youtube_personal/lib/services/preferred_language_service.dart

import '../models/video_item.dart';

class PreferredLanguageOption {
  const PreferredLanguageOption({
    required this.code,
    required this.label,
  });

  final String code;
  final String label;
}

class PreferredLanguageService {
  PreferredLanguageService._();

  static const List<PreferredLanguageOption> options =
      <PreferredLanguageOption>[
    PreferredLanguageOption(code: 'auto', label: 'Auto'),
    PreferredLanguageOption(code: 'fr', label: 'Français (fr)'),
    PreferredLanguageOption(code: 'en', label: 'English (en)'),
    PreferredLanguageOption(code: 'ar', label: 'العربية (ar)'),
    PreferredLanguageOption(code: 'de', label: 'Deutsch (de)'),
    PreferredLanguageOption(code: 'es', label: 'Español (es)'),
  ];

  static bool isSupported(String code) {
    return options.any((option) => option.code == code);
  }

  static String labelFor(String code) {
    for (final option in options) {
      if (option.code == code) {
        return option.label;
      }
    }
    return 'Auto';
  }

  static String nativeSearchHint(String code) {
    return switch (code) {
      'fr' => 'français',
      'en' => 'English',
      'ar' => 'عربي',
      'de' => 'Deutsch',
      'es' => 'español',
      _ => '',
    };
  }

  static String dubbingSearchHint(String code) {
    return switch (code) {
      'fr' => 'VF doublage français version française',
      'en' => 'English dub dubbed',
      'ar' => 'مدبلج عربي دبلجة عربية',
      'de' => 'deutsche Synchronisation Deutsch',
      'es' => 'doblaje español versión española',
      _ => '',
    };
  }

  /// Returns a positive local boost. This is deliberately a preference,
  /// never a hard filter: videos in other languages remain eligible.
  static int scoreVideo(VideoItem video, String preferredLanguage) {
    if (preferredLanguage == 'auto') {
      return 0;
    }

    final raw = '${video.title} ${video.description} ${video.author}';
    final text = _normalize(raw);

    var score = 0;

    if (_hasDubbingMarker(text, preferredLanguage)) {
      score += 18;
      score += _popularityBonus(video.viewCount);
    }

    if (_looksLikeLanguage(text, preferredLanguage)) {
      score += 12;
    }

    return score.clamp(0, 32).toInt();
  }

  static bool isLikelyDubbed(VideoItem video, String preferredLanguage) {
    if (preferredLanguage == 'auto') {
      return false;
    }

    final text = _normalize(
      '${video.title} ${video.description} ${video.author}',
    );

    return _hasDubbingMarker(text, preferredLanguage);
  }

  static List<VideoItem> rankVideos(
    List<VideoItem> videos,
    String preferredLanguage,
  ) {
    if (preferredLanguage == 'auto' || videos.length < 2) {
      return videos;
    }

    final indexed = videos.asMap().entries.toList(growable: false);

    final sorted = [...indexed]
      ..sort((a, b) {
        final aScore = scoreVideo(a.value, preferredLanguage);
        final bScore = scoreVideo(b.value, preferredLanguage);

        final scoreCompare = bScore.compareTo(aScore);
        if (scoreCompare != 0) {
          return scoreCompare;
        }

        return a.key.compareTo(b.key);
      });

    return sorted.map((entry) => entry.value).toList(growable: false);
  }


  static int _popularityBonus(int? viewCount) {
    if (viewCount == null || viewCount <= 0) {
      return 0;
    }

    if (viewCount >= 10000000) {
      return 8;
    }
    if (viewCount >= 1000000) {
      return 6;
    }
    if (viewCount >= 100000) {
      return 4;
    }
    if (viewCount >= 10000) {
      return 2;
    }
    return 1;
  }

  static bool _hasDubbingMarker(String text, String code) {
    final markers = switch (code) {
      'fr' => const <String>[
          ' vf ',
          '[vf]',
          '(vf)',
          'version francaise',
          'version française',
          'double en francais',
          'doublé en français',
          'doublage francais',
          'doublage français',
          'french dub',
          'french dubbed',
          'audio francais',
          'audio français',
        ],
      'en' => const <String>[
          'english dub',
          'english dubbed',
          'dubbed in english',
          'english audio',
        ],
      'ar' => const <String>[
          'مدبلج عربي',
          'مدبلجة عربية',
          'دبلجة عربية',
          'باللغة العربية',
        ],
      'de' => const <String>[
          'deutsche synchronisation',
          'deutsch synchronisiert',
          'deutsche synchro',
          'german dub',
        ],
      'es' => const <String>[
          'doblaje espanol',
          'doblaje español',
          'version espanola',
          'versión española',
          'spanish dub',
          'audio espanol',
          'audio español',
        ],
      _ => const <String>[],
    };

    return markers.any(text.contains);
  }

  static bool _looksLikeLanguage(String text, String code) {
    if (code == 'ar') {
      final arabic = RegExp(r'[\u0600-\u06FF]');
      final matches = arabic.allMatches(text).length;
      return matches >= 5;
    }

    final words = text
        .split(RegExp(r'\s+'))
        .where((word) => word.length >= 2)
        .toList(growable: false);

    final markers = switch (code) {
      'fr' => const <String>{
          'le', 'la', 'les', 'un', 'une', 'des', 'du', 'de', 'dans', 'avec',
          'pour', 'sur', 'est', 'sont', 'comment', 'pourquoi', 'nouveau',
          'nouvelle', 'francais', 'français', 'voici', 'tout', 'tous',
        },
      'en' => const <String>{
          'the', 'and', 'with', 'for', 'from', 'this', 'that', 'how', 'why',
          'what', 'new', 'best', 'review', 'explained', 'english',
        },
      'de' => const <String>{
          'der', 'die', 'das', 'und', 'mit', 'für', 'fur', 'von', 'ist', 'wie',
          'warum', 'neu', 'neue', 'deutsch', 'erklart', 'erklärt',
        },
      'es' => const <String>{
          'el', 'la', 'los', 'las', 'un', 'una', 'con', 'para', 'por', 'como',
          'cómo', 'nuevo', 'nueva', 'espanol', 'español', 'explicado',
        },
      _ => const <String>{},
    };

    var hits = 0;
    for (final word in words) {
      if (markers.contains(word)) {
        hits++;
      }
    }

    return hits >= 2;
  }

  static String _normalize(String value) {
    return ' ${value.toLowerCase()}'
        .replaceAll(RegExp(r'[^a-z0-9à-ÿ\u0600-\u06ff\[\]\(\)]+'), ' ')
        .replaceAll(RegExp(r'\s+'), ' ')
        .trimRight() +
        ' ';
  }
}