RawArchiver/Rac.Core/Tools/LinkProcessor.cs

using Alba.CsCss.Style;
using AngleSharp;
using AngleSharp.Html.Dom;
using AngleSharp.Html.Parser;
using System;
using System.IO;
using System.Linq;
using System.Text;

namespace Rac.Tools
{
    // This class can change to instance class
    //  Implement different driver (StringParser/WebBrowserEmu/Etc)

    public class LinkProcessor
    {
        private static readonly CssLoader CssLoader = new CssLoader();

        public static string[] FromCss(string css)
        {
            return CssLoader.GetUris(css).ToArray();
        }

        public static string[] ExtractLinks(IHtmlDocument doc)
        {
            //extract urls from embedded style
            var links = doc.QuerySelectorAll("style").Select(p => FromCss(p.TextContent)).SelectMany(p => p).ToList();

            //extract urls from page
            links.AddRange(doc.QuerySelectorAll("*[src],*[href]")
                .Select(e => e.Attributes["src"]?.Value ?? e.Attributes["href"]?.Value)
                .Where(p => p != null));

            return links.ToArray();
        }

        public static string ReplaceCssLinks(string css, Func<string, string> func)
        {
            var links = FromCss(css).OrderByDescending(p => p.Length).Distinct().Where(s => string.IsNullOrWhiteSpace(s) == false).ToArray();
            foreach (var link in links)
            {
                if (link != "//")
                    css = css.Replace(link, func(link));
            }
            return css;
        }

        public static string ReplaceHtmlLinks(byte[] html, Func<string, string> func, ref Encoding encoding)
        {
            IHtmlDocument doc;
            if (null == encoding)
            {
                using var stream = new MemoryStream(html);
                doc = new HtmlParser().ParseDocument(stream);
            }
            else
            {
                var source = encoding.GetString(html);
                doc = new HtmlParser().ParseDocument(source);
            }

            var styles = doc.QuerySelectorAll("style");
            foreach (var style in styles)
            {
                style.TextContent = ReplaceCssLinks(style.TextContent, func);
            }

            foreach (var srcEl in doc.QuerySelectorAll("*[src]")) if (srcEl.Attributes["src"]?.Value != "//") srcEl.Attributes["src"].Value = func(srcEl.Attributes["src"].Value);
            foreach (var srcEl in doc.QuerySelectorAll("*[href]")) if (srcEl.Attributes["href"]?.Value != "//") srcEl.Attributes["href"].Value = func(srcEl.Attributes["href"].Value);

            if (null == encoding) encoding = Encoding.GetEncoding(doc.CharacterSet);
            return doc.ToHtml();
        }
    }
}