"""
Test cases for RBTChunker.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
"""
import pytest
from rbt_mcp_server.chunking.rbt_chunker import RBTChunker
from rbt_mcp_server.chunking.models import ChunkMetadata
class TestRBTChunker:
"""
Test suite for RBTChunker following TDD approach.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
"""
@pytest.fixture
def chunker(self):
"""Create RBTChunker instance."""
return RBTChunker()
@pytest.fixture
def sample_rbt_document(self):
"""
Sample RBT document with complete structure.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
"""
return """---
id: TASK-002-RBTChunker
group_id: knowledge-smith
type: Task
title: 實作 RBTChunker
blueprint: BP-graphiti-chunk-mcp
requirement: REQ-graphiti-chunk-mcp
---
<!-- info-section -->
> status: Pending
> update_date: 2025-10-08
> dependencies: []
<!-- id: sec-root -->
# Task: 實作 RBTChunker - 根據 RBT 結構分塊
<!-- id: sec-goal-dependencies -->
## 1. 任務目標與前置
<!-- id: blk-goal-content, type: paragraph -->
這是第一個 section 的內容。
<!-- id: sec-goal -->
### 1.1 目標
<!-- id: blk-goal-list, type: list -->
- 實作 RBTChunker
- 解析文件結構
- 生成穩定的 chunk_id
<!-- id: sec-dependencies -->
### 1.2 前置任務
<!-- id: blk-dependencies, type: paragraph -->
無前置任務。
<!-- id: sec-implementation -->
## 2. 實作指引
<!-- id: blk-implementation-intro, type: paragraph -->
這是實作指引的內容。
<!-- id: blk-implementation-steps, type: list -->
**實作步驟**:
- 步驟 1
- 步驟 2
- 步驟 3
<!-- id: blk-code-example, type: code, language: python -->
```python
def example():
return "code example"
```
<!-- id: blk-table-example, type: table -->
| Header 1 | Header 2 |
| --- | --- |
| Cell 1 | Cell 2 |
| Cell 3 | Cell 4 |
"""
@pytest.fixture
def nested_sections_document(self):
"""
RBT document with nested sections.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
"""
return """---
id: TEST-001
group_id: knowledge-smith
type: Test
title: Test Document
---
<!-- info-section -->
> status: Draft
> update_date: 2025-10-08
<!-- id: sec-root -->
# Test: Nested Sections
<!-- id: sec-level1 -->
## 1. Section Level 1
<!-- id: blk-level1-content, type: paragraph -->
Content at level 1.
<!-- id: sec-level2 -->
### 1.1 Section Level 2
<!-- id: blk-level2-content, type: paragraph -->
Content at level 2.
<!-- id: sec-level3 -->
#### 1.1.1 Section Level 3
<!-- id: blk-level3-content, type: paragraph -->
Content at level 3.
<!-- id: sec-another -->
## 2. Another Section
<!-- id: blk-another-content, type: paragraph -->
More content.
"""
@pytest.fixture
def mixed_summary_document(self):
"""
RBT document with mixed summary presence.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
"""
return """---
id: TEST-002
group_id: knowledge-smith
type: Test
title: Mixed Summary Document
---
<!-- info-section -->
> status: Draft
> update_date: 2025-10-08
<!-- id: sec-root -->
# Test: Mixed Summaries
<!-- id: sec-with-summary -->
## 1. Section With Summary
<!-- summary: This is a summary for this section. -->
<!-- id: blk-with-summary-content, type: paragraph -->
Content goes here.
<!-- id: sec-without-summary -->
## 2. Section Without Summary
<!-- id: blk-without-summary-content, type: paragraph -->
Content without summary.
"""
def test_chunk_complete_document(self, chunker, sample_rbt_document):
"""
Test Case 1: 解析完整 RBT 文件.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
Given: 一份包含 metadata, info, 多個 sections 的 RBT 文件
When: 調用 chunk()
Then: 返回對應數量的 ChunkMetadata，每個 chunk_id 穩定且正確
"""
chunks = chunker.chunk(
document_content=sample_rbt_document,
project_id="knowledge-smith",
feature_id="graphiti-chunk-mcp",
doc_type="TASK"
)
# Verify we got chunks
assert len(chunks) > 0, "Should generate at least one chunk"
# Verify all chunks have required fields
for chunk in chunks:
assert isinstance(chunk, ChunkMetadata)
assert chunk.metadata is not None
assert chunk.metadata["chunk_id"] is not None
assert chunk.metadata["parent_document_id"] is not None
assert chunk.metadata["project_id"] == "knowledge-smith"
assert chunk.metadata["feature_id"] == "graphiti-chunk-mcp"
assert chunk.metadata["doc_type"] == "TASK"
assert chunk.metadata["section_id"] is not None
assert chunk.content is not None
# Verify parent_document_id format
expected_parent_id = "knowledge-smith+graphiti-chunk-mcp+TASK"
for chunk in chunks:
assert chunk.metadata["parent_document_id"] == expected_parent_id
# Verify chunk_id format
for chunk in chunks:
expected_chunk_id = f"knowledge-smith+graphiti-chunk-mcp+{chunk.metadata['section_id']}"
assert chunk.metadata["chunk_id"] == expected_chunk_id
def test_chunk_nested_sections(self, chunker, nested_sections_document):
"""
Test Case 2: 處理巢式 sections.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
Given: RBT 文件包含巢式 sections
When: 調用 chunk()
Then: 所有 sections（包含巢式）都被正確分塊
"""
chunks = chunker.chunk(
document_content=nested_sections_document,
project_id="test-project",
feature_id="test-feature",
doc_type="TEST"
)
# Should have chunks for all sections including nested ones
# We expect: sec-level1, sec-level2, sec-level3, sec-another
assert len(chunks) >= 4, f"Expected at least 4 chunks, got {len(chunks)}"
# Verify depth metadata
section_ids = [chunk.metadata["section_id"] for chunk in chunks]
assert "sec-level1" in section_ids
assert "sec-level2" in section_ids
assert "sec-level3" in section_ids
assert "sec-another" in section_ids
def test_section_summary_handling(self, chunker, mixed_summary_document):
"""
Test Case 3: section_summary 處理.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
Given: 部分 sections 有 summary，部分沒有
When: 調用 chunk()
Then: 有 summary 的 chunk 包含 section_summary，沒有的為 None
"""
chunks = chunker.chunk(
document_content=mixed_summary_document,
project_id="test-project",
feature_id="test-feature",
doc_type="TEST"
)
# Find chunks by section_id
chunk_map = {chunk.metadata["section_id"]: chunk for chunk in chunks}
# Section with summary should have section_summary
if "sec-with-summary" in chunk_map:
chunk_with = chunk_map["sec-with-summary"]
# It might have a summary or be None depending on parsing
# The key is that it's handled consistently
assert "section_summary" in chunk_with.metadata
# Section without summary should have None
if "sec-without-summary" in chunk_map:
chunk_without = chunk_map["sec-without-summary"]
assert "section_summary" in chunk_without.metadata
# Can be None or empty string, but should exist
def test_chunk_id_stability(self, chunker, sample_rbt_document):
"""
Test Case 4: chunk_id 穩定性.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
Given: 同一份 RBT 文件多次分塊
When: 多次調用 chunk()
Then: 相同 section 的 chunk_id 保持不變
"""
# Chunk the document twice
chunks1 = chunker.chunk(
document_content=sample_rbt_document,
project_id="knowledge-smith",
feature_id="graphiti-chunk-mcp",
doc_type="TASK"
)
chunks2 = chunker.chunk(
document_content=sample_rbt_document,
project_id="knowledge-smith",
feature_id="graphiti-chunk-mcp",
doc_type="TASK"
)
# Verify same number of chunks
assert len(chunks1) == len(chunks2), "Should generate same number of chunks"
# Verify chunk_ids are identical
chunk_ids1 = [chunk.metadata["chunk_id"] for chunk in chunks1]
chunk_ids2 = [chunk.metadata["chunk_id"] for chunk in chunks2]
assert chunk_ids1 == chunk_ids2, "Chunk IDs should be stable across multiple runs"
# Verify section_ids are also stable
section_ids1 = [chunk.metadata["section_id"] for chunk in chunks1]
section_ids2 = [chunk.metadata["section_id"] for chunk in chunks2]
assert section_ids1 == section_ids2, "Section IDs should be stable across multiple runs"
def test_chunk_content_generation(self, chunker, sample_rbt_document):
"""
Test chunk content includes section title and blocks.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
"""
chunks = chunker.chunk(
document_content=sample_rbt_document,
project_id="knowledge-smith",
feature_id="graphiti-chunk-mcp",
doc_type="TASK"
)
# All chunks should have non-empty content
for chunk in chunks:
chunk_id = chunk.metadata["chunk_id"]
assert chunk.content, f"Chunk {chunk_id} has empty content"
assert len(chunk.content) > 0, f"Chunk {chunk_id} has zero-length content"
# Content should contain section information
for chunk in chunks:
# Content should typically start with section title (##)
# Or contain meaningful text from blocks
chunk_id = chunk.metadata["chunk_id"]
assert chunk.content.strip(), f"Chunk {chunk_id} has only whitespace"
def test_chunk_metadata_fields(self, chunker, sample_rbt_document):
"""
Test chunk metadata contains expected fields.
@REQ: REQ-graphiti-chunk-mcp
@BP: BP-graphiti-chunk-mcp
@TASK: TASK-002-RBTChunker
"""
chunks = chunker.chunk(
document_content=sample_rbt_document,
project_id="knowledge-smith",
feature_id="graphiti-chunk-mcp",
doc_type="TASK"
)
for chunk in chunks:
# Verify metadata structure
assert isinstance(chunk.metadata, dict)
assert "chunk_id" in chunk.metadata
assert "parent_document_id" in chunk.metadata
assert "project_id" in chunk.metadata
assert "feature_id" in chunk.metadata
assert "doc_type" in chunk.metadata
assert "section_id" in chunk.metadata
assert "section_title" in chunk.metadata
assert "section_summary" in chunk.metadata
# Verify types
assert isinstance(chunk.metadata["chunk_id"], str)
assert isinstance(chunk.metadata["parent_document_id"], str)
assert isinstance(chunk.metadata["project_id"], str)
assert isinstance(chunk.metadata["doc_type"], str)
# Info chunk has special "info" field
if chunk.metadata["section_id"] == "sec-info":
assert "info" in chunk.metadata
assert isinstance(chunk.metadata["info"], dict)