QMAX/server/QMax.Api/Services/MessageTextSanitizer.cs

namespace QMax.Api.Services;

public static class MessageTextSanitizer
{
    public static string? CleanChatPreview(string? value)
    {
        if (IsGenericMediaLabel(value))
        {
            return "\u041c\u0435\u0434\u0438\u0430";
        }

        return IsQuestionMarkArtifact(value)
            ? "\u0421\u043e\u043e\u0431\u0449\u0435\u043d\u0438\u0435"
            : value;
    }

    public static string? CleanMessageText(string? value, bool hasAttachments)
    {
        if (hasAttachments && IsGenericAttachmentLabel(value))
        {
            return null;
        }

        if (!IsQuestionMarkArtifact(value))
        {
            return value;
        }

        return hasAttachments
            ? null
            : "\u0421\u043e\u043e\u0431\u0449\u0435\u043d\u0438\u0435";
    }

    public static bool IsGenericMediaLabel(string? value)
    {
        return value?.Trim().ToLowerInvariant() switch
        {
            "\u043c\u0435\u0434\u0438\u0430" => true,
            "media" => true,
            "\u0444\u043e\u0442\u043e" => true,
            "photo" => true,
            "image" => true,
            "gif" => true,
            "\u0432\u0438\u0434\u0435\u043e" => true,
            "video" => true,
            "\u0430\u0443\u0434\u0438\u043e" => true,
            "audio" => true,
            "\u0433\u043e\u043b\u043e\u0441\u043e\u0432\u043e\u0435" => true,
            "voice" => true,
            "\u0441\u0442\u0438\u043a\u0435\u0440" => true,
            "sticker" => true,
            "\u044d\u043c\u043e\u0434\u0437\u0438" => true,
            "emoji" => true,
            "\u043f\u0440\u0438\u043a\u0440\u0435\u043f\u043b\u0435\u043d\u043d\u044b\u0435 \u0444\u043e\u0442\u043e" => true,
            "\u043f\u0440\u0438\u043a\u0440\u0435\u043f\u043b\u0451\u043d\u043d\u044b\u0435 \u0444\u043e\u0442\u043e" => true,
            "attached photos" => true,
            _ => false
        };
    }

    public static bool IsGenericAttachmentLabel(string? value)
    {
        if (IsGenericMediaLabel(value))
        {
            return true;
        }

        return value?.Trim().ToLowerInvariant() switch
        {
            "\u0441\u043a\u0430\u0447\u0430\u0442\u044c" => true,
            "download" => true,
            "\u0437\u0430\u0433\u0440\u0443\u0437\u0438\u0442\u044c" => true,
            "file" => true,
            "\u0444\u0430\u0439\u043b" => true,
            "document" => true,
            "\u0434\u043e\u043a\u0443\u043c\u0435\u043d\u0442" => true,
            "contact" => true,
            "\u043a\u043e\u043d\u0442\u0430\u043a\u0442" => true,
            _ => false
        };
    }

    public static bool IsQuestionMarkArtifact(string? value)
    {
        var text = value?.Trim();
        if (string.IsNullOrWhiteSpace(text))
        {
            return false;
        }

        var questionMarks = text.Count(c => c is '?' or '\uFFFD');
        if (questionMarks < 4)
        {
            return false;
        }

        var meaningfulCharacters = text.Count(c => char.IsLetterOrDigit(c) && c != '?');
        return meaningfulCharacters == 0;
    }
}