claude-os/tests/TESTING_GUIDE.md

Claude OS Testing Guide

🎯 Overview

This guide provides comprehensive information about testing Claude OS, a production-grade RAG system.

---

📋 Table of Contents

1. Quick Start
2. Test Structure
3. Running Tests
4. Writing Tests
5. CI/CD Integration
6. Troubleshooting

---

🚀 Quick Start

Prerequisites

1. Python 3.11+ with pytest
2. Ollama running locally (for integration tests)

Run All Tests

./run_tests.sh

That's it! The script will:

• ✅ Check dependencies
• ✅ Create test database if needed
• ✅ Run all tests
• ✅ Show coverage report

---

📁 Test Structure

tests/
├── __init__.py                 # Test package
├── conftest.py                 # Shared fixtures
├── pytest.ini                  # Pytest configuration
├── README.md                   # Quick reference
├── TEST_COVERAGE.md            # Coverage details
├── TESTING_GUIDE.md            # This file
│
├── test_pg_vector.py           # Vector operations
├── test_rag_engine.py          # RAG engine
├── test_embeddings.py          # Embedding generation
├── test_document_processing.py # Document ingestion
├── test_api.py                 # API endpoints
└── test_pg_manager.py          # Database operations

Test Categories

Category	Marker	Description
Unit	@pytest.mark.unit	Fast, isolated tests
Integration	@pytest.mark.integration	Tests with external services
Vector	@pytest.mark.vector	Vector operations
RAG	@pytest.mark.rag	RAG engine tests
API	@pytest.mark.api	API endpoint tests
Embeddings	@pytest.mark.embeddings	Embedding generation
Slow	@pytest.mark.slow	Tests >5 seconds

---

🏃 Running Tests

Basic Commands

# All tests
./run_tests.sh

# With coverage report
./run_tests.sh --coverage

# Verbose output
./run_tests.sh --verbose

By Category

# Unit tests only (fast)
./run_tests.sh --unit

# Integration tests
./run_tests.sh --integration

# Vector operations
./run_tests.sh --vector

# RAG engine
./run_tests.sh --rag

# API endpoints
./run_tests.sh --api

Advanced Usage

# Specific test file
pytest tests/test_pg_vector.py

# Specific test function
pytest tests/test_pg_vector.py::TestPGVectorOperations::test_vector_storage

# Stop on first failure
pytest -x

# Run last failed tests
pytest --lf

# Run tests matching pattern
pytest -k "vector"

---

✍️ Writing Tests

Test Template

"""
Tests for [component name].
"""

import pytest


@pytest.mark.unit  # or integration, vector, etc.
class TestMyComponent:
    """Test [component] functionality."""
    
    def test_basic_functionality(self, sample_kb, clean_db):
        """Test that [feature] works correctly."""
        # Arrange
        input_data = "test input"
        
        # Act
        result = my_function(input_data)
        
        # Assert
        assert result == expected_output
        assert result.status == "success"

Using Fixtures

def test_with_fixtures(self, sample_kb, sample_documents, clean_db):
    """Test using multiple fixtures."""
    # sample_kb: Pre-created knowledge base
    # sample_documents: 5 documents with embeddings
    # clean_db: Clean database connection
    
    result = query_kb(sample_kb["name"], "test query")
    assert len(result["sources"]) > 0

Mocking External Services

from unittest.mock import patch, MagicMock

@patch('app.core.rag_engine.Ollama')
def test_with_mock_llm(self, mock_llm):
    """Test with mocked LLM."""
    # Setup mock
    mock_instance = MagicMock()
    mock_instance.chat.return_value.message.content = "Test response"
    mock_llm.return_value = mock_instance
    
    # Test
    result = my_function()
    assert result == "Test response"

Best Practices

1. One assertion per test (when possible)
2. Descriptive test names (test_vector_similarity_with_empty_kb)
3. Use fixtures for common setup
4. Mock external services in unit tests
5. Test edge cases (empty input, invalid data, etc.)
6. Clean up after tests (use fixtures with cleanup)

---

🔄 CI/CD Integration

GitHub Actions

Tests run automatically on:

• ✅ Push to main or develop
• ✅ Pull requests
• ✅ Manual workflow dispatch

See .github/workflows/tests.yml for configuration.

Local Pre-commit Hook

# .git/hooks/pre-commit
#!/bin/bash
./run_tests.sh --unit
if [ $? -ne 0 ]; then
    echo "Tests failed! Commit aborted."
    exit 1
fi

Make it executable:

chmod +x .git/hooks/pre-commit

---

🐛 Troubleshooting

Ollama Connection Errors

# Check Ollama is running
curl http://localhost:11434/api/tags

# Start Ollama
ollama serve

Import Errors

# Ensure you're in project root
cd /path/to/claude-os

# Install dependencies
pip install -r requirements.txt

Fixture Not Found

Make sure conftest.py is in the tests/ directory and contains the fixture definition.

Test Database Pollution

# Drop and recreate test database
dropdb codeforge_test
createdb codeforge_test
psql -d codeforge_test -c "CREATE EXTENSION vector"
psql -d codeforge_test -f app/core/schema.sql

---

📊 Coverage Reports

Generate HTML Report

./run_tests.sh --coverage
open htmlcov/index.html  # macOS
xdg-open htmlcov/index.html  # Linux

Coverage Goals

• Overall: 85%
• Critical paths: 90%+
• New code: 80%+

---

🎓 Learning Resources

• pytest documentation
• pytest fixtures
• unittest.mock
• Test-Driven Development

---

📝 Checklist for New Features

• Write tests first (TDD)
• Use appropriate markers
• Add fixtures if reusable
• Test happy path
• Test edge cases
• Test error handling
• Run ./run_tests.sh
• Verify >80% coverage
• Update documentation

---

Remember: Good tests are the foundation of reliable software! 🚀