tests/test_scoring.py

"""Tests for reward and final score computation."""

import pytest

from hr_env.server.scoring import compute_final_score, compute_quarterly_reward


class TestQuarterlyReward:
    def test_positive_for_improvement(self):
        prev = {"hcva": 100, "hcroi": 1.5, "qips": {"composite": 0.6}, "five_indexes": {"cost": -0.05, "quality": 0.1, "human_reactions": 0.05, "quantity": 0.02}}
        curr = {"hcva": 110, "hcroi": 1.6, "qips": {"composite": 0.65}, "five_indexes": {"cost": -0.03, "quality": 0.12, "human_reactions": 0.08, "quantity": 0.03}}
        reward = compute_quarterly_reward(curr, prev)
        assert reward > 0, f"Reward should be positive for improvement, got {reward}"

    def test_negative_for_decline(self):
        prev = {"hcva": 110, "hcroi": 1.6, "qips": {"composite": 0.65}, "five_indexes": {"cost": 0.0, "quality": 0.0, "human_reactions": 0.0, "quantity": 0.0}}
        curr = {"hcva": 90, "hcroi": 1.3, "qips": {"composite": 0.55}, "five_indexes": {"cost": 0.1, "quality": -0.1, "human_reactions": -0.1, "quantity": -0.05}}
        reward = compute_quarterly_reward(curr, prev)
        assert reward < 0, f"Reward should be negative for decline, got {reward}"

    def test_near_zero_for_no_change(self):
        metrics = {"hcva": 100, "hcroi": 1.5, "qips": {"composite": 0.6}, "five_indexes": {"cost": 0.0, "quality": 0.0, "human_reactions": 0.0, "quantity": 0.0}}
        reward = compute_quarterly_reward(metrics, metrics)
        assert abs(reward) < 0.1, f"Reward should be near zero, got {reward}"


class TestFinalScore:
    def test_bounded_zero_one(self):
        history = [
            {"hcva": 100, "hcroi": 1.5, "qips": {"composite": 0.6}, "five_indexes": {"cost": 0, "quality": 0, "human_reactions": 0}, "employee_value": 0.5, "profit": 1000000},
            {"hcva": 105, "hcroi": 1.55, "qips": {"composite": 0.62}, "five_indexes": {"cost": -0.02, "quality": 0.05, "human_reactions": 0.03}, "employee_value": 0.52, "profit": 1100000},
            {"hcva": 110, "hcroi": 1.6, "qips": {"composite": 0.64}, "five_indexes": {"cost": -0.01, "quality": 0.03, "human_reactions": 0.02}, "employee_value": 0.55, "profit": 1200000},
        ]
        score = compute_final_score(history)
        assert 0 <= score <= 1.0, f"Score {score} out of [0, 1] range"

    def test_higher_for_improvement_trajectory(self):
        improving = [
            {"hcva": 100, "hcroi": 1.5, "qips": {"composite": 0.5}, "five_indexes": {"cost": 0, "quality": 0, "human_reactions": 0}, "employee_value": 0.4, "profit": 500000},
            {"hcva": 120, "hcroi": 1.7, "qips": {"composite": 0.6}, "five_indexes": {"cost": -0.05, "quality": 0.1, "human_reactions": 0.1}, "employee_value": 0.5, "profit": 700000},
            {"hcva": 140, "hcroi": 1.9, "qips": {"composite": 0.7}, "five_indexes": {"cost": -0.03, "quality": 0.08, "human_reactions": 0.05}, "employee_value": 0.6, "profit": 900000},
        ]
        declining = [
            {"hcva": 140, "hcroi": 1.9, "qips": {"composite": 0.7}, "five_indexes": {"cost": 0, "quality": 0, "human_reactions": 0}, "employee_value": 0.6, "profit": 900000},
            {"hcva": 120, "hcroi": 1.5, "qips": {"composite": 0.55}, "five_indexes": {"cost": 0.1, "quality": -0.1, "human_reactions": -0.1}, "employee_value": 0.45, "profit": 500000},
            {"hcva": 100, "hcroi": 1.2, "qips": {"composite": 0.4}, "five_indexes": {"cost": 0.15, "quality": -0.15, "human_reactions": -0.12}, "employee_value": 0.35, "profit": 200000},
        ]
        score_improving = compute_final_score(improving)
        score_declining = compute_final_score(declining)
        assert score_improving > score_declining

    def test_empty_history(self):
        score = compute_final_score([])
        assert score == 0.0

    def test_single_quarter(self):
        score = compute_final_score([{"hcva": 100}])
        assert score == 0.0

    def test_profitable_scores_higher(self):
        profitable = [
            {"hcva": 100, "hcroi": 1.5, "qips": {"composite": 0.6}, "five_indexes": {"cost": 0, "quality": 0, "human_reactions": 0}, "employee_value": 0.5, "profit": 1000000},
            {"hcva": 100, "hcroi": 1.5, "qips": {"composite": 0.6}, "five_indexes": {"cost": 0, "quality": 0, "human_reactions": 0}, "employee_value": 0.5, "profit": 1000000},
        ]
        unprofitable = [
            {"hcva": 100, "hcroi": 1.5, "qips": {"composite": 0.6}, "five_indexes": {"cost": 0, "quality": 0, "human_reactions": 0}, "employee_value": 0.5, "profit": -500000},
            {"hcva": 100, "hcroi": 1.5, "qips": {"composite": 0.6}, "five_indexes": {"cost": 0, "quality": 0, "human_reactions": 0}, "employee_value": 0.5, "profit": -500000},
        ]
        assert compute_final_score(profitable) > compute_final_score(unprofitable)