Hacking Session: Prototyping a data model for the IUPAC Commission on Isotopic Abundances and Atomic Weights (CIAAW)

[StroemPhi]

context

The IUPAC Commission on Isotopic Abundances and Atomic Weights (CIAAW) currently provides the standard atomic weights and isotopic abundances values as HTML websites (https://ciaaw.org/isotopic-abundances.htm). IUPAC proposes to digitize the processes behind the publication of these values. We could help them by building a data model that is grounded in chemical ontologies.

expected outcome of the workshop

• Develop a data model prototype for CIAAW that can be used to derive the tables currently publishing these values on the CIAAW website.

prerequisites issues/tasks

• get all necessary information that needs to be represented by this data model prototype from CIAAW/IUPAC
• gather other needed prerequisits

implementation issues/tasks

• discuss in how far LinkML can be utilized as a basis for this prototype by testing this approach.

---

How to use this discussion page

I've created issue #20 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 #20 shortly before the workshop.

[egonw]

Count me in.

[StroemPhi]

Me parking thoughts here:

1. What is the current workflow of the curators of CIAAW?

   • What tools are currently used to source and ingest the data?
     • What are the interface we have / how much tooling is actually needed?
   • How are the samples/measurements sourced?
     • Do they come from trusted labs?
   • What provenance trail is needed regarding the sampling process?
   • Can we assume a CSV based data curation? Or what format is the data coming in?
   • How often does the data get updated and how exactly (batch operation, single direct curation)?

2. What data is processed in the backend and what of that is supposed to be exposed in the frontend?

3. What is the expected way the data is to be published?

   • html, rdf/json-ld, csv/tsv, ??? --> Of course in the context of our workshop we'd want rdf 😉
   • via API or dump?

4. What are the classes/slots we'd need in the datamodel?

# If we were to do it as a LinkML schema

classes:
   Element:
     attributes:
       hasIsotope:
         range: Isotope
   Isotope:
     is_a: Element
     attributes:
       hasAtomicMass:
         range: AtomicMass
       hasAbundance:
         range: IsotopicAbundance
   AtomicMass:
   IsotopicAbundance:
   RepresentativeIsotopicComposition:
   
   # to be continued