/**
 * THE HIVE SCRUBBER
 * Sanitizes local rules and signals before they are broadcast to the Global Registry.
 * Ensures strict anonymity.
 */

export interface ScrubberResult {
    sanitizedContent: string;
    redactionCount: number;
    riskScore: number; // 0-100 (If too high, do not broadcast)
}

export class HiveScrubber {

    // Patterns that definitely identify a project and MUST be removed
    private static SENSITIVE_PATTERNS = [
        /(api_key|secret|token|password)[\s]*[:=][\s]*['"][a-zA-Z0-9_\-]+['"]/gi, // Secrets
        /[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}/g, // Emails
        /(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\. (?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)\.(?:25[0-5]|2[0-4][0-9]|[01]?[0-9][0-9]?)/g, // IPs
        /postgres:\/\/[^:]+:[^@]+@/g, // DB Connection Strings
    ];

    // Generic replacements for project-specific terms to keep the rule abstract
    private static ABSTRACTION_MAP: Record<string, string> = {
        'Vibeline': '{PROJECT_NAME}',
        'Rigstate': '{FRAMEWORK}',
        'Steinhofve': '{USER}',
    };

    /**
     * Scrubs a string (rule content or log excerpt) of sensitive data.
     */
    public static scrub(content: string, customTerms: string[] = []): ScrubberResult {
        let scrubbed = content;
        let count = 0;
        let risk = 0;

        // 1. Remove Hard Credentials (High Risk)
        this.SENSITIVE_PATTERNS.forEach(pattern => {
            if (pattern.test(scrubbed)) {
                scrubbed = scrubbed.replace(pattern, '[REDACTED_CREDENTIAL]');
                count++;
                risk += 50; // High penalty for finding secrets
            }
        });

        // 2. Abstract Project Specifics (Medium Risk)
        Object.entries(this.ABSTRACTION_MAP).forEach(([term, replacement]) => {
            const regex = new RegExp(term, 'gi');
            if (regex.test(scrubbed)) {
                scrubbed = scrubbed.replace(regex, replacement);
                count++;
            }
        });

        // 3. Scrub Custom Terms (e.g. Table names passed from context)
        customTerms.forEach(term => {
            const regex = new RegExp(term, 'gi');
            if (regex.test(scrubbed)) {
                scrubbed = scrubbed.replace(regex, '{ENTITY}');
                count++;
            }
        });

        return {
            sanitizedContent: scrubbed,
            redactionCount: count,
            riskScore: Math.min(risk, 100)
        };
    }
}