spacy-entity-relation-extractor-task-008: add integration tests for SpaCy components

Add integration test files for SpacyEntityRelationExtractor and
HierarchicalTextSplitter that exercise the real en_core_web_sm model.
All tests are auto-skipped (not failed) when the model is not installed.

Author: jexp

.plans/progress-spacy-entity-relation-extractor.txt

@@ -40,3 +40,7 @@ Files: ~src/neo4j_graphrag/experimental/components/__init__.py, ~src/neo4j_graph
 ## TASK-010 — 2026-06-24T17:00:00Z
 Example script for end-to-end spacy KG pipeline.
 Files: +examples/customize/build_graph/components/spacy_kg_pipeline.py
+
+## TASK-008 — 2026-06-24T17:00:00Z
+Integration tests for SpacyEntityRelationExtractor and HierarchicalTextSplitter (skip when en_core_web_sm absent).
+Files: +tests/unit/experimental/components/test_spacy_extractor_integration.py, +tests/unit/experimental/components/text_splitters/test_hierarchical_splitter_integration.py

.plans/tasks-spacy-entity-relation-extractor.yml

@@ -176,7 +176,7 @@ tasks:
       Use pytest fixtures that skip the test if `en_core_web_sm` is not installed.
       Verify that real SpaCy output produces a populated Neo4jGraph and correct
       TextChunks respectively.
-    status: in_progress
+    status: completed
     dependencies:
       - task-005
       - task-007
@@ -185,7 +185,7 @@ tasks:
       - "With model installed, SpacyEntityRelationExtractor.run() returns Neo4jGraph with at least one node and one relationship for a 2-sentence input"
       - "With model installed, HierarchicalTextSplitter.run() returns correct chunks for a multi-section document"
       - "Confidence values on relationships are floats in [0, 1]"
-    completed_at: null
+    completed_at: "2026-06-24T17:00:00Z"
 
   - id: task-009
     title: "Export new components from experimental __init__"

tests/unit/experimental/components/test_spacy_extractor_integration.py

@@ -0,0 +1,122 @@
+#  Copyright (c) "Neo4j"
+#  Neo4j Sweden AB [https://neo4j.com]
+#  #
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#  #
+#      https://www.apache.org/licenses/LICENSE-2.0
+#  #
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+"""Integration tests for SpacyEntityRelationExtractor using the real en_core_web_sm model.
+
+All tests in this module are automatically skipped when either spaCy or the
+``en_core_web_sm`` model is not installed.  No mocking is performed — these
+tests exercise the full NLP pipeline end-to-end.
+"""
+
+from __future__ import annotations
+
+import pytest
+
+spacy = pytest.importorskip("spacy")
+
+
+@pytest.fixture(scope="module")
+def nlp():  # type: ignore[return]
+    try:
+        return spacy.load("en_core_web_sm")
+    except OSError:
+        pytest.skip(
+            "en_core_web_sm not installed — run: python -m spacy download en_core_web_sm"
+        )
+
+
+@pytest.mark.asyncio
+async def test_run_returns_populated_graph(nlp) -> None:  # noqa: ANN001
+    """SpacyEntityRelationExtractor.run() returns a Neo4jGraph with nodes and relationships."""
+    from neo4j_graphrag.experimental.components.spacy_entity_relation_extractor import (
+        SpacyEntityRelationExtractor,
+    )
+    from neo4j_graphrag.experimental.components.types import TextChunk, TextChunks
+
+    extractor = SpacyEntityRelationExtractor(
+        model="en_core_web_sm",
+        use_linear_extractor=True,
+        create_lexical_graph=False,
+    )
+    chunks = TextChunks(
+        chunks=[
+            TextChunk(
+                text="Apple acquired Beats Electronics. Tim Cook announced the deal.",
+                index=0,
+            )
+        ]
+    )
+    result = await extractor.run(chunks=chunks)
+
+    assert len(result.nodes) >= 1, "Expected at least one entity node"
+    assert len(result.relationships) >= 1, "Expected at least one relationship"
+
+
+@pytest.mark.asyncio
+async def test_confidence_values_in_range(nlp) -> None:  # noqa: ANN001
+    """All relationship confidence values are floats in [0, 1]."""
+    from neo4j_graphrag.experimental.components.spacy_entity_relation_extractor import (
+        SpacyEntityRelationExtractor,
+    )
+    from neo4j_graphrag.experimental.components.types import TextChunk, TextChunks
+
+    extractor = SpacyEntityRelationExtractor(
+        model="en_core_web_sm",
+        use_linear_extractor=True,
+        create_lexical_graph=False,
+    )
+    chunks = TextChunks(
+        chunks=[
+            TextChunk(
+                text="Apple acquired Beats Electronics. Tim Cook announced the deal.",
+                index=0,
+            )
+        ]
+    )
+    result = await extractor.run(chunks=chunks)
+
+    for rel in result.relationships:
+        conf = rel.properties.get("confidence")
+        assert isinstance(conf, float), f"confidence should be float, got {type(conf)}"
+        assert 0.0 <= conf <= 1.0, f"confidence out of range [0,1]: {conf}"
+
+
+@pytest.mark.asyncio
+async def test_use_linear_extractor_false_no_related_to(nlp) -> None:  # noqa: ANN001
+    """use_linear_extractor=False produces no RELATED_TO relationships."""
+    from neo4j_graphrag.experimental.components.spacy_entity_relation_extractor import (
+        SpacyEntityRelationExtractor,
+    )
+    from neo4j_graphrag.experimental.components.types import TextChunk, TextChunks
+
+    extractor = SpacyEntityRelationExtractor(
+        model="en_core_web_sm",
+        use_linear_extractor=False,
+        create_lexical_graph=False,
+    )
+    chunks = TextChunks(
+        chunks=[
+            TextChunk(
+                text="Apple acquired Beats Electronics. Tim Cook announced the deal.",
+                index=0,
+            )
+        ]
+    )
+    result = await extractor.run(chunks=chunks)
+
+    related_to_rels = [r for r in result.relationships if r.type == "RELATED_TO"]
+    assert related_to_rels == [], (
+        f"Expected no RELATED_TO relationships when use_linear_extractor=False, "
+        f"but got: {related_to_rels}"
+    )

tests/unit/experimental/components/text_splitters/test_hierarchical_splitter_integration.py

@@ -0,0 +1,99 @@
+#  Copyright (c) "Neo4j"
+#  Neo4j Sweden AB [https://neo4j.com]
+#  #
+#  Licensed under the Apache License, Version 2.0 (the "License");
+#  you may not use this file except in compliance with the License.
+#  You may obtain a copy of the License at
+#  #
+#      https://www.apache.org/licenses/LICENSE-2.0
+#  #
+#  Unless required by applicable law or agreed to in writing, software
+#  distributed under the License is distributed on an "AS IS" BASIS,
+#  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+#  See the License for the specific language governing permissions and
+#  limitations under the License.
+"""Integration tests for HierarchicalTextSplitter using the real en_core_web_sm model.
+
+All tests in this module are automatically skipped when either spaCy or the
+``en_core_web_sm`` model is not installed.  No mocking is performed — these
+tests exercise the full NLP pipeline end-to-end.
+"""
+
+from __future__ import annotations
+
+import pytest
+
+spacy = pytest.importorskip("spacy")
+
+
+@pytest.fixture(scope="module")
+def nlp():  # type: ignore[return]
+    try:
+        return spacy.load("en_core_web_sm")
+    except OSError:
+        pytest.skip(
+            "en_core_web_sm not installed — run: python -m spacy download en_core_web_sm"
+        )
+
+
+@pytest.mark.asyncio
+async def test_markdown_split_real(nlp) -> None:  # noqa: ANN001
+    """HierarchicalTextSplitter with markdown strategy returns sequential chunks for a 3-section doc."""
+    from neo4j_graphrag.experimental.components.text_splitters.hierarchical_splitter import (
+        HierarchicalTextSplitter,
+    )
+
+    text = (
+        "# Introduction\n"
+        "This section introduces the topic and provides background information.\n\n"
+        "# Methods\n"
+        "This section describes the experimental methods used in the study.\n\n"
+        "# Conclusion\n"
+        "This section summarises the findings and suggests future work.\n"
+    )
+
+    splitter = HierarchicalTextSplitter(
+        header_strategy="markdown",
+        max_chunk_size=200,
+        chunk_overlap=20,
+        drop_verbless_sentences=False,
+    )
+    result = await splitter.run(text)
+
+    assert len(result.chunks) >= 2, (
+        f"Expected at least 2 chunks for a 3-section markdown doc, got {len(result.chunks)}"
+    )
+    # Indices must be sequential starting from 0.
+    for i, chunk in enumerate(result.chunks):
+        assert chunk.index == i, f"chunk {i} has non-sequential index {chunk.index}"
+
+
+@pytest.mark.asyncio
+async def test_drop_verbless_sentences_real(nlp) -> None:  # noqa: ANN001
+    """drop_verbless_sentences=True drops verbless fragments using the real SpaCy model."""
+    from neo4j_graphrag.experimental.components.text_splitters.hierarchical_splitter import (
+        HierarchicalTextSplitter,
+    )
+
+    # "Overview" is a single-word verbless fragment.
+    # The second sentence contains a real verb ("covers").
+    text = "Overview\n\nThis section covers the main concepts of the system in detail."
+
+    splitter = HierarchicalTextSplitter(
+        header_strategy="blank_line",
+        max_chunk_size=500,
+        chunk_overlap=0,
+        drop_verbless_sentences=True,
+        model="en_core_web_sm",
+    )
+    result = await splitter.run(text)
+
+    # At least one chunk must survive.
+    assert len(result.chunks) >= 1, "Expected at least one chunk after filtering"
+
+    # The verbless word "Overview" should not appear as a standalone sentence
+    # in any chunk, while the main sentence content should be present.
+    all_text = " ".join(chunk.text for chunk in result.chunks)
+    assert "covers" in all_text, (
+        "Expected the verbal sentence to survive verbless-sentence filtering"
+    )