CodeMeta-Metadata for Software and displayFormat for controlledVocabularies

[Dorothea Iglezakis]

After we had the first software dataset for our new Dataverse instance (https://darus.uni-stuttgart.de), we created a metadata block for software based on CodeMeta (https://codemeta.github.io/terms/). In this context, we noticed that Dataverse ignored the displayFormat information for the codeMetaDevelopmentStatus field. 

Is this because the field has a controlled vocabulary?  

Is anyone else interested in or working at metadata for research software? 

Thanks and best wishes,

Dorothea Iglezakis

University of Stuttgart

[Philip Durbin]

Hi Dorothea,

First of all, thanks for working on a custom metadata block for software! I feel like https://github.com/IQSS/dataverse/issues/2739 and https://github.com/IQSS/dataverse/issues/5372 about making it easier to deposit code into Dataverse from GitHub have been a bit blocked because we'd like to express metadata about software and haven't had a place to put it, if you will. I'm glad you're looking at CodeMeta and I'd be happy to hear if there are another other standards for software metadata that you are evaluating.

With regard to your specific question about codeMetaDevelopmentStatus and displayFormat, I loaded up your codemeta.tsv file on a test server and I think I see what you're up to, which is really cool! It looks like you're taking a controlled vocabulary of terms like "concept", "active", "unsupported", etc. and trying to put icons next to these terms. For others, here's how these icons look:

- https://www.repostatus.org/badges/latest/concept.svg
- https://www.repostatus.org/badges/latest/active.svg
- https://www.repostatus.org/badges/latest/unsupported.svg

I could be wrong about this but I don't believe it's supported to add images to metadata fields like you're trying to do. Could you please open an issue about this at https://github.com/IQSS/dataverse/issues so developers can discuss it and possibly estimate the amount of time it would take to support this? Or maybe we'll figure out it *is* supported and help you with the syntax. :)

I'd also like to put on your radar that at least two other installations of Dataverse have put some effort into supporting metadata for software.

- Johns Hopkins created https://github.com/jhu-sheridan-libraries/dataverse-ansible-role/blob/v1.1/roles/dataverse/files/dv-mdblocks/jhuSoftware.tsv and you can see it being used in this dataset, for example: https://archive.data.jhu.edu/dataset.xhtml?persistentId=doi:10.7281/T1/D06YVM
- CIMMYT changed "datasets" to "datasets & software" at https://data.cimmyt.org so I assume they're thinking about software a lot! We're working with them on merging other pull requests right now but I'll try to remember to ask them if they've done any work on a metadata block for software.

Thanks!

Phil

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/d1e7a772-e9fe-41eb-892a-af9ab4aa90c4%40googlegroups.com.

--

Philip Durbin
Software Developer for http://dataverse.org
http://www.iq.harvard.edu/people/philip-durbin

[Philipp at UiT]

Phil, maybe you could add the Stuttgart Dataverse to the Dataverse installations map? (https://darus.uni-stuttgart.de)

Best, Philipp

onsdag 16. oktober 2019 14.36.27 UTC+2 skrev Philip Durbin følgende:

Hi Dorothea,

First of all, thanks for working on a custom metadata block for software! I feel like https://github.com/IQSS/dataverse/issues/2739 and https://github.com/IQSS/dataverse/issues/5372 about making it easier to deposit code into Dataverse from GitHub have been a bit blocked because we'd like to express metadata about software and haven't had a place to put it, if you will. I'm glad you're looking at CodeMeta and I'd be happy to hear if there are another other standards for software metadata that you are evaluating.

With regard to your specific question about codeMetaDevelopmentStatus and displayFormat, I loaded up your codemeta.tsv file on a test server and I think I see what you're up to, which is really cool! It looks like you're taking a controlled vocabulary of terms like "concept", "active", "unsupported", etc. and trying to put icons next to these terms. For others, here's how these icons look:

- https://www.repostatus.org/badges/latest/concept.svg
- https://www.repostatus.org/badges/latest/active.svg
- https://www.repostatus.org/badges/latest/unsupported.svg

I could be wrong about this but I don't believe it's supported to add images to metadata fields like you're trying to do. Could you please open an issue about this at https://github.com/IQSS/dataverse/issues so developers can discuss it and possibly estimate the amount of time it would take to support this? Or maybe we'll figure out it *is* supported and help you with the syntax. :)

I'd also like to put on your radar that at least two other installations of Dataverse have put some effort into supporting metadata for software.

- Johns Hopkins created https://github.com/jhu-sheridan-libraries/dataverse-ansible-role/blob/v1.1/roles/dataverse/files/dv-mdblocks/jhuSoftware.tsv and you can see it being used in this dataset, for example: https://archive.data.jhu.edu/dataset.xhtml?persistentId=doi:10.7281/T1/D06YVM
- CIMMYT changed "datasets" to "datasets & software" at https://data.cimmyt.org so I assume they're thinking about software a lot! We're working with them on merging other pull requests right now but I'll try to remember to ask them if they've done any work on a metadata block for software.

Thanks!

Phil

On Mon, Oct 14, 2019 at 4:46 AM Dorothea Iglezakis <dorothea...@ub.uni-stuttgart.de> wrote:

After we had the first software dataset for our new Dataverse instance (https://darus.uni-stuttgart.de), we created a metadata block for software based on CodeMeta (https://codemeta.github.io/terms/). In this context, we noticed that Dataverse ignored the displayFormat information for the codeMetaDevelopmentStatus field. 

Is this because the field has a controlled vocabulary?  

Is anyone else interested in or working at metadata for research software? 

Thanks and best wishes,

Dorothea Iglezakis

University of Stuttgart

--
You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-community+unsub...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/d1e7a772-e9fe-41eb-892a-af9ab4aa90c4%40googlegroups.com.

[Philip Durbin]

Well, being on the map of Dataverse installations is opt-in so I would never simply add an installation. I did reach out to Dorothea on the side and they do want to be on the map some day but have other priorities at the moment. No rush. No pressure. :)

Dorothea, I'm mostly writing to say thanks for creating the "I want to display values of a controlled vocabulary as images" issue at https://github.com/IQSS/dataverse/issues/6289

Finally, speaking of metadata blocks, I thought I'd repeat something Stefan Kasberger of pyDataverse fame said[2] on Thursday in IRC:

"would be great, to collect the metadata blocks in a github repo or so. i plan to have all config related files in a private repo, so i can use it for build/test activities. a master repo from iqss, developed and maintaned together, would be very helpful for that. maybe just one for metadata blocks to beginn with. i like granular repos for just one purpose as a foundation... at least to have a list of metadata blocks already created, so you have orientation when you want to work one out for your own institution. and maybe later on then agree upon some standards"

The idea of having a master list of metadata blocks is a good one. (Seven blocks are enabled out of the box but there are others[2] out there in various states of completion.) My instinct is always to start with a spreadsheet when making lists (I'm thinking about some sort of "status" or "maturity" column) but I'm open to ideas, such as the one above about a dedicated repo.

Thanks,

Phil

1. http://irclog.iq.harvard.edu/dataverse/2019-10-17#i_108936

2. A metadata block for Darwin Core comes to mind: https://github.com/IQSS/dataverse/issues/6243

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/d1e7a772-e9fe-41eb-892a-af9ab4aa90c4%40googlegroups.com.

--

Philip Durbin
Software Developer for http://dataverse.org
http://www.iq.harvard.edu/people/philip-durbin

--

You received this message because you are subscribed to the Google Groups "Dataverse Users Community" group.

To unsubscribe from this group and stop receiving emails from it, send an email to dataverse-commu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dataverse-community/035a8206-4406-4010-9751-1547cad0a5bf%40googlegroups.com.