Hacking Session: Building a SKOS representation of the IUPAC Goldbook

[StroemPhi]

context

IUPAC plans on updating its workflows around the Gold Book. As part of this they would like to provide it as a SKOS vocabulary besides the current website. The workshop cold be used to build a prototype of such a SKOS vocabulary using the Goldbooks web API.

Figure 1. Example Representation of a Gold Book Term in SKOS (see below)

expected outcome of the workshop

• develop a prototype Gold Book SKOS vocabulary
• identify key needs and issues in using the terminology for IUPAC project to more systematically address
  • unique ID per term
  • classes of potential interest to the semantic community (e.g., instruments, techniques, etc.)

prerequisites issues/tasks

• identify key SKOS properties representing the structure and provenance of the terminology (see Figure 1 below)
  • skos:collection for formal source glossary
  • identify existing components of term entries (e.g., preferred terms, synonyms, definition, notes, examples, ‘see also’ related terms)
  • also considering symbols, units, quantities, can we define/develop notation types (see example 21, https://www.w3.org/TR/skos-reference/#notations)
    • ChemicalSymbolNotation
    • QuantitySymbolNotation
    • UnitSymbolNotation
• determine the grouping concepts needed, to allow clustering of concepts around a core concept e.g. solubility
  • analyze IUPAC glossary sources to identify key scientific grouping concepts / broader terms (in progress by SJC, LRM)
  • review mapping to use cases that reference Gold Book terminology to identify potential grouping concepts (e.g., CHMO, ChemDCAT-AP, PaNET)

implementation issues/tasks

• Trying to reuse the pipeline around the Voc4Cat to build the SKOS representation of the Goldbook via fetching its terms from its web API (https://goldbook.iupac.org/pages/api)
  • Other approaches suggested in comments
• handle the different meaning possible for a concept based on a specific context

Figure 1. Example Representation of a Gold Book Term in SKOS

---

How to use this discussion page

I've created issue #22 with the intention to help the session chair better structure the hacking session during the workshop.
Please keep using this discussion for brainstorming before the workshop to add any comments, thoughts, links and tasks you think should be discussed / worked on during the session.

Being able to reply to posts directly keeps things a bit more organized here.

We'll destill this then into more actionable items within #22 shortly before the workshop.

[cthoyt]

I already did something like this to support the SSSOM mapping from CHMO and wrote it up in https://cthoyt.com/2026/06/03/owl-to-skos.html.

here's several IUPAC Goldbook produced by https://github.com/biopragmatics/pyobo/blob/main/src/pyobo/sources/goldbook.py, which fetches terms from the Goldbook web API. The results are committed to the OBO Database Ingest resource: https://github.com/biopragmatics/obo-db-ingest/tree/main/export/goldbook:

• OBO: https://github.com/biopragmatics/obo-db-ingest/raw/refs/heads/main/export/goldbook/goldbook.obo
• OFN: https://github.com/biopragmatics/obo-db-ingest/raw/refs/heads/main/export/goldbook/goldbook.ofn
• OWL: https://github.com/biopragmatics/obo-db-ingest/raw/refs/heads/main/export/goldbook/goldbook.owl
• SKOS: https://github.com/biopragmatics/obo-db-ingest/raw/refs/heads/main/export/goldbook/goldbook.skos.ttl

Note: this comment was updated following discussion below to reflect updates suggested by David and Philip

[StroemPhi]

Great blogpost! Thank you for this input. I couldn't open the OFN from which you want to derive the SKOS, since I cannot load it into Protégé, to check what you've done in a more readable way though :(

[StroemPhi]

For the SKOS vocabulary, I'd use https://www.w3.org/TR/skos-reference/skos.html#altLabel as predicate for the synonyms and rather https://www.w3.org/TR/skos-reference/skos.html#definition instead of skos:scopeNote for the definition.

[dalito]

I was also going to comment on the unconventional use of SKOS. You could look at the vocpub profile (or our vp4cat version of it for voc4cat) or what Jacob supports in his SKOS/jskos chain.

In general I think that vocabularies with looser semantics have its place. The Goldbook is in some parts not yet conforming with the idea of a machine readable vocabulary/terminology: It puts sometimes two or more meanings under the same DOI which is OK if the idea is to publish a definition with bibliographic information. However, it breaks the idea of a terminology to provide a unique ID per term. So an initial step would be to identify and split such entries.

I like the idea to devote some time in the workshop to the topic.

[cthoyt]

i updated the goldbook workflow to make sure the OFN and OWL are valid in biopragmatics/pyobo#510. I also fixed the minor suggestion about the definition

now, OBO Database Ingest also has the SKOS artifact, see:

• OBO: https://github.com/biopragmatics/obo-db-ingest/raw/refs/heads/main/export/goldbook/goldbook.obo
• OFN: https://github.com/biopragmatics/obo-db-ingest/raw/refs/heads/main/export/goldbook/goldbook.ofn
• OWL: https://github.com/biopragmatics/obo-db-ingest/raw/refs/heads/main/export/goldbook/goldbook.owl
• SKOS: https://github.com/biopragmatics/obo-db-ingest/raw/refs/heads/main/export/goldbook/goldbook.skos.ttl

[Wiljes]

@dalito wrote

{The Goldbook} puts sometimes two or more meanings under the same DOI.

Could you please privide an example for this?

[egonw]

ping @stuchalk

[StroemPhi]

Thank you @egonw and yes @stuchalk we'd need your input here in terms of getting proper requirements of how the workflow should be.
Here are some of my top of the head questions

• Should this be an independent repo in which some chronjob is fetching the data from the GoldBook Website and then produces the SKOS thesaurus?
  • How is the SKOS version envisioned to be published?
• What PURL scheme should be used, w3id?
• How should we deal with terms that have multiple meaning? --> prevents us from using the current DOI's also as IRIs for the created SKOS concepts, as @dalito pointed out
• How much schould we try to model relations in terms of using skos:broader and skos:narrower? And if we should, what would be the best approach to do this automatically?
• Should we use skos:Collection to group terms that are semantically related? And if we should, what would be the best approach to do this automatically?

@ALL the others: what questions do you have that should be asked before or at the workshop?

[LeahRaeMcEwen]

To
Thank you for the great questions @StroemPhi. Some thoughts below from the IUPAC crew:

• Should this be an independent repo in which some chronjob is fetching the data from the GoldBook Website and then produces the SKOS thesaurus?
  -- LRM/SJC: yes initially
• How is the SKOS version envisioned to be published?
  -- LRM/SCJ: e.g, VocBench, Skosmos, others? We are reviewing possible tools
• What PURL scheme should be used, w3id?
  -- LRM/SJC: how/why PURL scheme useful in this context, what is the need? Can we use current Gold Book ids as the basis?
• How should we deal with terms that have multiple meaning? --> prevents us from using the current DOI's also as IRIs for the created SKOS concepts, as @dalito pointed out
  -- LRM/SJC: practical suggestion to pull these out of the working set for now
  -- LRM/SJC: separately trying to address these at the IUPAC source (but lengthy time) >> one practice we are considering is to add context qualifier to term for distinct definitions
• How much should we try to model relations in terms of using skos:broader and skos:narrower? And if we should, what would be the best approach to do this automatically?
  -- LRM/SJC: suggest to use “skos:related” as a default to start
  -- LRM/SJC: we can do some analysis of the source glossaries, which are arranged broadly by scientific area
  -- LRM/SCJ: it would be great to also look at other projects that are referencing Gold Book terms to hear what types of classes they would like to see, and strategize how we can design further analysis (e.g., PaNeT interested in instruments, techniques)
• Should we use skos:Collection to group terms that are semantically related? And if we should, what would be the best approach to do this automatically?
  -- LRM/SJC: suggest to use “skos:collection” for the source glossaries

[StroemPhi]

I just updated the requirements provided by IUPAC ;)

[Wiljes]

@StroemPhi Where can I find these requirements?

[CPI-Darren]

I may have missed this @StroemPhi but if there isn't a mechanism for dealing with gaps can we address this in the workshop?

Gap - a user is building an ontology using chemical terms. The user wants a chemical term and finds it isn't in the Gold book ... how do we bring it to the attention of IUPAC?

[dalito]

An idea was also to discuss e.g. GitHub based workflows for contributing.

It would be great if you could list also other gaps you were already thinking of.

[StroemPhi]

Good question! @LeahRaeMcEwen will give a brief update of the current status of the Gold Book before this session, which presumably will also adress their plans on how to address this aspect of addig new terms. But I agree, this might be a good requirement to also somehow adress in this session.

[Wiljes]

@StroemPhi wrote:

IUPAC plans on updating its workflows around the Gold Book.

Do we have more information on this? What are these IUPAC workflows? What role does the Goldbook play in these workflows? In which way is a SKOS representation of the Goldbook expected to support or improve these workflows?

Do we have other use cases in which a SKOS representation of the Goldbook would be beneficial?

This would help to define the scope of the SKOS concept scheme and to come up with initial competency questions.

[egonw]

ping @LeahRaeMcEwen