Initial roadmap for high-level cyberinfrastructure components of analysis system

[BenGalewsky]

Assumptions

High-level cyberinfrastructure components include:

• INSPIRE
• HEPData
• CERN Analysis Preservation framework
• CERN Open Data
• REANA
• INDICO
• Zenodo
• RECAST
• GitHub and GitLab at CERN

These projects are pre-existing and funded through a variety of mechanisms (sometimes including IRIS-HEP). IRIS-HEP's role is largely about

• integration with analysis system tools (eg. pyhf and HEPData)
• enhancing interoperability between those components (eg. RECAST, CAP, REANA, HEPData)

Acceptance criteria

• figure or other documentation describing interaction of components needed to achieve specific analysis user stories (eg. on-boarding, reinterpretation, publishing, preservation).
• pointeres to specific engagements with these cyberinfrastructure components (eg. via GitHub issues)

[matthewfeickert]

@cranmer @gordonwatts Can you give some feedback on what the goals of this Milestone are? It wasn't really clear to anyone in the AS meeting on 2020-02-05, so as they are due soon it would be good to understand how to move forward.

Edit: Comments from @gordonwatts and @msneubauer later in the meeting.

What infrastructure are you going to host all the different parts of the AS software pipeline on?

c.f. River at Chicago

@mweinberg2718 can you add some comments here as well about what you've done with Coffea and River?

[gordonwatts]

I would guess we need text that is about 1/2 a page for this - since it is mostly Hardware. There were some conclusions from the NYU blueprint meeting (@msneubauer), that might be relevant. It might be useful to use the ecosystem coherency to help here too.

[cranmer]

Had nice discussion with @lukasheinrich and @Sinclert about this. This should focus mainly on HEPData, INSPIRE, CAP, RECAST, etc. pyhf and fitting service are also related.

[cranmer]

See issue with HEPData: HEPData/hepdata#164

[cranmer]

See issue: HEPData/hepdata#163

[cranmer]

see issue: indico/indico#4293

[cranmer]

See HEPData Advisory Board agenda: https://conference.ippp.dur.ac.uk/event/875/

[cranmer]

For reference, many ideas of this road map first discussed at AAHEP7 in 2014
https://indico.cern.ch/event/262430/sessions/48420/#20140402

[cranmer]

CERN press on publishing pyhf on HEPData:
https://home.cern/news/news/knowledge-sharing/new-open-release-allows-theorists-explore-lhc-data-new-way

[cranmer]

Initial roadmap tied to reinterpretation user story:
https://github.com/iris-hep/as-user-facing/blob/master/reinterpretation.md