conftest.py

#!/usr/bin/env python3
"""
Pytest configuration and fixtures for Case File Analysis tests.
"""

import pytest
import sys
import os
import warnings

# Add src to path for imports
sys.path.insert(0, os.path.join(os.path.dirname(__file__), '..', 'src'))

# Suppress warnings for cleaner test output
warnings.filterwarnings("ignore", category=UserWarning)
warnings.filterwarnings("ignore", category=DeprecationWarning)


@pytest.fixture(scope="session")
def sample_case_text():
    """Fixture providing sample case file text for testing."""
    return """
    Case No: 2024-CR-001
    Status: Active
    Date of Incident: March 15, 2024
    Time of Incident: 10:30 PM
    Location: 123 Main Street, Downtown
    Crime: Burglary
    
    John Smith reported that Tom Wilson broke into his house. 
    The suspect stole $2,500 in cash and jewelry worth $15,000.
    Mary Johnson witnessed the incident.
    """


@pytest.fixture(scope="session")
def complex_case_text():
    """Fixture providing complex case file text for integration testing."""
    return """
    Case No: 2024-BG-001
    Status: Under Investigation
    Date of Incident: March 15, 2024
    Time of Incident: 2:30 AM
    Location: 123 Main Street, Downtown
    Crime Type: Burglary
    
    Incident Report:
    On the above date and time, victim John Smith (DOB: 01/15/1980) reported a burglary
    at his residence located at 123 Main Street, Downtown. The suspect, identified as
    Tom Wilson (DOB: 05/22/1985), allegedly broke into the residence through the back door
    using a crowbar.
    
    Stolen Items:
    - Gold necklace valued at $2,500.00
    - Cash totaling $500.00  
    - iPhone 13 Pro valued at $800.00
    - Wedding ring worth $1,200.00
    
    Witnesses:
    - Mary Johnson (neighbor) observed Wilson near the property at approximately 2:15 AM
    - Security camera footage shows suspect vehicle (License: ABC-123)
    
    Evidence Collected:
    - Fingerprints on crowbar
    - Footprints in backyard
    - Security footage from 2:00-3:00 AM
    
    Investigating Officer: Detective Sarah Connor
    Report Filed: March 15, 2024 at 8:30 AM
    
    Follow-up Actions:
    - Interview additional neighbors
    - Process forensic evidence
    - Issue warrant for suspect arrest
    """


@pytest.fixture(scope="function")
def temp_directory():
    """Fixture providing temporary directory for test files."""
    import tempfile
    import shutil
    
    temp_dir = tempfile.mkdtemp()
    yield temp_dir
    shutil.rmtree(temp_dir)


@pytest.fixture(scope="session")
def parser_minimal():
    """Fixture providing LangGraphParser with minimal configuration for fast testing."""
    try:
        from casefile_parser.LangGraphParser import LangGraphParser
        parser = LangGraphParser(
            max_enrichment_steps=3,
            model_config={
                "model_id": "claude-3-haiku-20240307",
                "temperature": 0.0,
                "max_tokens": 4096
            }
        )
        return parser
    except Exception as e:
        pytest.skip(f"Could not initialize LangGraphParser: {e}")


@pytest.fixture(scope="session", autouse=True)
def setup_test_environment():
    """Automatically set up test environment."""
    # Set environment variables for testing
    os.environ['TESTING'] = 'true'
    
    # Reduce logging verbosity during tests
    import logging
    logging.getLogger().setLevel(logging.ERROR)
    
    yield
    
    # Cleanup after all tests
    if 'TESTING' in os.environ:
        del os.environ['TESTING']


@pytest.fixture
def mock_nlp():
    """Fixture providing mocked spaCy nlp object."""
    from unittest.mock import MagicMock
    
    mock_nlp = MagicMock()
    mock_doc = MagicMock()
    mock_nlp.return_value = mock_doc
    mock_doc.ents = []
    mock_doc.sents = []
    
    return mock_nlp


# Pytest markers for test categorization
pytest_plugins = []

def pytest_configure(config):
    """Configure pytest with custom markers."""
    config.addinivalue_line(
        "markers", "slow: mark test as slow running (integration tests)"
    )
    config.addinivalue_line(
        "markers", "integration: mark test as integration test"
    )
    config.addinivalue_line(
        "markers", "unit: mark test as unit test"
    )
    config.addinivalue_line(
        "markers", "parser: mark test as parser-related"
    )
    config.addinivalue_line(
        "markers", "entities: mark test as entity extraction related"
    )
    config.addinivalue_line(
        "markers", "relations: mark test as relation extraction related"
    )
    config.addinivalue_line(
        "markers", "utilities: mark test as utility function related"
    )


def pytest_collection_modifyitems(config, items):
    """Modify test collection to add markers automatically."""
    for item in items:
        # Add markers based on test file names
        if "test_integration" in item.nodeid:
            item.add_marker(pytest.mark.integration)
            item.add_marker(pytest.mark.slow)
        elif "test_casefile_parser" in item.nodeid:
            item.add_marker(pytest.mark.parser)
            item.add_marker(pytest.mark.unit)
        elif "test_entity_extractor" in item.nodeid:
            item.add_marker(pytest.mark.entities)
            item.add_marker(pytest.mark.unit)
        elif "test_relational_inference" in item.nodeid:
            item.add_marker(pytest.mark.relations)
            item.add_marker(pytest.mark.unit)
        elif "test_utilities" in item.nodeid:
            item.add_marker(pytest.mark.utilities)
            item.add_marker(pytest.mark.unit)
        
        # Add slow marker for tests that might take longer
        if any(keyword in item.nodeid.lower() for keyword in ['performance', 'large', 'batch']):
            item.add_marker(pytest.mark.slow)


@pytest.fixture
def skip_if_no_spacy():
    """Skip tests if spaCy model is not available (not autouse to avoid slow imports)."""
    try:
        import spacy
        spacy.load("en_core_web_sm")
    except (ImportError, OSError):
        pytest.skip("spaCy en_core_web_sm model not available")


# Pytest hooks for better test reporting
def pytest_runtest_setup(item):
    """Setup for each test run."""
    # Clear any cached imports or state
    pass


def pytest_runtest_teardown(item, nextitem):
    """Teardown after each test run."""
    # Clean up any test artifacts
    pass


# Custom assertion helpers
class CaseFileAssertions:
    """Custom assertions for case file testing."""
    
    @staticmethod
    def assert_valid_entity_result(result):
        """Assert that entity extraction result is valid."""
        assert isinstance(result, dict), "Result should be a dictionary"
        for category, spans in result.items():
            assert isinstance(category, str), "Category should be string"
            assert isinstance(spans, list), "Spans should be a list"
            for span in spans:
                assert isinstance(span, tuple), "Span should be tuple"
                assert len(span) == 2, "Span should have start and end"
                assert isinstance(span[0], int), "Start should be integer"
                assert isinstance(span[1], int), "End should be integer"
                assert span[0] <= span[1], "Start should be <= end"
    
    @staticmethod
    def assert_valid_relations_result(relations):
        """Assert that relation extraction result is valid."""
        assert isinstance(relations, list), "Relations should be a list"
        for relation in relations:
            assert isinstance(relation, tuple), "Relation should be tuple"
            assert len(relation) == 3, "Relation should have head, tail, info"
            head, tail, info = relation
            assert isinstance(head, str), "Head should be string"
            assert isinstance(tail, str), "Tail should be string"
            assert isinstance(info, dict), "Info should be dictionary"
            assert 'relation' in info, "Info should contain relation type"


# Make assertions available to all test files
@pytest.fixture
def assertions():
    """Fixture providing custom assertions."""
    return CaseFileAssertions()