required by DOW (“visual workflow editor”, “training workflows”).
• Improve flexibility and ease of use of IIS.

Forseen Integration with other Work Packages and Tasks¶

• The functionalities delivered in this tasks influence all other tasks in this Work Package.

Communication Strategy: when and how to raise awareness among consortium of updates in task¶

• We plan to continue the tradition established in OpenAIRE+ project of holding monthly technical “status update” teleconferences for parties interested in developments in the work package.
• Communication with technical teams through openaire-mining@bwnmail.icm.edu.pl mailing list, issue tracker, mail, and wiki pages (or some similar system).
• The code will be published and later developed in the most popular open code repository: GitHub.
• At least one scientific paper describing functionality of the system is planned to be published.