Registry WG Session, IVOA Interoperability Virtual meeting, April 2022

Friday April 29 - 15:00 UTC

Notes combined from Etherpad: Theresa Dower, Pierre le Sidaner, Markus Demleitner. Thank you!

Claudia Martens: IVOA resources in B2Find and OpenAire via DataCite

B2Find search portal - search and ingest research metadata across various scientific communities.

One of harvesting methods is OAI-PMH, which works with the IVOA registries

• schema mapping: DataCite (old), DublinCore, other standards

• schema mapping EUDAT Core: experimental format through GAVO registry, includes instrument, etc. Markus is involved in using the EUDAT core metadata, it's a job in progress. Specially for coverage for time, and sky coverage. instrument and resourcetype is now inside
• système no-sql with facet
• Current portal: http://b2find.eudat.eu/dataset?groups=ivoa .New B2Find site coming June 2022 (or later)

OpenAire: another EU-wide search project; papers, 'research output'

• Everything in B2Find published in OAI with various metadata prefixes; harvested to OpenAire as DataCite
• OpenAire use multiple data id not only DOI as Datacite, but it's the one we already provide (via GAVO)

• Some IVOA metadata lost in translation; can be improved when they support EUDAT Core?

Do IVOA records count as 'in' European Open Science Cloud because GAVO registry -> B2Find -> OpenAire?

• Search Portal (yes), Marketplace (EOSC members have no relevant services), not working well.

• Update in September: may make this more useful?

• FAIRCORE 4 EOSC: proposal to expand service definitions, will this make ours usable?

• So, yes, but groups becoming formal 'EOSC Provider' to onboard services and products would improve inclusion in upcoming projects. How do we do that?

Questions/comments:

* Call from Markus to add temporal coverage to registry records - to pull into these searches

* Mark Allen: working with EOSC policy and other EU projects: how do we onboard IVOA / registry as a whole? Some projects and data services already in (SIMBAD), but how do we do this as a whole?

* Tess Jaffe: similar to some NASA open science plans?

Markus Demleitner: RegTAP 1.2

RegTAP 1.2 Working Draft soon: STC Tables and GloTS replacements

STC tables:

• Example implementation at: dc.g-vo.org/tap

• We need more implementations before standard! And this means MOC support!

MOC takeup issues:

• MOC in ADQL (not in 2.1?)

• MOC in VOTable (this goes in DALI -- is in DALI 1.2 per Pat Dowler)

• MOC in DB: we have these in pgSphere. what about SQL Server?

Replacing GloTS (evolution of Glots used for finding services, harvest /tables from tap services):

• Can we offer a view of TAP-queryable tables?
• Needs to filter duplicate tables (from tableset of TAP service and from resources with aux TAP capabilities)

• proposal: view server side? requires UNION, not supported in ADQL for client side

• Open problems:
  • Some services do not maintain their tablesets in the registry (problem of maintaining tableset on registry resource, difficult if you don't have an integrated publishing registry)

• • tableset not 100% well defined - can contain output tables we want to keep out of tap_table (???)

• • There's no way to get originating resource

Questions/comments:

• Dower: See the sql/etc for the view build?

Hendrik Heinl: Documentation and Tutorials in the Registry

From Edu note to TextRegExt in RegTAP?

Based on Edu note: https://ivoa.net/documents/Notes/EDU/index.html

Use cases: Curated tutorial collections

• There is new demand

• Especially how do we find tutorials on how to publish?

• These become outdated - how can we track this usefully?

Proposal: TextRegExt - from Chapter 2 of the Note

One rendering of this data: https://dc.g-vo.org/VOTT

Questions/Comments:

• Janet Evans: Can harvest these all into a web collection

• How do we curate these to be useful and validated?

• Markus: sort by contentlevel, validation tags, author, etc

Chloé Azria: Registering EPN-TAP services using DaCHS

Work translating planetary schema/vocabulary (EPN Core) to TAP/resources tags

UAT keyword subjects (https://astrothesaurus.org)

Provenance issues: only one naming authority, several servers

Findability issues with first pass of resource names/titles

• don't find own services w/default TOPCAT inclusion
• problem discovering service EPN-TAP as collection as they have each one specific subject. The more convenient client is TOPCAT, when venus is a subject in VODataResource, it's not visible in the query
• Also a Topcat feature, when you select a resource inside a TAP VODataService topcat select the first table alphabetically not the one dedicated to the selected ressource

• debugging with TOPCAT / ADQL

Questions/Comments:

• Mark Taylor: will follow up (with Markus) on TOPCAT findability, what tags show up where?

• Claudia: Can we rely on everyone using UAT for subjects?

• Markus / Theresa / Chloe: Not technically required but we are all working on it!

• Baptiste: the issues with search are also in TAP Handle - this may be the CatalogResource type?

• Markus: pyvo pull request for registry search - way to fix findability, maybe?

More info on that: https://blog.g-vo.org/towards-data-discovery-in-pyvo.html