import numpy as np
from scipy import stats
import json

class APAReport:
    """Generate APA 7th edition compliant statistical reports."""

    @staticmethod
    def format_number(value, decimals=2):
        """APA: no leading zero for p-values, effect sizes; zero before decimal for others."""
        if abs(value) < 1 and decimals == 2:
            return f".{str(round(value, decimals)).split('.')[1]}"
        return f"{round(value, decimals)}"

    @staticmethod
    def t_test_report(group1, group2, group1_name="Group 1", group2_name="Group 2",
                      alpha=0.05):
        """Generate APA-formatted t-test report."""
        n1, n2 = len(group1), len(group2)
        m1, m2 = np.mean(group1), np.mean(group2)
        sd1, sd2 = np.std(group1, ddof=1), np.std(group2, ddof=1)

        # Welch's t-test
        t_stat, p_value = stats.ttest_ind(group1, group2, equal_var=False)
        df = (sd1**2/n1 + sd2**2/n2)**2 / \
             ((sd1**2/n1)**2/(n1-1) + (sd2**2/n2)**2/(n2-1))

        # Pooled SD for Cohen's d
        sp = np.sqrt(((n1-1)*sd1**2 + (n2-1)*sd2**2) / (n1+n2-2))
        d = (m1 - m2) / sp

        # CI for mean difference
        se_diff = np.sqrt(sd1**2/n1 + sd2**2/n2)
        t_crit = stats.t.ppf(1 - alpha/2, df)
        ci_low = (m1 - m2) - t_crit * se_diff
        ci_high = (m1 - m2) + t_crit * se_diff

        # Format according to APA
        report = {
            "descriptive": {
                group1_name: {"M": round(m1, 2), "SD": round(sd1, 2), "n": n1},
                group2_name: {"M": round(m2, 2), "SD": round(sd2, 2), "n": n2}
            },
            "test": {
                "statistic": f"t({round(df, 1)}) = {round(t_stat, 2)}",
                "p_value": f"p = {APAReport.format_number(p_value)}" if p_value >= .001 else "p < .001",
                "effect_size": f"d = {APAReport.format_number(d)}",
                "ci_95": f"95% CI [{APAReport.format_number(ci_low)}, {APAReport.format_number(ci_high)}]"
            },
            "narrative": (
                f"{group1_name} (M = {round(m1,2)}, SD = {round(sd1,2)}) "
                f"{'significantly higher' if m1 > m2 else 'significantly lower'} "
                f"than {group2_name} (M = {round(m2,2)}, SD = {round(sd2,2)}), "
                f"t({round(df,1)}) = {round(t_stat,2)}, "
                f"{'p < .001' if p_value < .001 else f'p = {APAReport.format_number(p_value)}'}, "
                f"d = {APAReport.format_number(d)}, "
                f"95% CI [{APAReport.format_number(ci_low)}, {APAReport.format_number(ci_high)}]."
            )
        }
        return report

    @staticmethod
    def anova_report(groups, group_names=None, alpha=0.05):
        """Generate APA-formatted one-way ANOVA report."""
        if group_names is None:
            group_names = [f"Group {i+1}" for i in range(len(groups))]

        f_stat, p_value = stats.f_oneway(*groups)

        # Effect size (eta-squared)
        all_data = np.concatenate(groups)
        grand_mean = np.mean(all_data)
        ss_between = sum(len(g) * (np.mean(g) - grand_mean)**2 for g in groups)
        ss_total = np.sum((all_data - grand_mean)**2)
        eta_sq = ss_between / ss_total

        # Report
        report = {
            "test": f"F({len(groups)-1}, {len(all_data)-len(groups)}) = {round(f_stat, 2)}",
            "p_value": f"p = {APAReport.format_number(p_value)}" if p_value >= .001 else "p < .001",
            "effect_size": f"η² = {APAReport.format_number(eta_sq)}",
            "group_means": {name: round(np.mean(g), 2) for name, g in zip(group_names, groups)}
        }
        return report

# Example usage
np.random.seed(42)
treatment = np.random.normal(82, 7, 50)
control = np.random.normal(75, 8, 50)

t_report = APAReport.t_test_report(treatment, control, "Treatment", "Control")
print("=== APA T-Test Report ===")
print(t_report["narrative"])
print(json.dumps(t_report["test"], indent=2))

print("\n=== APA ANOVA Report ===")
g1 = np.random.normal(100, 15, 30)
g2 = np.random.normal(108, 15, 30)
g3 = np.random.normal(112, 15, 30)
anova = APAReport.anova_report([g1, g2, g3], ["Control", "Treatment A", "Treatment B"])
print(anova)