Well Named Graphs

.github/workflows/test.yml

@@ -16,15 +16,15 @@ jobs:
         python-version: ["3.10", "3.11", "3.12"]
 
     steps:
-      - uses: actions/checkout@v4
+      - uses: actions/checkout@v6
 
       - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
+        uses: actions/setup-python@v6
         with:
           python-version: ${{ matrix.python-version }}
 
       - name: Install uv
-        uses: astral-sh/setup-uv@v2
+        uses: astral-sh/setup-uv@v7
 
       - name: Install dependencies
         run: uv sync

README.md

@@ -362,4 +362,6 @@ The `example_projects` folder contains contrived examples, but this has also bee
 1. check and raise error or at least warning if default_union is set in underlying Dataset of DatasetView
 1. document and/or fix serialisation: canon longTurtle is not great with the way it orders things, so we might need to call out to riot unfortunately.
 1. consider changing the distinction from interal/external to data/vocabulary (where vocab includes taxonomies or ontologies) - basically the ABox/TBox distinction where CBox is part of TBox.
-1. add support for ASK query
+1. add better output support for ASK query
+1. add option to remove project name from named graphs, for easier specification:
+   1. e.g. `<urn:pythinfer:inferences:owl>` which is easy to remember and specify on command-line.

example_projects/eg0-basic/expected/expected-0-merged.trig

@@ -0,0 +1,30 @@
+@prefix : <http://example.org/> .
+@prefix dcterms: <http://purl.org/dc/terms/> .
+@prefix foaf: <http://xmlns.com/foaf/0.1/> .
+@prefix owl: <http://www.w3.org/2002/07/owl#> .
+@prefix pythinfer: <http://pythinfer.local/> .
+
+@base <http://pythinfer.local/eg0-basic/> .
+
+<file/basic-data.ttl> {
+    :Bob a foaf:Person ;
+        foaf:knows :Alice ;
+        foaf:name "Bob Jones" .
+
+    :Alice a foaf:Person ;
+        foaf:age 30 ;
+        foaf:name "Alice Smith" .
+}
+
+<file/basic-model.ttl> {
+    foaf:knows a owl:SymmetricProperty .
+}
+
+<provenance> {
+    <file/basic-data.ttl> a pythinfer:SourceGraph ;
+        dcterms:source <file:///$PROJ_FOLDER/pythinfer/example_projects/eg0-basic/basic-data.ttl> ;
+        .
+    <file/basic-model.ttl> a pythinfer:SourceGraph ;
+        dcterms:source <file:///$PROJ_FOLDER/pythinfer/example_projects/eg0-basic/basic-model.ttl> ;
+        .
+}

example_projects/eg0-basic/derived/expected_inferred_wanted.trig → example_projects/eg0-basic/expected/expected-2-inferred-wanted.trig

@@ -1,6 +1,6 @@
 @prefix : <http://example.org/> .
 @prefix foaf: <http://xmlns.com/foaf/0.1/> .
 
-<inferences_owl> {
+<http://pythinfer.local/eg0-basic/inferences/owl> {
     :Alice foaf:knows :Bob .
 }

example_projects/eg1-ancestors/derived/expected_merged.trig → example_projects/eg1-ancestors/expected/expected-0-merged.trig

@@ -4,8 +4,11 @@
 @prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
 @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
 @prefix skos: <http://www.w3.org/2004/02/skos/core#> .
+@prefix pythinfer: <http://pythinfer.local/> .
 
-<ancestors-model.ttl> {
+@base <http://pythinfer.local/eg1-ancestors/> .
+
+<file/ancestors-model.ttl> {
     ex:childOf rdfs:label "child of" ;
         owl:inverseOf ex:parentOf .
 
@@ -33,7 +36,7 @@
         rdfs:label "Person" .
 }
 
-<ancestors-data.ttl> {
+<file/ancestors-data.ttl> {
     ex:Alice a ex:Person ;
         ex:parentOf ex:Bob,
             ex:Carol .
@@ -48,3 +51,9 @@
 
     ex:Eve a ex:Person .
 }
+
+<provenance> {
+    <file/ancestors-data.ttl> a pythinfer:SourceGraph .
+    <file/ancestors-model.ttl> a pythinfer:SourceGraph .
+    <file/../../vocabs-external/skos.ttl> a pythinfer:SourceGraph .
+}

example_projects/eg1-ancestors/derived/expected_inferred_wanted.trig → example_projects/eg1-ancestors/expected/expected-2-inferred-wanted.trig

@@ -1,6 +1,7 @@
 @prefix ex: <http://example.org/ancestor/> .
 
-<inferences_owl> {
+@base <http://pythinfer.local/eg1-ancestors/> .
+<inferences/owl> {
     ex:David ex:childOf ex:Bob ;
         ex:descendantOf ex:Alice,
             ex:Bob .

example_projects/eg2-projects/derived/expected_inferred_wanted.trig → example_projects/eg2-projects/expected/expected-2-inferred-wanted.trig

@@ -7,7 +7,9 @@
 @prefix ptp: <http://www.example.org/pythinfer/project/> .
 @prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
 
-<inferences_sparql> {
+@base <http://pythinfer.local/eg2-projects/> .
+
+<inferences/sparql> {
     # This is the most important SPARQL-based inference: the relationship between projA and projB
     eg:relationship-projA-projB a ptp:ProjectRelationship ;
         ptp:hasParticipant eg:projA,
@@ -47,7 +49,7 @@
 }
 
 
-<inferences_owl> {
+<inferences/owl> {
     eg:projA a dcat:Catalog,
             prov:Activity ;
         ptp:hasDataSource _:b0,

src/pythinfer/infer.py

@@ -9,6 +9,7 @@
 from owlrl import DeductiveClosure
 from owlrl.OWLRL import OWLRL_Semantics
 from rdflib import (
+    DCTERMS,
     OWL,
     RDF,
     RDFS,
@@ -30,11 +31,9 @@
     export_dataset,
     load_sparql_inference_queries,
 )
+from pythinfer.inout import PYTHINFER_NS
 from pythinfer.rdflibplus import DatasetView
 
-IRI_EXTERNAL_INFERENCES: URIRef = URIRef("inferences_external")  # type: ignore[bad-assignment]
-IRI_OWL_INFERENCES: URIRef = URIRef("inferences_owl")  # type: ignore[bad-assignment]
-IRI_SPARQL_INFERENCES: URIRef = URIRef("inferences_sparql")  # type: ignore[bad-assignment]
 
 MAX_REASONING_ROUNDS = 5
 SCRIPT_DIR = Path(__file__).parent
@@ -259,7 +258,7 @@ def filter_triples(
 
 
 def _generate_external_inferences(
-    ds: Dataset, external_graph_ids: list[IdentifiedNode]
+    ds: Dataset, external_graph_ids: list[IdentifiedNode], project: Project
 ) -> Graph:
     """Generate inferences from external vocabularies only (step 2).
 
@@ -269,6 +268,7 @@ def _generate_external_inferences(
     Args:
         ds: Dataset containing all graphs.
         external_graph_ids: List of graph identifiers that are external.
+        project: The project configuration.
 
     Returns:
         Graph containing external inferences.
@@ -286,13 +286,21 @@ def _generate_external_inferences(
     info("  Temporary dataset created with %d triples in default graph", len(temp_ds))
 
     # Create inferences graph in temp dataset (must share same store)
-    temp_inferences = temp_ds.graph(IRI_EXTERNAL_INFERENCES)
+    iri_external = project.inference_gid("external")
+    temp_inferences = temp_ds.graph(iri_external)
+    g_provenance = ds.graph(project.provenance_gid)
 
     apply_owlrl_inference(temp_ds, temp_inferences)
 
-    g_external_inferences = ds.graph(IRI_EXTERNAL_INFERENCES)
+    g_external_inferences = ds.graph(iri_external)
     for s, p, o in temp_inferences:
         g_external_inferences.add((s, p, o))
+
+    # Add provenance metadata for external inference graph
+    g_provenance.add((iri_external, RDF.type, PYTHINFER_NS["InferenceGraph"]))
+    g_provenance.add(
+        (iri_external, PYTHINFER_NS["inferenceEngine"], Literal("owlrl"))
+    )
     info("  External inferences generated: %d triples", len(g_external_inferences))
     return g_external_inferences
 
@@ -406,16 +414,38 @@ def run_inference_backend(
     sparql_queries = load_sparql_inference_queries(project.paths_sparql_inference or [])
 
     # Step 2: Generate external inferences (once - this is the "noise floor")
-    g_external_inferences = _generate_external_inferences(ds, external_graph_ids)
+    g_external_inferences = _generate_external_inferences(
+        ds, external_graph_ids, project
+    )
 
     # Steps 3-5: Iterate full inferences + heuristics until convergence
     info(
         "Steps 3-5: Iterating full inferences + heuristics (max %d iterations)...",
         MAX_REASONING_ROUNDS,
     )
 
-    g_inferences_owl = ds.graph(IRI_OWL_INFERENCES)
-    g_inferences_sparql = ds.graph(IRI_SPARQL_INFERENCES)
+    iri_owl = project.inference_gid("owl")
+    iri_sparql = project.inference_gid("sparql")
+    g_inferences_owl = ds.graph(iri_owl)
+    g_inferences_sparql = ds.graph(iri_sparql)
+    g_provenance = ds.graph(project.provenance_gid)
+
+    # Add provenance metadata for inference graphs
+    g_provenance.add((iri_owl, RDF.type, PYTHINFER_NS["InferenceGraph"]))
+    g_provenance.add(
+        (iri_owl, PYTHINFER_NS["inferenceEngine"], Literal(project.owl_backend))
+    )
+
+    g_provenance.add(
+        (iri_sparql, RDF.type, PYTHINFER_NS["InferenceGraph"])
+    )
+    g_provenance.add(
+        (
+            iri_sparql,
+            PYTHINFER_NS["inferenceEngine"],
+            Literal("SPARQL CONSTRUCT"),
+        )
+    )
     iteration = 0
     previous_triple_count = len(ds)  # Count triples in entire dataset
 
@@ -491,18 +521,19 @@ def run_inference_backend(
         len(g_inferences_owl) + len(g_inferences_sparql),
     )
 
+    iri_external = project.inference_gid("external")
     all_external_ids: list[IdentifiedNode] = [
         *external_graph_ids,
-        IRI_EXTERNAL_INFERENCES,
+        iri_external,
     ]
 
     output_file = output or project.path_output / f"{INFERRED_WANTED_FILESTEM}.trig"
     output_file.parent.mkdir(parents=True, exist_ok=True)
 
     output_ds = DatasetView(
         ds,
-        [IRI_OWL_INFERENCES, IRI_SPARQL_INFERENCES]
-        + ([IRI_EXTERNAL_INFERENCES] if export_external_inferences else []),
+        [iri_owl, iri_sparql]
+        + ([iri_external] if export_external_inferences else []),
     )
 
     export_dataset(