Add a gold standard example using data from chemotion repository #139
Conversation
|
|
||
| This folder contains two files: | ||
| - PASTA.eln an export of the standard example of an installation with samples, measurements, devices, ... | ||
| - A gold‑standard sibling triplet consists of an ELN file, a JSON‑LD file, and a Turtle file. The example shows |
There was a problem hiding this comment.
Shouldn't the GoldStandard be something cross-implementation? This indicates that is somehow bound to Pasta
There was a problem hiding this comment.
From my point of view that would make sense. But I cannot / do-not-want-to speak for the entire consortium.
- You can cross-link it from OSL
- if everybody agrees, we can have it cross-implementation
examples/PASTA/README.md
Outdated
| { | ||
| "@type": "Organization", | ||
| "name": "Chemie, Ludwig-Maximilians-Universit\u00e4t M\u00fcnchen, Deutschland", | ||
| "@id": "Organization_Chemie_LudwigMaximiliansUniversitt_Mnchen_Deutschland" |
There was a problem hiding this comment.
For the GoldStandard we can do better and use ROR Ids
There was a problem hiding this comment.
Good catch; did exactly that.
| ``` | ||
|
|
||
| ### goldStandard.eln | ||
| ```json |
There was a problem hiding this comment.
In general this seems much to extensiv for an easy to grasp GoldStandard - e.g. there no additional learning from more than two person objects. Also we should sort the graph, e.g. based on @type
There was a problem hiding this comment.
- The conventional examples are for the learning / understanding. This is a small but real and complete example.
- Since it is real dataset, I also would not like to kick people off the list, ignoring their contribution.
There was a problem hiding this comment.
Is there a major difference to the RO-Create JSON-LD other than beeing framed / nested (by JSON-LDs framing)?
There was a problem hiding this comment.
Yeah, some keys are renamed, some @types have to be changed. There is a readme file to list all those changes.
examples/PASTA/goldStandard.ttl
Outdated
| @prefix schema1: <http://schema.org/> . | ||
| @prefix xsd: <http://www.w3.org/2001/XMLSchema#> . | ||
|
|
||
| <file:///home/steffen/FZJ/DataScience/Repositories/TheELNConsortium/GoldStandard/10.14272/reaction/SA-FUHFF-UHFFFADPSC-GLVNZYODMK-UHFFFADPSC-NUHFF-NSOPS-NUHFF-ZZZ> a schema1:Study ; |
There was a problem hiding this comment.
Shouldn't it be https://www.chemotion-repository.net/home/publications/collections/4916 ?
There was a problem hiding this comment.
I had used rdflib to do the job. Now I did a replace of the first string. Great.
There was a problem hiding this comment.
What is the source of this SHACL shapes?
There was a problem hiding this comment.
The idea was to show a prove-of-concept: json-ld can be validated. Hence, I needed a shapes file that catches errors. And this does that job: I changed the data a few times and the shapes files caught it.
How was it generated: some-LLM created
Generate SHACL shapes from a JSON-LD file by analyzing its RDF structure.
- This implementation first collects the kinds of values observed for each property (Literal/IRI/BlankNode) across all instances of a class and only emits a
sh:datatypeorsh:nodeKindwhen the predicate has a single consistent kind. If kinds are mixed, no kind/datatype constraint is added.
3 participants
@simontaurus @raederan Hey this is the gold-standard example we talked about. Since I was in the vibe, I also created the turtle file. Do they make sense?