A void-generator (For VoID) is to be figured out before a human-readable schema by ShEx is conducted.

[candlecao]

This follows the #27, and https://pypi.org/project/sparql-llm/, in which there is a "SPARQL endpoint schema loader" section with code:

from sparql_llm import SparqlVoidShapesLoader

loader = SparqlVoidShapesLoader("https://sparql.uniprot.org/sparql/")
docs = loader.load()
print(len(docs))
print(docs[0].metadata)

I tried it but it didn't work out while prompted, and fed back: Could not retrieve VoID description for endpoint https://sparql.uniprot.org/sparql/: No VoID description found in the endpoint

In https://pypi.org/project/sparql-llm/, it also tips: Checkout the void-generator project to automatically generate VoID description for your endpoint.

[candlecao]

About the use of VoID, please see the excerpt from the paper: LLM-based SPARQL Query Generation from Natural Language over Federated Knowledge Graphs

The VoID generator is available open source6 and can be used to generate statistics over any SPARQL endpoint. This information allows us to generate simple, human-readable Shape Expressions (ShEx) for each class.

[candlecao]

First, I understood VOID: https://www.w3.org/TR/void/.
To vividly illustrate it, see this example:
There are 2 Graphs in a SPARQL Endpoint, such as:

1. Fruits Graph <https://example.org/void-generator/test/fruits>

prefix ex:<https://example.org/void-generator/test/>


ex:red_apple a ex:Fruit, ex:Apple ;
    ex:color "red" ;
    ex:grows_on ex:apple_tree .

ex:green_apple a ex:Fruit, ex:Apple ;
    ex:color "green" ;
    ex:grows_on ex:apple_tree .

ex:rotton_apple a ex:Fruit, ex:Apple ;
    ex:color "brownish" .

ex:green_pear a ex:Fruit, ex:Pear ;
    ex:color "green" ;
    ex:growns_on ex:pear_tree .

2. Fruit Trees Graph <https://example.org/void-generator/test/trees>

prefix ex:<https://example.org/void-generator/test/>


ex:apple_tree a ex:FruitTree .

ex:pear_tree a ex:FruitTree .

ex:lemon_tree a ex:FruitTree .

And the corresponding VoID info is:

@prefix void: <http://rdfs.org/ns/void#> .
@prefix dcterms: <http://purl.org/dc/terms/> .
@prefix ex: <https://example.org/void-generator/test/> .

# Fruits Graph
<https://example.org/void-generator/test/fruits>
    a void:Dataset ;
    dcterms:title "Fruits Graph" ;
    dcterms:description "This graph contains information about different types of fruits and their properties." ;
    void:uriSpace "https://example.org/void-generator/test/" ;
    void:exampleResource ex:red_apple ;
    void:triples 15 ;
    void:classes 3 ;  # Fruit, Apple, Pear
    void:properties 3 ;  # color, grows_on, growns_on .

# Fruit Trees Graph
<https://example.org/void-generator/test/trees>
    a void:Dataset ;
    dcterms:title "Fruit Trees Graph" ;
    dcterms:description "This graph contains information about different types of fruit trees." ;
    void:uriSpace "https://example.org/void-generator/test/" ;
    void:exampleResource ex:apple_tree ;
    void:triples 3 ;
    void:classes 1 ;  # FruitTree
    void:properties 0 .  # No specific properties in this graph

# Linkset: Relationship between Fruits and Fruit Trees
<https://example.org/void-generator/test/fruits-trees-linkset>
    a void:Linkset ;
    dcterms:title "Linkset between Fruits and Fruit Trees" ;
    dcterms:description "This linkset describes the relationship between fruits and the trees they grow on." ;
    void:subjectsTarget <https://example.org/void-generator/test/fruits> ;
    void:objectsTarget <https://example.org/void-generator/test/trees> ;
    void:linkPredicate ex:grows_on ;
    void:triples 2 .  # red_apple -> apple_tree, green_apple -> apple_tree

[candlecao]

That would be useful for federated query.

[candlecao]

The README.md https://github.com/DDMAL/void-generator/blob/main/README.md has clearly demonstrated it.
Notice: When generating a VoID file for a graph in a SPARQL endpoint, please ensure the file is uploaded to the corresponding endpoint. Only after this step, and with the VoID file uploaded to the endpoint, can we extract the schema (shapes) for the graph. (refer to https://pypi.org/project/sparql-llm/)