fix(indicium): take the asb: base from indicium instead of hardcoding it

Author: lfnothias

pyproject.toml

@@ -134,7 +134,7 @@ graph = [
     "pyoxigraph>=0.4",
 ]
 
-# The indicium stack — indicium, indicium-adapters,
+# The indicium stack — indicium-standard, indicium-adapters,
 # indicium-adapters-metabolomics — backs validate_claims(), build_claim_graph()
 # and export_astra(). Those packages are private and absent from PyPI, so they
 # are deliberately NOT declared here, in any extra, and NOT wired up through
@@ -146,7 +146,9 @@ graph = [
 # `uv-resolve` CI job fail if they come back.
 #
 # The PyPI name `indicium` belongs to an unrelated project (a key-value store,
-# latest 0.1.0a3), so a bare `indicium` requirement installs the wrong package.
+# latest 0.1.0a3) that also ships a top-level `indicium` module, so a bare
+# `indicium` requirement installs the wrong package and clobbers the import.
+# The standard publishes as `indicium-standard`; the import name is unchanged.
 #
 # Maintainers holding the private sibling checkouts install them into the
 # synced venv, without editing this file:

src/perspicacite/indicium_layer/queries.py

@@ -10,9 +10,17 @@
 
 from typing import Any
 
+from indicium import ASB_BASE
+
 # ---------- Namespaces (single source of truth across the package) ----------
 
-ASB_NS = "https://asb.holobiomics.org/ns/asb#"
+# Read from indicium rather than restated, so the graphs this package builds are
+# always targeted by the SHACL shapes that validate them. A hardcoded copy went
+# stale when indicium 2.0.0 moved the base: the shapes then matched nothing and
+# every claim graph reported conforms=True without a single constraint checked.
+# Importing indicium at module scope is safe here — this package is the indicium
+# adapter layer, and its builder already requires it.
+ASB_NS = ASB_BASE
 CITO_NS = "http://purl.org/spar/cito/"
 PROV_NS = "http://www.w3.org/ns/prov#"
 FABIO_NS = "http://purl.org/spar/fabio/"

src/perspicacite/indicium_layer/store.py

@@ -25,14 +25,16 @@
 from rdflib import Dataset, URIRef
 from rdflib import Literal as RdflibLiteral
 
+from perspicacite.indicium_layer.queries import ASB_NS
+
 # rdflib ≥ 6.0 deprecates ConjunctiveGraph in favour of Dataset.
 # Dataset is a drop-in for our usage (add/query named graphs).
 ConjunctiveGraph = Dataset  # alias for backward compat in type hints below
 
 _RDF_SUBJECT = URIRef("http://www.w3.org/1999/02/22-rdf-syntax-ns#subject")
 _RDF_PREDICATE = URIRef("http://www.w3.org/1999/02/22-rdf-syntax-ns#predicate")
 _RDF_OBJECT = URIRef("http://www.w3.org/1999/02/22-rdf-syntax-ns#object")
-_ASB_CONFIDENCE = URIRef("https://asb.holobiomics.org/ns/asb#confidence")
+_ASB_CONFIDENCE = URIRef(f"{ASB_NS}confidence")
 _PROV_WAS_GENERATED_BY = URIRef("http://www.w3.org/ns/prov#wasGeneratedBy")
 _XSD_DECIMAL = "http://www.w3.org/2001/XMLSchema#decimal"
 

src/perspicacite/pipeline/claims.py

@@ -140,18 +140,21 @@ def _coerce_claim(c: dict, qualifiers: frozenset[str] = _QUALIFIERS) -> dict | N
     return claim
 
 
-_ASB = "https://asb.holobiomics.org/ns/asb#"
-
-
 def claims_to_graph(claims: list[dict]):
     """Serialize claim dicts to an rdflib Graph using the asb: vocabulary that
     indicium's SHACL targets (a asb:Claim with asb:context/subject/...)."""
     import rdflib
 
+    # Imported lazily beside rdflib: this module is core pipeline and must stay
+    # importable without the optional indicium stack. Taking the base from
+    # indicium keeps the graph in the namespace its SHACL shapes target; a
+    # hardcoded copy goes stale on a base move and then validates vacuously.
+    from indicium import ASB_BASE, INDICIUM_BASE
+
     g = rdflib.Graph()
-    asb = rdflib.Namespace(_ASB)
+    asb = rdflib.Namespace(ASB_BASE)
     SSSOM = rdflib.Namespace("https://w3id.org/sssom/")
-    INDICIUM = rdflib.Namespace("https://w3id.org/indicium/")
+    INDICIUM = rdflib.Namespace(INDICIUM_BASE)
     DCT = rdflib.Namespace("http://purl.org/dc/terms/")
     for i, c in enumerate(claims):
         cid = c.get("id") or f"pos:{i}"

src/perspicacite/rag/modes/reasoning/provenance.py

@@ -27,6 +27,7 @@
 if TYPE_CHECKING:
     from collections.abc import AsyncIterator
 
+from perspicacite.indicium_layer.queries import ASB_NS
 from perspicacite.logging import get_logger
 from perspicacite.models.rag import RAGRequest, SourceReference, StreamEvent
 from perspicacite.rag.modes.base import BaseRAGMode
@@ -58,7 +59,7 @@
 # Internal helpers
 # ---------------------------------------------------------------------------
 
-_CLAIM_NS = "https://asb.holobiomics.org/ns/asb#"
+_CLAIM_NS = ASB_NS
 _RDF_TYPE = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type"
 
 

src/perspicacite/rag/modes/reasoning/typed_contradiction.py

@@ -33,6 +33,7 @@
 if TYPE_CHECKING:
     from collections.abc import AsyncIterator
 
+from perspicacite.indicium_layer.queries import ASB_NS
 from perspicacite.logging import get_logger
 from perspicacite.models.rag import RAGRequest, SourceReference, StreamEvent
 from perspicacite.rag.modes.base import BaseRAGMode
@@ -78,7 +79,7 @@
 # Internal helpers — shared with provenance pattern
 # ---------------------------------------------------------------------------
 
-_CLAIM_NS = "https://asb.holobiomics.org/ns/asb#"
+_CLAIM_NS = ASB_NS
 _RDF_TYPE = "http://www.w3.org/1999/02/22-rdf-syntax-ns#type"
 
 

tests/unit/test_anchor_builder.py

@@ -4,6 +4,7 @@
 import pytest
 
 from perspicacite.indicium_layer.builder import build_claim_graph
+from perspicacite.indicium_layer.queries import ASB_NS
 from perspicacite.indicium_layer.store import ClaimGraphStore
 
 
@@ -57,13 +58,13 @@ async def test_builder_binds_quote_to_content_matched_passage(tmp_path, monkeypa
         )
         # The verified quote_exact must be present on some Evidence node.
         rows = store.select(
-            'SELECT ?q WHERE { ?e <https://asb.holobiomics.org/ns/asb#quoteExact> ?q }'
+            f'SELECT ?q WHERE {{ ?e <{ASB_NS}quoteExact> ?q }}'
         )
         quotes = {r["q"] for r in rows}
         assert "compound A inhibits enzyme B" in quotes
         # anchorStatus verified is recorded.
         status_rows = store.select(
-            'SELECT ?s WHERE { ?e <https://asb.holobiomics.org/ns/asb#anchorStatus> ?s }'
+            f'SELECT ?s WHERE {{ ?e <{ASB_NS}anchorStatus> ?s }}'
         )
         assert "verified" in {r["s"] for r in status_rows}
     finally:
@@ -109,12 +110,12 @@ async def test_builder_does_not_launder_unverified_quote(tmp_path, monkeypatch):
         )
         # No quoteExact may be laundered onto any Evidence node.
         quote_rows = store.select(
-            'SELECT ?q WHERE { ?e <https://asb.holobiomics.org/ns/asb#quoteExact> ?q }'
+            f'SELECT ?q WHERE {{ ?e <{ASB_NS}quoteExact> ?q }}'
         )
         assert quote_rows == []
         # The claim is still kept (fail-open) and tagged unverified.
         status_rows = store.select(
-            'SELECT ?s WHERE { ?e <https://asb.holobiomics.org/ns/asb#anchorStatus> ?s }'
+            f'SELECT ?s WHERE {{ ?e <{ASB_NS}anchorStatus> ?s }}'
         )
         assert "unverified" in {r["s"] for r in status_rows}
     finally:

tests/unit/test_claim_graph_store.py

@@ -1,9 +1,10 @@
 """Unit tests for indicium_layer.store.ClaimGraphStore (rdflib backend)."""
 
+from perspicacite.indicium_layer.queries import ASB_NS
 from perspicacite.indicium_layer.store import ClaimGraphStore
 
 PROV_WAS_GEN_BY = "http://www.w3.org/ns/prov#wasGeneratedBy"
-ASB_CONFIDENCE = "https://asb.holobiomics.org/ns/asb#confidence"
+ASB_CONFIDENCE = f"{ASB_NS}confidence"
 RDF_SUBJECT = "http://www.w3.org/1999/02/22-rdf-syntax-ns#subject"
 CITO_SUPPORTS = "http://purl.org/spar/cito/supports"
 
@@ -32,11 +33,11 @@ def test_literal_with_datatype():
     xsd_decimal = "http://www.w3.org/2001/XMLSchema#decimal"
     store.add(
         "kb://kbA/x",
-        "https://asb.holobiomics.org/ns/asb#confidence",
+        ASB_CONFIDENCE,
         ("literal", "0.82", xsd_decimal),
     )
     rows = store.select(
-        "SELECT ?v WHERE { <kb://kbA/x> <https://asb.holobiomics.org/ns/asb#confidence> ?v }"
+        f"SELECT ?v WHERE {{ <kb://kbA/x> <{ASB_CONFIDENCE}> ?v }}"
     )
     assert rows[0]["v"] == "0.82"
     store.close()

tests/unit/test_claims_extraction.py

@@ -309,8 +309,10 @@ def enrich_claim(self, c): return c
 def test_validate_claims_enrichment_and_shacl_combined():
     """validate_claims() with an adapter that has both enrich_claim and shacl_shapes
     must call shacl_shapes() once (wiring integration check)."""
-    import rdflib
     from unittest.mock import MagicMock, patch
+
+    import rdflib
+
     from perspicacite.pipeline.claims import validate_claims
 
     mock_graph = rdflib.Graph()
@@ -346,6 +348,7 @@ async def test_extract_claims_mcp_tool_passes_adapter_not_manual_enrich():
     """
     import json
     from unittest.mock import AsyncMock, MagicMock, patch
+
     from perspicacite.pipeline.claims import extract_claims
 
     adapter = _MockAdapter()
@@ -383,8 +386,9 @@ async def test_extract_claims_mcp_tool_passes_adapter_not_manual_enrich():
 def test_claims_to_graph_serializes_ontology_terms():
     """ontology_terms in a claim dict must appear as asb:{slot}_ontology_term triples."""
     import rdflib
+    from indicium import ASB_BASE as _ASB
+
     from perspicacite.pipeline.claims import claims_to_graph
-    _ASB = "https://asb.holobiomics.org/ns/asb#"
     asb = rdflib.Namespace(_ASB)
     claim = {
         "context": "in vivo", "subject": "glucose", "qualifier": "quantifies",
@@ -403,8 +407,9 @@ def test_claims_to_graph_serializes_ontology_terms():
 def test_claims_to_graph_no_ontology_terms_no_error():
     """Claims without ontology_terms must not raise and produce no extra triples."""
     import rdflib
+    from indicium import ASB_BASE as _ASB
+
     from perspicacite.pipeline.claims import claims_to_graph
-    _ASB = "https://asb.holobiomics.org/ns/asb#"
     asb = rdflib.Namespace(_ASB)
     claim = {
         "context": "in vivo", "subject": "glucose", "qualifier": "quantifies",
@@ -419,8 +424,9 @@ def test_claims_to_graph_no_ontology_terms_no_error():
 def test_claims_to_graph_skips_none_ontology_terms():
     """None values in ontology_terms must be skipped — not serialized as literal 'None'."""
     import rdflib
+    from indicium import ASB_BASE as _ASB
+
     from perspicacite.pipeline.claims import claims_to_graph
-    _ASB = "https://asb.holobiomics.org/ns/asb#"
     asb = rdflib.Namespace(_ASB)
     claim = {
         "context": "in vivo", "subject": "glucose", "qualifier": "quantifies",
@@ -533,10 +539,11 @@ def test_compose_adapters_all_unknown_gives_none():
 @pytest.mark.unit
 def test_claims_to_graph_emits_anchor_status_and_quote_exact():
     import rdflib
+    from indicium import ASB_BASE
 
     from perspicacite.pipeline.claims import claims_to_graph
 
-    asb = rdflib.Namespace("https://asb.holobiomics.org/ns/asb#")
+    asb = rdflib.Namespace(ASB_BASE)
     claims = [{
         "id": "c1", "context": "in vitro", "subject": "A",
         "qualifier": "inhibits", "relation": "inhibits", "object": "B",
@@ -551,10 +558,11 @@ def test_claims_to_graph_emits_anchor_status_and_quote_exact():
 @pytest.mark.unit
 def test_claims_to_graph_unverified_emits_status_but_not_quote():
     import rdflib
+    from indicium import ASB_BASE
 
     from perspicacite.pipeline.claims import claims_to_graph
 
-    asb = rdflib.Namespace("https://asb.holobiomics.org/ns/asb#")
+    asb = rdflib.Namespace(ASB_BASE)
     claims = [{
         "id": "c2", "context": "in vitro", "subject": "A",
         "qualifier": "inhibits", "relation": "inhibits", "object": "B",

tests/unit/test_claims_validation.py

@@ -1,8 +1,10 @@
 import pytest
 import rdflib
+
 from perspicacite.pipeline.claims import claims_to_graph, validate_claims
 
-ASB = rdflib.Namespace("https://asb.holobiomics.org/ns/asb#")
+indicium = pytest.importorskip("indicium")
+ASB = rdflib.Namespace(indicium.ASB_BASE)
 
 
 @pytest.mark.unit
@@ -58,12 +60,13 @@ def test_claim_missing_slots_is_rejected():
 
 @pytest.mark.unit
 def test_claims_to_graph_emits_sssom_mapping():
-    import rdflib
     import indicium
+    import rdflib
+
     from perspicacite.pipeline.claims import claims_to_graph
 
     SSSOM = rdflib.Namespace("https://w3id.org/sssom/")
-    asb = rdflib.Namespace("https://asb.holobiomics.org/ns/asb#")
+    asb = ASB
 
     pytest.importorskip("indicium")