bakery-ia/services/procurement/tests/test_supplier_performance_predictor.py

"""
Tests for Supplier Performance Predictor
"""

import pytest
import pandas as pd
import numpy as np
from datetime import datetime, timedelta
from app.ml.supplier_performance_predictor import SupplierPerformancePredictor


@pytest.fixture
def sample_order_history_good_supplier():
    """Generate sample order history for a reliable supplier."""
    dates = pd.date_range(start='2024-01-01', end='2024-12-31', freq='W')

    orders = []
    for i, date in enumerate(dates):
        expected_delivery = date + timedelta(days=3)

        # Good supplier: 95% on-time, occasional 1-day delay
        if np.random.random() < 0.95:
            actual_delivery = expected_delivery
        else:
            actual_delivery = expected_delivery + timedelta(days=1)

        # Good quality: 98% no issues
        quality_issues = np.random.random() > 0.98
        quality_score = np.random.uniform(90, 100) if not quality_issues else np.random.uniform(70, 85)

        # Good quantity accuracy: 99% accurate
        quantity_accuracy = np.random.uniform(0.98, 1.02)

        orders.append({
            'order_id': f'order-{i}',
            'order_date': date,
            'expected_delivery_date': expected_delivery,
            'actual_delivery_date': actual_delivery,
            'order_quantity': 100,
            'received_quantity': int(100 * quantity_accuracy),
            'quality_issues': quality_issues,
            'quality_score': quality_score,
            'order_value': 500.0
        })

    return pd.DataFrame(orders)


@pytest.fixture
def sample_order_history_poor_supplier():
    """Generate sample order history for an unreliable supplier."""
    dates = pd.date_range(start='2024-01-01', end='2024-12-31', freq='W')

    orders = []
    for i, date in enumerate(dates):
        expected_delivery = date + timedelta(days=3)

        # Poor supplier: 60% on-time, frequent delays of 2-5 days
        if np.random.random() < 0.60:
            actual_delivery = expected_delivery
        else:
            actual_delivery = expected_delivery + timedelta(days=np.random.randint(2, 6))

        # Poor quality: 20% issues
        quality_issues = np.random.random() > 0.80
        quality_score = np.random.uniform(85, 100) if not quality_issues else np.random.uniform(50, 75)

        # Poor quantity accuracy: frequent short deliveries
        if np.random.random() < 0.25:
            quantity_accuracy = np.random.uniform(0.75, 0.95)  # Short delivery
        else:
            quantity_accuracy = np.random.uniform(0.95, 1.05)

        orders.append({
            'order_id': f'order-{i}',
            'order_date': date,
            'expected_delivery_date': expected_delivery,
            'actual_delivery_date': actual_delivery,
            'order_quantity': 100,
            'received_quantity': int(100 * quantity_accuracy),
            'quality_issues': quality_issues,
            'quality_score': quality_score,
            'order_value': 500.0
        })

    return pd.DataFrame(orders)


@pytest.mark.asyncio
async def test_analyze_good_supplier(sample_order_history_good_supplier):
    """Test analysis of a reliable supplier."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='good-supplier',
        order_history=sample_order_history_good_supplier,
        min_orders=10
    )

    # Check structure
    assert 'tenant_id' in results
    assert 'supplier_id' in results
    assert 'reliability_score' in results
    assert 'metrics' in results
    assert 'predictions' in results
    assert 'risk_assessment' in results
    assert 'insights' in results

    # Check metrics calculated
    metrics = results['metrics']
    assert metrics['total_orders'] == len(sample_order_history_good_supplier)
    assert 'on_time_rate' in metrics
    assert 'quality_issue_rate' in metrics
    assert 'avg_quantity_accuracy' in metrics

    # Good supplier should have high reliability score
    reliability_score = results['reliability_score']
    assert reliability_score >= 85, f"Expected high reliability, got {reliability_score}"

    # Risk should be low
    risk_assessment = results['risk_assessment']
    assert risk_assessment['risk_level'] in ['low', 'medium']


@pytest.mark.asyncio
async def test_analyze_poor_supplier(sample_order_history_poor_supplier):
    """Test analysis of an unreliable supplier."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='poor-supplier',
        order_history=sample_order_history_poor_supplier,
        min_orders=10
    )

    # Poor supplier should have low reliability score
    reliability_score = results['reliability_score']
    assert reliability_score < 75, f"Expected low reliability, got {reliability_score}"

    # Risk should be high or critical
    risk_assessment = results['risk_assessment']
    assert risk_assessment['risk_level'] in ['medium', 'high', 'critical']

    # Should have risk factors
    assert len(risk_assessment['risk_factors']) > 0

    # Should generate insights
    insights = results['insights']
    assert len(insights) > 0

    # Should have at least one alert or prediction
    alert_insights = [i for i in insights if i['type'] in ['alert', 'prediction']]
    assert len(alert_insights) > 0


@pytest.mark.asyncio
async def test_performance_metrics_calculation(sample_order_history_good_supplier):
    """Test detailed metrics calculation."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='test-supplier',
        order_history=sample_order_history_good_supplier
    )

    metrics = results['metrics']

    # Check all key metrics present
    required_metrics = [
        'total_orders',
        'on_time_orders',
        'delayed_orders',
        'on_time_rate',
        'avg_delivery_delay_days',
        'avg_quantity_accuracy',
        'short_deliveries',
        'short_delivery_rate',
        'quality_issues',
        'quality_issue_rate',
        'avg_quality_score',
        'delivery_consistency',
        'quantity_consistency'
    ]

    for metric in required_metrics:
        assert metric in metrics, f"Missing metric: {metric}"

    # Check metrics are reasonable
    assert 0 <= metrics['on_time_rate'] <= 100
    assert 0 <= metrics['avg_quantity_accuracy'] <= 200  # Allow up to 200% over-delivery
    assert 0 <= metrics['quality_issue_rate'] <= 100
    assert 0 <= metrics['avg_quality_score'] <= 100


@pytest.mark.asyncio
async def test_reliability_score_calculation():
    """Test reliability score calculation with known inputs."""
    predictor = SupplierPerformancePredictor()

    # Perfect metrics
    perfect_metrics = {
        'on_time_rate': 100.0,
        'avg_quantity_accuracy': 100.0,
        'avg_quality_score': 100.0,
        'delivery_consistency': 100.0,
        'quantity_consistency': 100.0,
        'quality_issue_rate': 0.0,
        'short_delivery_rate': 0.0
    }

    perfect_score = predictor._calculate_reliability_score(perfect_metrics)
    assert perfect_score >= 95, f"Expected perfect score ~100, got {perfect_score}"

    # Poor metrics
    poor_metrics = {
        'on_time_rate': 50.0,
        'avg_quantity_accuracy': 85.0,
        'avg_quality_score': 70.0,
        'delivery_consistency': 50.0,
        'quantity_consistency': 60.0,
        'quality_issue_rate': 20.0,  # Should apply penalty
        'short_delivery_rate': 25.0  # Should apply penalty
    }

    poor_score = predictor._calculate_reliability_score(poor_metrics)
    assert poor_score < 70, f"Expected poor score <70, got {poor_score}"


@pytest.mark.asyncio
async def test_delay_probability_prediction(sample_order_history_poor_supplier):
    """Test delay probability prediction."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='test-supplier',
        order_history=sample_order_history_poor_supplier
    )

    predictions = results['predictions']

    # Should have delay probability
    assert 'next_order_delay_probability' in predictions
    assert 0 <= predictions['next_order_delay_probability'] <= 1.0

    # Poor supplier should have higher delay probability
    assert predictions['next_order_delay_probability'] > 0.3

    # Should have confidence score
    assert 'confidence' in predictions
    assert 0 <= predictions['confidence'] <= 100


@pytest.mark.asyncio
async def test_risk_assessment(sample_order_history_poor_supplier):
    """Test procurement risk assessment."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='test-supplier',
        order_history=sample_order_history_poor_supplier
    )

    risk_assessment = results['risk_assessment']

    # Check structure
    assert 'risk_level' in risk_assessment
    assert 'risk_score' in risk_assessment
    assert 'risk_factors' in risk_assessment
    assert 'recommendation' in risk_assessment

    # Risk level should be valid
    assert risk_assessment['risk_level'] in ['low', 'medium', 'high', 'critical']

    # Risk score should be 0-100
    assert 0 <= risk_assessment['risk_score'] <= 100

    # Should have risk factors for poor supplier
    assert len(risk_assessment['risk_factors']) > 0

    # Recommendation should be string
    assert isinstance(risk_assessment['recommendation'], str)
    assert len(risk_assessment['recommendation']) > 0


@pytest.mark.asyncio
async def test_insight_generation_low_reliability(sample_order_history_poor_supplier):
    """Test insight generation for low reliability supplier."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='poor-supplier',
        order_history=sample_order_history_poor_supplier
    )

    insights = results['insights']

    # Should generate insights
    assert len(insights) > 0

    # Check for low reliability alert
    reliability_insights = [i for i in insights
                           if 'reliability' in i.get('title', '').lower()]

    if reliability_insights:
        insight = reliability_insights[0]
        assert insight['type'] in ['alert', 'recommendation']
        assert insight['priority'] in ['high', 'critical']
        assert 'actionable' in insight
        assert insight['actionable'] is True
        assert 'recommendation_actions' in insight
        assert len(insight['recommendation_actions']) > 0


@pytest.mark.asyncio
async def test_insight_generation_high_delay_risk(sample_order_history_poor_supplier):
    """Test insight generation for high delay probability."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='poor-supplier',
        order_history=sample_order_history_poor_supplier
    )

    insights = results['insights']

    # Check for delay risk prediction
    delay_insights = [i for i in insights
                     if 'delay' in i.get('title', '').lower()]

    if delay_insights:
        insight = delay_insights[0]
        assert 'confidence' in insight
        assert 'metrics_json' in insight
        assert 'recommendation_actions' in insight


@pytest.mark.asyncio
async def test_insight_generation_excellent_supplier(sample_order_history_good_supplier):
    """Test that excellent suppliers get positive insights."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='excellent-supplier',
        order_history=sample_order_history_good_supplier
    )

    insights = results['insights']

    # Should have positive insight for excellent performance
    positive_insights = [i for i in insights
                        if 'excellent' in i.get('title', '').lower()]

    if positive_insights:
        insight = positive_insights[0]
        assert insight['type'] == 'insight'
        assert insight['impact_type'] == 'positive_performance'


def test_compare_suppliers():
    """Test supplier comparison functionality."""
    predictor = SupplierPerformancePredictor()

    # Mock analysis results
    suppliers_analysis = [
        {
            'supplier_id': 'supplier-1',
            'reliability_score': 95,
            'risk_assessment': {'risk_level': 'low', 'risk_score': 10}
        },
        {
            'supplier_id': 'supplier-2',
            'reliability_score': 60,
            'risk_assessment': {'risk_level': 'high', 'risk_score': 75}
        },
        {
            'supplier_id': 'supplier-3',
            'reliability_score': 80,
            'risk_assessment': {'risk_level': 'medium', 'risk_score': 40}
        }
    ]

    comparison = predictor.compare_suppliers(suppliers_analysis)

    # Check structure
    assert 'suppliers_compared' in comparison
    assert 'top_supplier' in comparison
    assert 'top_supplier_score' in comparison
    assert 'bottom_supplier' in comparison
    assert 'bottom_supplier_score' in comparison
    assert 'ranked_suppliers' in comparison
    assert 'recommendations' in comparison

    # Check ranking
    assert comparison['suppliers_compared'] == 3
    assert comparison['top_supplier'] == 'supplier-1'
    assert comparison['top_supplier_score'] == 95
    assert comparison['bottom_supplier'] == 'supplier-2'
    assert comparison['bottom_supplier_score'] == 60

    # Ranked suppliers should be in order
    ranked = comparison['ranked_suppliers']
    assert ranked[0]['supplier_id'] == 'supplier-1'
    assert ranked[-1]['supplier_id'] == 'supplier-2'

    # Should have recommendations
    assert len(comparison['recommendations']) > 0


@pytest.mark.asyncio
async def test_insufficient_data_handling():
    """Test handling of insufficient order history."""
    predictor = SupplierPerformancePredictor()

    # Only 5 orders (less than min_orders=10)
    small_history = pd.DataFrame([
        {
            'order_date': datetime(2024, 1, i),
            'expected_delivery_date': datetime(2024, 1, i+3),
            'actual_delivery_date': datetime(2024, 1, i+3),
            'order_quantity': 100,
            'received_quantity': 100,
            'quality_issues': False,
            'quality_score': 95.0,
            'order_value': 500.0
        }
        for i in range(1, 6)
    ])

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='new-supplier',
        order_history=small_history,
        min_orders=10
    )

    # Should return insufficient data response
    assert results['orders_analyzed'] == 0
    assert results['reliability_score'] is None
    assert results['risk_assessment']['risk_level'] == 'unknown'
    assert 'Insufficient' in results['risk_assessment']['risk_factors'][0]


def test_get_supplier_reliability_score():
    """Test getting cached reliability scores."""
    predictor = SupplierPerformancePredictor()

    # Initially no score
    assert predictor.get_supplier_reliability_score('supplier-1') is None

    # Set a score
    predictor.reliability_scores['supplier-1'] = 85

    # Should retrieve it
    assert predictor.get_supplier_reliability_score('supplier-1') == 85


@pytest.mark.asyncio
async def test_metrics_no_nan_values(sample_order_history_good_supplier):
    """Test that metrics never contain NaN values."""
    predictor = SupplierPerformancePredictor()

    results = await predictor.analyze_supplier_performance(
        tenant_id='test-tenant',
        supplier_id='test-supplier',
        order_history=sample_order_history_good_supplier
    )

    metrics = results['metrics']

    # Check no NaN values
    for key, value in metrics.items():
        if isinstance(value, float):
            assert not np.isnan(value), f"Metric {key} is NaN"