Hypergraph is more powerful. Transform unstructured text into structured knowledge with LLMs. Graphs, hypergraphs, and spatio-temporal extractions — with one command.
test: cover relation_members extractors for graph and hypergraph
Frankie-Xu8f3446959da6
diff --git a/tests/template_engine/test_parser_identifiers.py b/tests/template_engine/test_parser_identifiers.pynew file mode 100644index 0000000..b7960be--- /dev/null+++ b/tests/template_engine/test_parser_identifiers.py@@ -0,0 +1,130 @@+"""Unit tests for template parsers - identifiers."""++import pytest+from pydantic import BaseModel++from hyperextract.utils.template_engine.parsers import parse_identifiers+from hyperextract.utils.template_engine.parsers.schemas import (+ GraphIdentifiersSchema,+)+++class Entity(BaseModel):+ name: str+++class Relation(BaseModel):+ source: str+ target: str+ type: str+++class CustomRelation(BaseModel):+ """Relation whose endpoint fields are not named source/target."""++ subject: str+ object: str+ type: str+++class NumericRelation(BaseModel):+ source: int+ target: int+ type: str+++class HyperRelation(BaseModel):+ name: str+ participants: list[str]+++class NestedHyperRelation(BaseModel):+ name: str+ group_a: list[str]+ group_b: list[str]+++def _graph_members_extractor(relation_members):+ """Build the members extractor via the public parse_identifiers API."""+ identifiers = GraphIdentifiersSchema(+ entity_id="name",+ relation_id="{type}",+ relation_members=relation_members,+ )+ _, _, members_extractor = parse_identifiers(identifiers, "graph")+ return members_extractor+++class TestGraphRelationMembers:+ """Dict-based relation_members for binary graph types."""++ def test_default_source_target_mapping(self):+ """Standard {source: source, target: target} mapping works."""+ extractor = _graph_members_extractor({"source": "source", "target": "target"})+ edge = Relation(source="A", target="B", type="knows")++ assert extractor(edge) == ("A", "B")++ def test_direction_is_preserved(self):+ """Directed edges must keep (source, target) order, not sorted order."""+ extractor = _graph_members_extractor({"source": "source", "target": "target"})+ edge = Relation(source="B", target="A", type="manages")++ assert extractor(edge) == ("B", "A")++ def test_custom_field_mapping(self):+ """Dict values map to custom edge field names, per documented contract."""+ extractor = _graph_members_extractor({"source": "subject", "target": "object"})+ edge = CustomRelation(subject="Alice", object="Bob", type="mentors")++ assert extractor(edge) == ("Alice", "Bob")++ def test_custom_field_mapping_preserves_direction(self):+ """Custom-mapped endpoints also keep source -> target order."""+ extractor = _graph_members_extractor({"source": "subject", "target": "object"})+ edge = CustomRelation(subject="Zoe", object="Adam", type="mentors")++ assert extractor(edge) == ("Zoe", "Adam")++ def test_missing_mapped_field_raises(self):+ """Mapping to a field absent from the edge schema raises AttributeError."""+ extractor = _graph_members_extractor({"source": "src", "target": "dst"})+ edge = Relation(source="A", target="B", type="knows")++ with pytest.raises(AttributeError, match="src"):+ extractor(edge)++ def test_non_string_values_coerced_to_str(self):+ """Endpoint values are coerced to strings for key comparison."""+ extractor = _graph_members_extractor({"source": "source", "target": "target"})+ edge = NumericRelation(source=2, target=1, type="links")++ assert extractor(edge) == ("2", "1")+++class TestHypergraphRelationMembers:+ """String and list relation_members for hypergraph types."""++ def test_string_members_returns_sorted_participants(self):+ """Unordered hyperedge participants are sorted for a stable key."""+ identifiers = GraphIdentifiersSchema(+ entity_id="name",+ relation_id="{name}",+ relation_members="participants",+ )+ _, _, extractor = parse_identifiers(identifiers, "hypergraph")+ edge = HyperRelation(name="meeting", participants=["Carol", "Alice", "Bob"])++ assert extractor(edge) == ("Alice", "Bob", "Carol")++ def test_list_members_returns_sorted_groups(self):+ """Nested hyperedge groups are each sorted for a stable key."""+ identifiers = GraphIdentifiersSchema(+ entity_id="name",+ relation_id="{name}",+ relation_members=["group_a", "group_b"],+ )+ _, _, extractor = parse_identifiers(identifiers, "hypergraph")+ edge = NestedHyperRelation(name="match", group_a=["Y", "X"], group_b=["B", "A"])++ assert extractor(edge) == (("X", "Y"), ("A", "B"))