Implemented file cache for document helper and some minor bug fixes

ChromiumHtmlToPdfLib/Browser.cs

@@ -743,11 +743,25 @@ internal void WriteToLog(string message)
     /// </summary>
     public void Dispose()
     {
-        _pageConnection.OnError -= OnOnError;
-        _pageConnection?.Dispose();
+        if (_pageConnection != null)
+            _pageConnection.OnError -= OnOnError;
+
+        if (_browserConnection != null)
+            _browserConnection.OnError -= OnOnError;
+
+        CloseAsync(CancellationToken.None).GetAwaiter().GetResult();
 
-        _browserConnection.OnError -= OnOnError;
-        _browserConnection?.Dispose();
+        if (_pageConnection != null)
+        {
+            _pageConnection.Dispose();
+            _pageConnection = null;
+        }
+
+        if (_browserConnection != null)
+        {
+            _browserConnection.Dispose();
+            _browserConnection = null;
+        }
     }
     #endregion
 
@@ -759,15 +773,21 @@ public void Dispose()
     public async ValueTask DisposeAsync()
     {
         if (_pageConnection != null)
-        {
             _pageConnection.OnError -= OnOnError;
+
+        if (_browserConnection != null)
+            _browserConnection.OnError -= OnOnError;
+
+        await CloseAsync(CancellationToken.None);
+
+        if (_pageConnection != null)
+        {
             await _pageConnection.DisposeAsync();
             _pageConnection = null;
         }
 
-        if (_pageConnection != null)
+        if (_browserConnection != null)
         {
-            _browserConnection.OnError -= OnOnError;
             await _browserConnection.DisposeAsync();
             _browserConnection = null;
         }

ChromiumHtmlToPdfLib/ChromiumHtmlToPdfLib.csproj

@@ -34,6 +34,7 @@
     <PackageReference Include="AngleSharp.Css" Version="0.17.0" />
     <PackageReference Include="AngleSharp.Io" Version="1.0.0" />
     <PackageReference Include="AngleSharp.Xml" Version="1.0.0" />
+    <PackageReference Include="FileCache" Version="3.3.0" />
     <PackageReference Include="HtmlSanitizer" Version="8.0.723" />
     <PackageReference Include="Microsoft.Extensions.Logging.Abstractions" Version="7.0.1" />
     <PackageReference Include="Microsoft.Win32.Registry" Version="5.0.0" />

ChromiumHtmlToPdfLib/Connection.cs

@@ -159,9 +159,9 @@ private async Task ReceiveLoop()
         {
             // Ignore
         }
-        catch (Exception e)
+        catch (Exception exception)
         {
-            WebSocketOnError(new ErrorEventArgs(e));
+            WebSocketOnError(new ErrorEventArgs(exception));
         }
         finally
         {
@@ -361,29 +361,25 @@ public async Task InternalDisposeAsync()
     {
         WriteToLog($"Disposing websocket connection to url '{_url}'");
 
-        try
-        {
-            WebSocketOnClosed(EventArgs.Empty);
-        }
-        catch (Exception exception)
-        {
-            WriteToLog($"Exception while disposing websocket connection to url '{_url}': {exception.Message}");
-        }
-
         if (_webSocket.State == WebSocketState.Open)
         {
             WriteToLog("Closing websocket");
+
             try
             {
                 await _webSocket.CloseAsync(WebSocketCloseStatus.NormalClosure, "Done", default);
-                WriteToLog("Websocket connection disposed gracefully");
             }
-            finally
+            catch 
             {
-                _receiveLoopCts.Cancel();
-                _webSocket.Dispose();
-                WriteToLog("Websocket connection disposed");
+                // Ignore
             }
+
+            WriteToLog("Websocket connection closed");
+
+            WebSocketOnClosed(EventArgs.Empty);
+            _receiveLoopCts.Cancel();
+            _webSocket.Dispose();
+            WriteToLog("Websocket connection disposed");
         }
     }
     #endregion

ChromiumHtmlToPdfLib/Converter.cs

@@ -191,7 +191,10 @@ private enum OutputFormat
     /// <summary>
     ///     <see cref="SetDiskCache"/>
     /// </summary>
-    private long _cacheSize;
+    /// <remarks>
+    ///     Default set to 1GB
+    /// </remarks>
+    private long _cacheSize = 1073741824;
 
     /// <summary>
     ///     <see cref="GetDocumentHelper"/>
@@ -647,9 +650,7 @@ private void StartChromiumHeadless()
                 processStartInfo.LoadUserProfile = true;
             }
             else
-            {
                 WriteToLog("Ignoring password and loading user profile because this is only supported on Windows");
-            }
         }
 
         if (!_userProfileSet)
@@ -687,8 +688,7 @@ private void StartChromiumHeadless()
 
             if (_conversionTimeout.HasValue)
                 if (!_chromiumWaitEvent.WaitOne(_conversionTimeout.Value))
-                    throw new ChromiumException(
-                        $"A timeout of '{_conversionTimeout.Value}' milliseconds exceeded, could not make a connection to the Chromium dev tools");
+                    throw new ChromiumException($"A timeout of '{_conversionTimeout.Value}' milliseconds exceeded, could not make a connection to the Chromium dev tools");
 
             _chromiumWaitEvent.WaitOne();
 
@@ -2610,7 +2610,6 @@ private async Task InternalDisposeAsync()
             try
             {
                 WriteToLog($"Closing {BrowserName} browser gracefully");
-                await _browser.CloseAsync(default);
 #if (NETSTANDARD2_0)
                 _browser.Dispose();
 #else

ChromiumHtmlToPdfLib/Helpers/DocumentHelper.cs

@@ -33,6 +33,7 @@
 using System.Linq;
 using System.Net;
 using System.Net.Http;
+using System.Runtime.Caching;
 using System.Text.RegularExpressions;
 using System.Threading;
 using System.Threading.Tasks;
@@ -95,12 +96,17 @@ internal class DocumentHelper : IDisposable
     /// <summary>
     ///     The cache folder
     /// </summary>
-    readonly DirectoryInfo _cacheDirectory;
+    private DirectoryInfo _cacheDirectory;
 
     /// <summary>
     ///     The cache size
     /// </summary>
-    private readonly long? _cacheSize;
+    private readonly long _cacheSize;
+
+    /// <summary>
+    ///     <see cref="FileCache"/>
+    /// </summary>
+    private FileCache _fileCache;
     #endregion
 
     #region Properties
@@ -133,6 +139,35 @@ internal int TimeLeft
             return (int)result;
         }
     }
+
+    /// <summary>
+    ///     Returns a file cache
+    /// </summary>
+    private FileCache FileCache
+    {
+        get
+        {
+            if (_fileCache != null)
+                return _fileCache;
+
+            _cacheDirectory = new DirectoryInfo(Path.Combine("d:\\", "HttpClientHandler"));
+            //_cacheDirectory = new DirectoryInfo(Path.Combine(_cacheDirectory.FullName, "HttpClientHandler"));
+
+            if (!_cacheDirectory.Exists)
+                _cacheDirectory.Create();
+
+            FileCache.DefaultCacheManager = FileCacheManagers.Hashed;
+
+            _fileCache = new FileCache(_cacheDirectory.FullName)
+            {
+                MaxCacheSize = _cacheSize,
+                AccessTimeout = TimeSpan.FromSeconds(10),
+                DefaultPolicy = new CacheItemPolicy { SlidingExpiration = TimeSpan.FromDays(1) },
+            };
+
+            return _fileCache;
+        }
+    }
     #endregion
 
     #region Constructor
@@ -149,7 +184,7 @@ internal int TimeLeft
     public DocumentHelper(DirectoryInfo tempDirectory,
         bool useCache,
         DirectoryInfo cacheDirectory, 
-        long? cacheSize,
+        long cacheSize,
         WebProxy webProxy,
         int? imageLoadTimeout,
         ILogger logger)
@@ -916,33 +951,48 @@ private async Task<Stream> OpenDownloadStream(Uri sourceUri, bool checkTimeout =
     {
         try
         {
-            var httpClientHandler = new HttpClientHandler();
-
-            if (_webProxy != null)
-                httpClientHandler.Proxy = _webProxy;
+            if (_useCache)
+            {
+                var item = FileCache.GetCacheItem(sourceUri.ToString());
+                if (item is { Value: not null })
+                {
+                    WriteToLog($"Returning stream for url '{sourceUri}' from the cache");
+                    return new MemoryStream((byte[])item.Value);
+                }
+            }
 
-            var handler = new FileCacheHandler(httpClientHandler, _cacheDirectory, _cacheSize.Value);
-            using var client = new HttpClient(handler);
+            var request = WebRequest.CreateHttp(sourceUri);
             var timeLeft = TimeLeft;
 
-            //if (_stopwatch != null && checkTimeout)
-            //{
-            //    if (timeLeft == 0)
-            //    {
-            //        WriteToLog($"Image load has timed out, skipping opening stream to url '{sourceUri}'");
-            //        return null;
-            //    }
-
-            //    request.Timeout = TimeLeft;
-            //}
-
+            if (_stopwatch != null && checkTimeout)
+            {
+                if (timeLeft == 0)
+                {
+                    WriteToLog($"Image load has timed out, skipping opening stream to url '{sourceUri}'");
+                    return null;
+                }
 
+                request.Timeout = TimeLeft;
+            }
 
             WriteToLog($"Opening stream to url '{sourceUri}'{(_stopwatch != null ? $" with a timeout of {timeLeft} milliseconds" : string.Empty)}");
-            var response = await client.GetAsync(sourceUri);
 
-            //WriteToLog($"Opened {(response.IsFromCache ? "cached " : string.Empty)}stream to url '{sourceUri}'");
-            return await response.Content.ReadAsStreamAsync();
+            var response = await request.GetResponseAsync();
+
+            if (!_useCache) 
+                return response.GetResponseStream();
+
+            var stream = response.GetResponseStream();
+
+            if (stream == null) 
+                return response.GetResponseStream();
+
+            var memoryStream = new MemoryStream();
+            await stream.CopyToAsync(memoryStream);
+            WriteToLog($"Adding item from url '{sourceUri}' to the cache");
+            FileCache.Add(sourceUri.ToString(), memoryStream.ToArray(), new CacheItemPolicy { SlidingExpiration = TimeSpan.FromDays(1) });
+            memoryStream.Position = 0;
+            return memoryStream;
         }
         catch (Exception exception)
         {