Nodecast/public/js/services/channelLanguageDetection.js

(function (root, factory) {
    const service = factory();
    if (typeof module === 'object' && module.exports) module.exports = service;
    if (root) root.ChannelLanguageDetection = service;
})(typeof globalThis !== 'undefined' ? globalThis : this, function () {
    const LANGUAGES = {
        fr: { code: 'fr', label: 'French' },
        en: { code: 'en', label: 'English' },
        es: { code: 'es', label: 'Spanish' }
    };
    const UNKNOWN = { code: 'unknown', label: 'Unknown', confidence: 0, status: 'unknown' };
    const EXPLICIT_CODES = { FR: 'fr', SFR: 'fr', 'FR-CAR': 'fr', EN: 'en', UK: 'en', US: 'en', AU: 'en', ES: 'es' };
    const WORDS = {
        fr: ['french', 'francais', 'france'],
        en: ['english', 'anglais', 'england', 'british'],
        es: ['spanish', 'espagnol', 'espanol', 'spain']
    };

    function normalized(value) {
        return String(value || '').normalize('NFKD').replace(/[\u0300-\u036f]/g, '').toLowerCase();
    }

    function addEvidence(scores, evidence, code, weight, type, value) {
        if (!code) return;
        scores[code] = (scores[code] || 0) + weight;
        evidence.push({ code, weight, type, value: String(value || '') });
    }

    function findExplicitPrefix(name) {
        const match = String(name || '').trim().match(/^(?:\|\s*([A-Z-]{2,8})\s*\||\[\s*([A-Z-]{2,8})\s*\]|([A-Z-]{2,8})\s*\|)/i);
        return match ? (match[1] || match[2] || match[3]).toUpperCase() : null;
    }

    function scoreText(scores, evidence, value, weight, type, includeCountryCodes) {
        const text = normalized(value);
        if (!text) return;
        Object.entries(WORDS).forEach(([code, terms]) => {
            if (terms.some(term => new RegExp(`(?:^|[^a-z])${term}(?:$|[^a-z])`, 'i').test(text))) {
                addEvidence(scores, evidence, code, weight, type, value);
            }
        });
        if (includeCountryCodes) {
            const tokens = text.toUpperCase().split(/[^A-Z]+/).filter(Boolean);
            const mapped = tokens.map(token => EXPLICIT_CODES[token]).filter(Boolean);
            [...new Set(mapped)].forEach(code => addEvidence(scores, evidence, code, weight, type, value));
        }
    }

    function detectLanguage(variant, context = {}) {
        const scores = {};
        const evidence = [];
        const prefix = findExplicitPrefix(variant.name);
        addEvidence(scores, evidence, EXPLICIT_CODES[prefix], 60, 'name-prefix', prefix);
        scoreText(scores, evidence, variant.name, 45, 'name-token', false);
        scoreText(scores, evidence, variant.groupTitle, 35, 'group', true);
        scoreText(scores, evidence, context.sourceName || variant.sourceName, 20, 'source', true);
        scoreText(scores, evidence, `${variant.tvgId || ''} ${variant.tvgName || ''}`, 20, 'metadata', true);

        const ranked = Object.entries(scores).sort((a, b) => b[1] - a[1]);
        const total = ranked.reduce((sum, entry) => sum + entry[1], 0);
        const percentages = Object.fromEntries(ranked.map(([code, score]) => [code, Math.round(score / total * 100)]));
        if (!ranked.length || total < 20) {
            return { language: { ...UNKNOWN }, languageScores: percentages, languageEvidence: evidence };
        }

        const [topCode, topScore] = ranked[0];
        const topConfidence = percentages[topCode];
        const secondConfidence = ranked[1] ? percentages[ranked[1][0]] : 0;
        if (topScore < 20 || topConfidence < 45 || topConfidence - secondConfidence < 10) {
            return { language: { ...UNKNOWN }, languageScores: percentages, languageEvidence: evidence };
        }

        let status = 'low';
        if (topScore >= 60 && topConfidence >= 75) status = 'high';
        else if (topScore >= 35 && topConfidence >= 60) status = 'medium';
        return {
            language: { ...LANGUAGES[topCode], confidence: topConfidence, status },
            languageScores: percentages,
            languageEvidence: evidence
        };
    }

    return { LANGUAGES, detectLanguage };
});