Testing Infrastructure

Relevant source files

The following files were used as context for generating this wiki page:

• tests/test_integration.py
• tests/test_mcp_tools.py
• tests/test_models.py
• tests/test_session_manager.py
• memory-bank/memories/67944906-43bc-4c85-9424-f2261850bc97.md
• memory-bank/memories/1eba30a1-6283-44a2-a372-82e9fe92460d.md

The testing infrastructure provides a comprehensive validation suite for the Sequential Thinking MCP v2 server. It covers unit testing for domain models and business logic, adapter-level testing for tool handlers, and full integration tests for end-to-end reasoning workflows. The suite ensures that session persistence, memory management, and structured reasoning remain robust during modernization efforts.

Sources: memory-bank/memories/67944906-43bc-4c85-9424-f2261850bc97.md:7-9, tests/test_integration.py:10-12

Testing Layers

The infrastructure divides testing into three distinct layers to ensure complete coverage of the 5-layer architecture.

Layer	Scope	Primary File
Model Validation	Validates data integrity, default values, and UUID generation for dataclasses.	tests/test_models.py
Logic & Storage	Tests UnifiedSessionManager for file operations and state transitions.	tests/test_session_manager.py
Tool Interface	Verifies MCPToolsHandler response formats and error catching.	tests/test_mcp_tools.py
Integration	Simulates complete user workflows and server restarts.	tests/test_integration.py

Sources: tests/test_models.py:5, tests/test_session_manager.py:11, tests/test_mcp_tools.py:9, tests/test_integration.py:10

Environment Management

The test suite uses pytest fixtures to manage temporary files. This prevents test execution from polluting the actual memory-bank/ directory.

Temporary Memory Bank Setup

The temp_memory_bank fixture creates a unique directory for each test class using tempfile.mkdtemp(). It recursively deletes the directory after the test finishes using shutil.rmtree().

@pytest.fixture
def temp_memory_bank(self):
    """Create a temporary memory bank for testing."""
    temp_dir = tempfile.mkdtemp()
    yield temp_dir
    shutil.rmtree(temp_dir)

Sources: tests/test_integration.py:13-18, tests/test_mcp_tools.py:12-17

Component Relationship Flow

This diagram illustrates how the testing infrastructure initializes the system components within a temporary environment.

The diagram shows the setup and teardown flow used in TestIntegration and TestMCPToolsHandler.
Sources: tests/test_integration.py:20-25, tests/test_mcp_tools.py:19-27

Domain Model Testing

The infrastructure validates the core data structures defined in models.py. Tests ensure that every object receives a unique ID and that defaults (like confidence levels) are applied correctly.

Key Model Validations

• UUID Uniqueness: Validates that calling constructors multiple times generates distinct identifiers.
• Default Values: Confirms that Thought and Memory objects start with a 0.8 confidence level.
• Enum Integrity: Verifies that SessionType correctly distinguishes between "general" and "coding" sessions.

Sources: tests/test_models.py:10-15, tests/test_models.py:157-175, tests/test_models.py:178-185

Session and Storage Validation

Testing for UnifiedSessionManager focuses on the persistence layer and state management.

Persistence Verification

Tests verify that the manager creates a structured filesystem upon initialization:

1. sessions/ for markdown snapshots.
2. memories/ for standalone memory artifacts.
3. patterns/ for reserved code patterns.
4. index.md for the session registry.

Sources: tests/test_session_manager.py:25-33

Workflow Logic

The suite tests complex state transitions, such as branching and merging.

The diagram shows the sequence of operations for testing branch lifecycle.
Sources: tests/test_session_manager.py:91-133

Tool Handler Testing

Tests in tests/test_mcp_tools.py ensure the MCPToolsHandler provides a stable interface for the MCP protocol.

Success and Error Patterns

Tool tests verify that successful operations return dictionaries with relevant IDs (e.g., thought_id, memory_id), while failures are caught and returned as error strings rather than raising unhandled exceptions.

Tool Method	Success Assertion	Error Condition
start_session	Returns session_id	Invalid parameters
add_thought	Returns thought_id	No active session
query_memories	Returns memories list	Non-existent tags
export_session	Returns filename	Invalid format or no session

Sources: tests/test_mcp_tools.py:29-57, tests/test_mcp_tools.py:186-218

Integration Testing

Integration tests simulate real-world reasoning tasks to verify that the 12 tools work together.

Persistence and Recovery

A critical test case simulates a server crash and restart:

1. Initialize a session and add data.
2. Create a new UnifiedSessionManager instance pointing to the same temporary path.
3. Call load_session(original_id).
4. Verify that all thoughts, memories, and architecture decisions are re-hydrated from the markdown files.

Sources: tests/test_integration.py:218-257

End-to-End Coding Workflow

The test_complete_workflow_coding_session validates a multi-step process:

1. Starting a "coding" type session.
2. Exploring package dependencies.
3. Recording architecture decisions with rationale and consequences.
4. Exporting the final session to a markdown file for human review.

Sources: tests/test_integration.py:27-142

The infrastructure achieves >85% coverage across unit, integration, and error handling scenarios, transforming a previously untested project into a production-ready system.

Sources: memory-bank/memories/67944906-43bc-4c85-9424-f2261850bc97.md:7-9, memory-bank/memories/1eba30a1-6283-44a2-a372-82e9fe92460d.md:16