Philip Durbin 🚀 (Jan 12 2024 at 19:38): Philip Durbin 🚀 (Jan 12 2024 at 19:38):I could tell you about astronomy features of Dataverse if you like. :grinning:
Jutta Schnabel (Jan 12 2024 at 19:40):I saw that there is a metadata set for astronomy :) - do you have any integrations with virtual observatory protocols?
Philip Durbin 🚀 (Jan 12 2024 at 19:43):@Jutta Schnabel well, we draw on VO standards for metadata at least:
"Astronomy and Astrophysics Metadata (see .tsv version): These metadata elements can be mapped/exported to the International Virtual Observatory Alliance’s (IVOA) VOResource Schema format and is based on Virtual Observatory (VO) Discovery and Provenance Metadata."
https://guides.dataverse.org/en/6.1/user/appendix.html#supported-metadata
Philip Durbin 🚀 (Jan 12 2024 at 19:43):What sort of integration do you have in mind?
Jutta Schnabel (Jan 12 2024 at 19:45):At the moment we are running a VO server in parallel, to make data accessible through VO protocols like TAP. But naturally it would be nice to "somehow" integrate the two. But at that point I really have no clue how that might look like.
Jutta Schnabel (Jan 12 2024 at 19:46):Speaking the same metadata language seems like a good start.
Philip Durbin 🚀 (Jan 12 2024 at 19:46):Table Access Protocol? This? https://www.ivoa.net/documents/TAP/
Philip Durbin 🚀 (Jan 12 2024 at 19:47):I mean, we love tabular data.
Philip Durbin 🚀 (Jan 12 2024 at 19:47):So maybe we should support this.
Jutta Schnabel (Jan 12 2024 at 19:47)::innocent:
Philip Durbin 🚀 (Jan 12 2024 at 19:48):XML-based? Is that right? From a quick look at https://www.ivoa.net/documents/TAP/20190927/REC-TAP-1.1.html
Jutta Schnabel (Jan 12 2024 at 19:49):I think yes ...
Jutta Schnabel (Jan 12 2024 at 19:50):I would believe the docs :)
Philip Durbin 🚀 (Jan 12 2024 at 19:51):How is the data stored in your VO server? In a database? As files?
Jutta Schnabel (Jan 12 2024 at 19:52):They are stored as files, but the content can be queried through the TAP service. Again I mostly know it from the user's perspective.
Philip Durbin 🚀 (Jan 12 2024 at 19:52):Gotcha. What file format? CSV? Excel? FITS?
Jutta Schnabel (Jan 12 2024 at 19:53):csv, text, fits
Jutta Schnabel (Jan 12 2024 at 19:53):(and some more)
Philip Durbin 🚀 (Jan 12 2024 at 19:53):Very interesting.
Philip Durbin 🚀 (Jan 12 2024 at 19:54):I wonder if it would make the most sense to keep running the VO server next to Dataverse.
Philip Durbin 🚀 (Jan 12 2024 at 19:54):And let it continue to be good at what it does.
Philip Durbin 🚀 (Jan 12 2024 at 19:54):And integrate the two somehow.
Jutta Schnabel (Jan 12 2024 at 19:54):At the moment that is a bit further down on the to-do-list.
Philip Durbin 🚀 (Jan 12 2024 at 19:54):Is the VO server open source?
Jutta Schnabel (Jan 12 2024 at 19:55):As there is a lot of infrastructure stuff going on, it is probably better then trying to add all the features. And make it findable through dataverse.
Jutta Schnabel (Jan 12 2024 at 19:56):yes, we are using DaCHS: https://dachs-doc.readthedocs.io/
Philip Durbin 🚀 (Jan 12 2024 at 19:56):Yeah, that's what we've always done. Try to integrate as much as possible.
Philip Durbin 🚀 (Jan 12 2024 at 19:58):Hey, they support Docker, for dev, at least: https://dachs-doc.readthedocs.io/install.html#docker
Jutta Schnabel (Jan 12 2024 at 19:58)::) This is how we are running it...
Philip Durbin 🚀 (Jan 12 2024 at 20:01):The docs don't seem to talk about integrating with data repositories like Dataverse. I guess we'd be the first. :grinning:
Jutta Schnabel (Jan 12 2024 at 20:08):I am not aware that there is any integration yet, the software packages used are generally focused only on serving the VO, not being an add-on for a larger data repository.
Philip Durbin 🚀 (Jan 12 2024 at 20:22):Would it make sense to have some files in both systems? Push or pull files to Dataverse for preservation or to VO for exploration?
Jutta Schnabel (Jan 12 2024 at 20:26):For us that is not clear yet, as it also gets complicated because we serve different science communities (astrophysicists and particle physicists) and they use partially the same data, but in different (software and analysis) environments. And in the VO, the data format development for this kind of data is ongoing, so there is no established best practice there yet.
Jutta Schnabel (Jan 12 2024 at 20:27):I suppose for now some kind of dublication is fine.
Philip Durbin 🚀 (Jan 12 2024 at 20:29):A common model for Dataverse is to be the place that a file is published and maybe even gets a DOI. A lot of systems push files into Dataverse for this reason. Often the other system is more for active data, if that makes sense, instead of published data.
Last updated: Nov 01 2025 at 14:11 UTC