gl/tools
SHA256
1
0
Fork 0
Code Actions Activity

.

Browse Source
This commit is contained in:
alex
2026-08-08 23:50:04 -04:00
parent 9177c8902b
commit 79894f654d
26 changed files with 748 additions and 411 deletions
Download Patch File Download Diff File Expand all files Collapse all files
publish/src/rdf/ontology.rs
+147 -83
View File
@@ -4,7 +4,10 @@ use gl_graph::vocab::gl;
use gl_search::language::LanguageCondition;
use iced::futures::TryFutureExt;
use oxigraph::model::vocab::{rdf, rdfs, xsd};
use oxigraph::model::{Dataset, Graph, GraphNameRef, LiteralRef, NamedNode, NamedNodeRef, NamedOrBlankNode, NamedOrBlankNodeRef, Term, TermRef, Triple, TripleRef};
use oxigraph::model::{
Dataset, Graph, GraphNameRef, LiteralRef, NamedNode, NamedNodeRef, NamedOrBlankNode,
NamedOrBlankNodeRef, Term, TermRef, Triple, TripleRef,
};
use oxigraph::sparql::{QueryResults, SparqlEvaluator};
use oxigraph::store::Store;
use std::collections::{BTreeMap, BTreeSet, HashMap};
@@ -14,48 +17,53 @@ use std::sync::LazyLock;
use tracing::{debug_span, field};
const ONTOLOGY_PREFIX: &str = "https://graphofliberty.org/2026/04/ont/";
const ONTOLOGY_GRAPH_NAME: GraphNameRef = GraphNameRef::NamedNode(NamedNodeRef::new_unchecked("https://graphofliberty.org/2026/04/ont"));
const ONTOLOGY_GRAPH_NAME: GraphNameRef = GraphNameRef::NamedNode(NamedNodeRef::new_unchecked(
"https://graphofliberty.org/2026/04/ont",
));
static PREFIXES: LazyLock<BTreeMap<String, String>> = LazyLock::new(|| {
BTreeMap::from_iter([
("rdf", "http://www.w3.org/1999/02/22-rdf-syntax-ns#"),
("rdfs", "http://www.w3.org/2000/01/rdf-schema#"),
("owl", "http://www.w3.org/2002/07/owl#"),
("xsd", "http://www.w3.org/2001/XMLSchema#"),
("ldp", "http://www.w3.org/ns/ldp#"),
("dc", "http://purl.org/dc/elements/1.1/"),
("posix", "http://www.w3.org/ns/posix/stat#"),
(
"ebucore",
"http://www.ebu.ch/metadata/ontologies/ebucore/ebucore#",
),
("prov", "http://www.w3.org/ns/prov#"),
("locid", "http://id.loc.gov/vocabulary/identifiers/"),
("loclang", "http://id.loc.gov/vocabulary/languages/"),
("premis", "http://www.loc.gov/premis/rdf/v1#"),
("premis3", "http://www.loc.gov/premis/rdf/v3/"),
("dcterms", "http://purl.org/dc/terms/"),
("fedora", "http://fedora.info/definitions/v4/repository#"),
("rdaa", "http://rdaregistry.info/Elements/a/"),
("rdac", "http://rdaregistry.info/Elements/c/"),
("rdae", "http://rdaregistry.info/Elements/e/"),
("rdai", "http://rdaregistry.info/Elements/i/"),
("rdam", "http://rdaregistry.info/Elements/m/"),
("rdan", "http://rdaregistry.info/Elements/n/"),
("rdap", "http://rdaregistry.info/Elements/p/"),
("rdaf", "http://rdaregistry.info/Elements/rof/"),
("rdat", "http://rdaregistry.info/Elements/t/"),
("rdau", "http://rdaregistry.info/Elements/u/"),
("rdaw", "http://rdaregistry.info/Elements/w/"),
("rdax", "http://rdaregistry.info/Elements/x/"),
("rdaco", "http://rdaregistry.info/termList/RDAContentType/"),
("rdact", "http://rdaregistry.info/termList/RDACarrierType/"),
("rdamt", "http://rdaregistry.info/termList/RDAMediaType/"),
("rdaft", "http://rdaregistry.info/termList/fileType/"),
("schema", "https://schema.org/"),
("gl", "http://fedora.quill.lan/rest/"),
("glo", ONTOLOGY_PREFIX),
].map(|(k, v)| (k.to_string(), v.to_string())))
BTreeMap::from_iter(
[
("rdf", "http://www.w3.org/1999/02/22-rdf-syntax-ns#"),
("rdfs", "http://www.w3.org/2000/01/rdf-schema#"),
("owl", "http://www.w3.org/2002/07/owl#"),
("xsd", "http://www.w3.org/2001/XMLSchema#"),
("ldp", "http://www.w3.org/ns/ldp#"),
("dc", "http://purl.org/dc/elements/1.1/"),
("posix", "http://www.w3.org/ns/posix/stat#"),
(
"ebucore",
"http://www.ebu.ch/metadata/ontologies/ebucore/ebucore#",
),
("prov", "http://www.w3.org/ns/prov#"),
("locid", "http://id.loc.gov/vocabulary/identifiers/"),
("loclang", "http://id.loc.gov/vocabulary/languages/"),
("premis", "http://www.loc.gov/premis/rdf/v1#"),
("premis3", "http://www.loc.gov/premis/rdf/v3/"),
("dcterms", "http://purl.org/dc/terms/"),
("fedora", "http://fedora.info/definitions/v4/repository#"),
("rdaa", "http://rdaregistry.info/Elements/a/"),
("rdac", "http://rdaregistry.info/Elements/c/"),
("rdae", "http://rdaregistry.info/Elements/e/"),
("rdai", "http://rdaregistry.info/Elements/i/"),
("rdam", "http://rdaregistry.info/Elements/m/"),
("rdan", "http://rdaregistry.info/Elements/n/"),
("rdap", "http://rdaregistry.info/Elements/p/"),
("rdaf", "http://rdaregistry.info/Elements/rof/"),
("rdat", "http://rdaregistry.info/Elements/t/"),
("rdau", "http://rdaregistry.info/Elements/u/"),
("rdaw", "http://rdaregistry.info/Elements/w/"),
("rdax", "http://rdaregistry.info/Elements/x/"),
("rdaco", "http://rdaregistry.info/termList/RDAContentType/"),
("rdact", "http://rdaregistry.info/termList/RDACarrierType/"),
("rdamt", "http://rdaregistry.info/termList/RDAMediaType/"),
("rdaft", "http://rdaregistry.info/termList/fileType/"),
("schema", "https://schema.org/"),
("gl", "http://fedora.quill.lan/rest/"),
("glo", ONTOLOGY_PREFIX),
]
.map(|(k, v)| (k.to_string(), v.to_string())),
)
});
pub struct OntologyBuilder {
@@ -82,18 +90,18 @@ impl OntologyBuilder {
.collect::<HashMap<String, String>>();
let mut indexed_by = HashMap::new();
for quad in store.quads_for_pattern(None, Some(gl::INDEXED_BY_FIELD), None, None)
.filter_map(Result::ok) {
for quad in store
.quads_for_pattern(None, Some(gl::INDEXED_BY_FIELD), None, None)
.filter_map(Result::ok)
{
if let NamedOrBlankNode::NamedNode(subject) = quad.subject
&& let Term::NamedNode(field) = quad.object {
&& let Term::NamedNode(field) = quad.object
{
indexed_by.insert(subject, field);
}
}
Ok(Ontology {
store,
prefixes,
})
Ok(Ontology { store, prefixes })
}
}
@@ -146,9 +154,7 @@ pub struct Ontology {
impl Ontology {
pub fn builder() -> OntologyBuilder {
OntologyBuilder {
path: None,
}
OntologyBuilder { path: None }
}
pub fn prefixes() -> &'static BTreeMap<String, String> {
@@ -166,9 +172,14 @@ impl Ontology {
self.store.clone()
}
pub fn query_for_indexable_triples(&self, language: &LanguageCondition, source: Option<Dataset>) -> impl Future<Output = error::Result<HashMap<NamedNode, IndexEntry>>> + 'static {
pub fn query_for_indexable_triples(
&self,
language: &LanguageCondition,
source: Option<Dataset>,
) -> impl Future<Output = error::Result<HashMap<NamedNode, IndexEntry>>> + 'static {
let language_filter = language.to_filter_expression("fieldValue");
let query = format!(r#"SELECT ?individual ?categoryId ?fieldName ?fieldValue
let query = format!(
r#"SELECT ?individual ?categoryId ?fieldName ?fieldValue
WHERE {{
?class a gl:SearchableClass ;
gl:categoryId ?categoryId ;
@@ -180,9 +191,11 @@ WHERE {{
?property ?fieldValue .
{language_filter}
}}"#);
}}"#
);
let mut sparql = SparqlEvaluator::new()
.with_prefix("gl", ONTOLOGY_PREFIX).unwrap()
.with_prefix("gl", ONTOLOGY_PREFIX)
.unwrap()
.parse_query(&query)
.expect("Unable to parse query");
sparql.dataset_mut().set_default_graph_as_union();
@@ -194,24 +207,40 @@ WHERE {{
sparql.on_queryable_dataset(source).execute()
} else {
sparql.on_store(&store).execute()
}.expect("Unable to execute indexing query");
}
.expect("Unable to execute indexing query");
let mut results = HashMap::new();
if let QueryResults::Solutions(solutions) = query_results {
let mut counter = 0usize;
for solution in solutions.filter_map(Result::ok) {
let individual = solution.get("individual").and_then(conversion::term_to_named_node);
let individual = solution
.get("individual")
.and_then(conversion::term_to_named_node);
let catalog_id = solution.get("categoryId").and_then(conversion::term_to_u64);
let field_name = solution.get("fieldName").and_then(conversion::term_as_str);
let field_value = solution.get("fieldValue").and_then(conversion::term_as_str);
if let (Some(individual), Some(catalog_id), Some(field_name), Some(field_value)) = (individual, catalog_id, field_name, field_value) {
results.entry(individual.to_owned())
if let (
Some(individual),
Some(catalog_id),
Some(field_name),
Some(field_value),
) = (individual, catalog_id, field_name, field_value)
{
results
.entry(individual.to_owned())
.and_modify(|entry: &mut IndexEntry| {
entry.fields.insert(field_name.to_owned(), field_value.to_owned());
}).or_insert(IndexEntry {
category_id: catalog_id,
fields: HashMap::from_iter([(field_name.to_owned(), field_value.to_owned())]),
});
entry
.fields
.insert(field_name.to_owned(), field_value.to_owned());
})
.or_insert(IndexEntry {
category_id: catalog_id,
fields: HashMap::from_iter([(
field_name.to_owned(),
field_value.to_owned(),
)]),
});
}
counter += 1;
}
@@ -220,27 +249,35 @@ WHERE {{
unreachable!()
}
results
}).map_err(error::Error::from)
})
.map_err(error::Error::from)
}
pub fn category_id(&self, class: &NamedNode) -> Option<u64> {
self.store.quads_for_pattern(Some(class.into()), Some(gl::CATEGORY_ID), None, None)
self.store
.quads_for_pattern(Some(class.into()), Some(gl::CATEGORY_ID), None, None)
.filter_map(Result::ok)
.map(conversion::quad_into_term)
.filter_map(conversion::term_into_u64)
.next()
}
pub fn fields_for_class(&self, class: &NamedNode, language: &LanguageCondition) -> error::Result<Vec<IndexField>> {
pub fn fields_for_class(
&self,
class: &NamedNode,
language: &LanguageCondition,
) -> error::Result<Vec<IndexField>> {
let language_filter = language.to_filter_expression("label");
let query = format!(r#"SELECT DISTINCT ?name ?label {{
let query = format!(
r#"SELECT DISTINCT ?name ?label {{
{class} gl:associatedProperty/gl:indexedByField ?field .
?field gl:fieldName ?name ;
gl:fieldLabel ?label .
{language_filter}
}}"#);
}}"#
);
let mut sparql = SparqlEvaluator::new()
.with_prefix("rdfs", "http://www.w3.org/2000/01/rdf-schema#")?
.with_prefix("gl", ONTOLOGY_PREFIX)?
@@ -253,7 +290,9 @@ WHERE {{
let name = solution.get("name").and_then(conversion::term_as_str);
let label = solution.get("label").and_then(conversion::term_as_str);
if let Some(name) = name && let Some(label) = label {
if let Some(name) = name
&& let Some(label) = label
{
results.push(IndexField {
name: name.to_string(),
label: label.to_string(),
@@ -265,7 +304,9 @@ WHERE {{
}
pub fn info(&self, iri: &NamedNode, language: &LanguageCondition) -> LabeledIri {
let label = self.store.quads_for_pattern(Some(iri.as_ref().into()), Some(rdfs::LABEL), None, None)
let label = self
.store
.quads_for_pattern(Some(iri.as_ref().into()), Some(rdfs::LABEL), None, None)
.filter_map(Result::ok)
.map(conversion::quad_into_term)
.filter(|term| language.primary_matches_term(term))
@@ -280,11 +321,16 @@ WHERE {{
}
pub fn subclasses_of(&self, class: &NamedNode) -> BTreeSet<NamedNode> {
self.store.quads_for_pattern(None, Some(rdfs::SUB_CLASS_OF), Some(class.into()), None)
self.store
.quads_for_pattern(None, Some(rdfs::SUB_CLASS_OF), Some(class.into()), None)
.filter_map(Result::ok)
.filter_map(|quad| if let NamedOrBlankNode::NamedNode(subject) = quad.subject {
Some(subject)
} else { None })
.filter_map(|quad| {
if let NamedOrBlankNode::NamedNode(subject) = quad.subject {
Some(subject)
} else {
None
}
})
.collect()
}
@@ -300,7 +346,8 @@ WHERE {{
.filter_map(|quad| match quad.subject {
NamedOrBlankNode::NamedNode(subject) => Some(subject),
_ => None,
}).collect()
})
.collect()
}
pub fn is_read_only<'a>(&self, triple: impl Into<TripleRef<'a>>) -> bool {
@@ -313,23 +360,35 @@ WHERE {{
};
self.store
.quads_for_pattern(Some(NamedOrBlankNodeRef::NamedNode(subject)), Some(gl::READ_ONLY), Some(true_term), None)
.quads_for_pattern(
Some(NamedOrBlankNodeRef::NamedNode(subject)),
Some(gl::READ_ONLY),
Some(true_term),
None,
)
.filter_map(Result::ok)
.count() >= 1
.count()
>= 1
}
pub fn exclude_read_only(&self) -> impl Fn(TripleRef<'_>) -> bool + 'static {
let true_term = TermRef::Literal(LiteralRef::new_typed_literal("true", xsd::BOOLEAN));
let read_only_graph = self.store
let read_only_graph = self
.store
.quads_for_pattern(None, Some(gl::READ_ONLY), Some(true_term), None)
.filter_map(Result::ok)
.map(Triple::from)
.collect::<Graph>();
move |triple| {
if triple.predicate == rdf::TYPE && let TermRef::NamedNode(class) = triple.object {
if triple.predicate == rdf::TYPE
&& let TermRef::NamedNode(class) = triple.object
{
read_only_graph.triples_for_subject(class).count() == 0
} else {
read_only_graph.triples_for_subject(triple.predicate).count() == 0
read_only_graph
.triples_for_subject(triple.predicate)
.count()
== 0
}
}
}
@@ -353,7 +412,12 @@ WHERE {{
if let Term::BlankNode(blank_node) = quad.object {
let iter = self
.store
.quads_for_pattern(Some(NamedOrBlankNodeRef::BlankNode(blank_node.as_ref())), None, None, None)
.quads_for_pattern(
Some(NamedOrBlankNodeRef::BlankNode(blank_node.as_ref())),
None,
None,
None,
)
.filter_map(Result::ok)
.map(Triple::from)
.map(move |mut triple| {
@@ -368,4 +432,4 @@ WHERE {{
None
}
}
}
}