Test Coverage¶

Summary of automated test coverage, reporting conventions, and quality gates for Neutryx Core.

Current Status (November 2025)¶

Total Tests: 500+ comprehensive tests across all modules Coverage: 85%+ line coverage, 90%+ for core modules Test Types: Unit, integration, regression, performance, security

Test Distribution by Module¶

Module	Tests	Focus Areas
Interest Rate Derivatives	87	IRS, OIS, swaptions, CMS, exotic IR
Regulatory Compliance	120	FRTB SA/IMA, DRC/RRAO, SA-CCR, SIMM, IFRS
Calibration	60	Joint calibration, Bayesian averaging, regularization
Risk Management	57	VaR, stress testing, limits, pre-trade controls
Portfolio Analytics	80	Backtesting, factor analysis, optimization
XVA & CCR	35	CVA, DVA, FVA, MVA, KVA, exposure profiles
Market Data	25	Vendor integration, validation, storage
Core Pricing	36	Monte Carlo, PDE solvers, Greeks
Total	500+	Comprehensive coverage across platform

Test Categories¶

1. Unit Tests (300+)¶

Test individual functions and components in isolation:

Pricing Functions: Black-Scholes, Heston, SABR analytical pricing
Model Implementations: Parameter validation, constraint checking
Greeks Calculation: Delta, gamma, vega accuracy tests
Calibration: Optimizer convergence, constraint satisfaction
Market Data: Parsing, validation, transformation logic
Utilities: Day count conventions, holiday calendars, date adjustments

Example:

def test_black_scholes_call_price():
    """Test BS call option pricing against known values."""
    price = bs_price(S=100, K=100, T=1.0, r=0.05, q=0.02, sigma=0.2, option_type="call")
    expected = 10.450583572185565
    assert jnp.isclose(price, expected, rtol=1e-6)

2. Integration Tests (120+)¶

Test end-to-end workflows and system interactions:

Pricing Workflows: Market data → calibration → pricing → Greeks
Risk Workflows: Position updates → VaR calculation → limit checking
XVA Workflows: Exposure simulation → CVA calculation → reporting
API Endpoints: REST/gRPC request handling and responses
Database Operations: Read/write operations with PostgreSQL, MongoDB, TimescaleDB
Regulatory Reports: FRTB, SA-CCR, SIMM calculation pipelines

Example:

def test_heston_calibration_workflow():
    """Test complete Heston calibration workflow."""
    market_data = load_market_data()
    model = HestonModel()
    calibrated = calibrate_model(model, market_data)
    assert calibrated.converged
    assert all(calibrated.params.values() > 0)  # Positive parameters

3. Regression Tests (50+)¶

Ensure pricing and risk calculations remain stable across versions:

Pricing Stability: Compare current vs reference prices
Greeks Stability: Verify sensitivities within tolerance
Calibration Stability: Parameter convergence to known values
Performance Regression: Monitor execution times
Numerical Precision: Track floating-point accuracy

Test Data: - Reference prices from QuantLib, Bloomberg - Historical market data snapshots - Known analytical solutions

4. Performance Tests (20+)¶

Benchmark critical paths and validate scalability:

Monte Carlo: Paths/second throughput
PDE Solvers: Grid points/second throughput
Calibration: Convergence speed
Batch Pricing: Scaling with batch size
GPU Acceleration: CPU vs GPU speedup
Memory Usage: Peak memory consumption

Targets: - 10-100x speedup with JIT compilation - Linear scaling for batch pricing - 100x+ speedup on GPU for Monte Carlo

5. Security Tests (10+)¶

Validate security controls and vulnerabilities:

Input Validation: SQL injection, XSS prevention
Authentication: OAuth/JWT token validation
Authorization: RBAC permission checks
Audit Logging: Event capture and integrity
Dependency Scanning: Known vulnerabilities (bandit, pip-audit)

Coverage Targets¶

Coverage by Module¶

Module	Target	Current	Status
Core Pricing	95%	96%	✅ Exceeds
Models	90%	92%	✅ Exceeds
Products	90%	88%	🔄 Close
Risk Management	85%	87%	✅ Exceeds
XVA & Valuations	85%	86%	✅ Exceeds
Calibration	90%	93%	✅ Exceeds
Market Data	80%	82%	✅ Exceeds
Infrastructure	75%	78%	✅ Exceeds
Overall	85%	87%	✅ Exceeds

Critical Path Coverage¶

Functions on critical paths require 95%+ coverage: - Option pricing engines - Greeks calculation - VaR methodologies - XVA calculations - Regulatory capital calculations

Running Tests¶

All Tests¶

pytest -v                              # Verbose output
pytest -q                              # Quiet output (default)
pytest --cov=neutryx --cov-report=html # Coverage report

By Category¶

# Unit tests only
pytest src/neutryx/tests/unit/ -v

# Integration tests
pytest src/neutryx/tests/integration/ -v

# Regression tests
pytest -m regression

# Performance tests
pytest -m performance

# Specific module
pytest src/neutryx/tests/products/ -v
pytest src/neutryx/tests/risk/ -v
pytest src/neutryx/tests/valuations/regulatory/ -v

By Feature¶

# Interest rate derivatives
pytest src/neutryx/tests/products/test_interest_rate.py -v
pytest src/neutryx/tests/products/test_swaptions.py -v

# Regulatory compliance
pytest src/neutryx/tests/valuations/regulatory/test_frtb.py -v
pytest src/neutryx/tests/valuations/regulatory/test_sa_ccr.py -v
pytest src/neutryx/tests/valuations/regulatory/test_simm.py -v

# Portfolio analytics
pytest src/neutryx/tests/portfolio/test_optimization.py -v
pytest src/neutryx/tests/portfolio/test_backtesting.py -v
pytest src/neutryx/tests/portfolio/test_factor_analysis.py -v

Test Quality Standards¶

1. Test Structure¶

Follow the Arrange-Act-Assert pattern:

def test_european_call_option():
    # Arrange: Setup inputs and expected results
    S, K, T, r, q, sigma = 100.0, 100.0, 1.0, 0.05, 0.02, 0.2
    expected_price = 10.450583572185565

    # Act: Execute the function under test
    price = european_option_price(S, K, T, r, q, sigma, "call")

    # Assert: Verify results
    assert jnp.isclose(price, expected_price, rtol=1e-6)

2. Test Isolation¶

Each test should be independent
Use fixtures for shared setup
Clean up resources in teardown
No test order dependencies

3. Test Documentation¶

Docstring explaining what is tested
Comments for complex test logic
Clear assertion messages
Link to requirements/issues

4. Test Data¶

Use realistic market data
Include edge cases and boundary conditions
Test both valid and invalid inputs
Document data sources

Continuous Integration¶

Tests run automatically on:

Pull Requests: All tests must pass
Main Branch: Full test suite + performance benchmarks
Nightly Builds: Extended tests + security scans
Release Tags: Complete validation suite

Quality Gates¶

Gate	Threshold	Current	Status
All tests pass	100%	100%	✅
Line coverage	>85%	87%	✅
Branch coverage	>80%	82%	✅
Performance regression	<5%	2%	✅
Security vulnerabilities	0 critical	0	✅
Code quality (ruff)	0 errors	0	✅

Test Maintenance¶

Adding New Tests¶

Identify the feature to test
Write test cases covering:
Happy path (normal operation)
Edge cases (boundary conditions)
Error cases (invalid inputs)
Ensure tests are deterministic
Add to appropriate test module
Verify CI passes

Updating Existing Tests¶

Update when requirements change
Preserve regression test data
Document breaking changes
Maintain backward compatibility when possible

Removing Tests¶

Deprecate before removal
Document reason for removal
Check for dependent tests
Update coverage reports

Test Performance¶

Execution Times (Approximate)¶

Test Suite	Tests	Time	Per Test
Unit Tests	300+	~60s	~200ms
Integration Tests	120+	~180s	~1.5s
Regression Tests	50+	~120s	~2.4s
Performance Tests	20+	~300s	~15s
Total	500+	~10min	~1.2s

Optimization Strategies¶

Use pytest-xdist for parallel execution
Cache expensive fixtures
Mock external dependencies
Use fast approximations for smoke tests
Profile slow tests with pytest-profiling

Test Data Management¶

Fixtures¶

@pytest.fixture
def market_data():
    """Sample market data for testing."""
    return {
        "spot": 100.0,
        "risk_free_rate": 0.05,
        "dividend_yield": 0.02,
        "volatility": 0.20
    }

Reference Data¶

Stored in tests/data/ directory
Versioned with git
Documented in README files
Updated when models change

Future Enhancements¶

[ ] Property-based testing with Hypothesis
[ ] Mutation testing for test quality
[ ] Contract testing for APIs
[ ] Chaos engineering tests
[ ] Load testing for distributed systems

Last Updated: November 2025 Test Count: 500+ Coverage: 87% Status: Production-ready ✅