EPUB CSS Parsing Issues (#690)

* WIP. Rewrote some of the Regex to better support css escaping. We now escape background-image, border-image, and list-style-image within css files.

* Added position relative to help with positioning on books that are just absolute positioned elements.

* When there is absolute positioning, like in some epub based comics, supress the bottom action bar since it wont render in the correct location.

* Fixed tests

* Commented out tests

API/Parser/Parser.cs

@@ -24,11 +24,25 @@ namespace API.Parser
         private const RegexOptions MatchOptions =
             RegexOptions.IgnoreCase | RegexOptions.Compiled | RegexOptions.CultureInvariant;
 
-        public static readonly Regex FontSrcUrlRegex = new Regex(@"(src:url\(.{1})" + "([^\"']*)" + @"(.{1}\))",
+        /// <summary>
+        /// Matches against font-family css syntax. Does not match if url import has data: starting, as that is binary data
+        /// </summary>
+        /// <remarks>See here for some examples https://developer.mozilla.org/en-US/docs/Web/CSS/@font-face</remarks>
+        public static readonly Regex FontSrcUrlRegex = new Regex(@"(?<Start>(src:\s?)?url\((?!data:).(?!data:))" + "(?<Filename>(?!data:)[^\"']*)" + @"(?<End>.{1}\))",
             MatchOptions, RegexTimeout);
-        public static readonly Regex CssImportUrlRegex = new Regex("@import\\s([\"|']|url\\([\"|'])(?<Filename>[^'\"]+)[\"|']\\)?;",
+        /// <summary>
+        /// https://developer.mozilla.org/en-US/docs/Web/CSS/@import
+        /// </summary>
+        public static readonly Regex CssImportUrlRegex = new Regex("(@import\\s([\"|']|url\\([\"|']))(?<Filename>[^'\"]+)([\"|']\\)?);",
+            MatchOptions | RegexOptions.Multiline, RegexTimeout);
+        /// <summary>
+        /// Misc css image references, like background-image: url(), border-image, or list-style-image
+        /// </summary>
+        /// Original prepend: (background|border|list-style)-image:\s?)?
+        public static readonly Regex CssImageUrlRegex = new Regex(@"(url\((?!data:).(?!data:))" + "(?<Filename>(?!data:)[^\"']*)" + @"(.\))",
             MatchOptions, RegexTimeout);
 
+
         private static readonly string XmlRegexExtensions = @"\.xml";
         private static readonly Regex ImageRegex = new Regex(ImageFileExtensions,
             MatchOptions, RegexTimeout);

API/Services/BookService.cs

@@ -140,13 +140,18 @@ namespace API.Services
             }
 
             stylesheetHtml = stylesheetHtml.Insert(0, importBuilder.ToString());
-            stylesheetHtml =
-                Parser.Parser.CssImportUrlRegex.Replace(stylesheetHtml, "$1" + apiBase + prepend + "$2" + "$3");
+            var importMatches = Parser.Parser.CssImportUrlRegex.Matches(stylesheetHtml);
+            foreach (Match match in importMatches)
+            {
+                if (!match.Success) continue;
+                var importFile = match.Groups["Filename"].Value;
+                stylesheetHtml = stylesheetHtml.Replace(importFile, apiBase + prepend + importFile);
+            }
+
+            // Check if there are any background images and rewrite those urls
+            EscapeCssImageReferences(ref stylesheetHtml, apiBase, book);
 
             var styleContent = RemoveWhiteSpaceFromStylesheets(stylesheetHtml);
-            styleContent =
-                Parser.Parser.FontSrcUrlRegex.Replace(styleContent, "$1" + apiBase + "$2" + "$3");
-
             styleContent = styleContent.Replace("body", ".reading-section");
 
             var stylesheet = await _cssParser.ParseAsync(styleContent);
@@ -165,6 +170,21 @@ namespace API.Services
             return RemoveWhiteSpaceFromStylesheets(stylesheet.ToCss());
         }
 
+        private static void EscapeCssImageReferences(ref string stylesheetHtml, string apiBase, EpubBookRef book)
+        {
+            var matches = Parser.Parser.CssImageUrlRegex.Matches(stylesheetHtml);
+            foreach (Match match in matches)
+            {
+                if (!match.Success) continue;
+
+                var importFile = match.Groups["Filename"].Value;
+                var key = CleanContentKeys(importFile);
+                if (!book.Content.AllFiles.ContainsKey(key)) continue;
+
+                stylesheetHtml = stylesheetHtml.Replace(importFile, apiBase + key);
+            }
+        }
+
         public ComicInfo GetComicInfo(string filePath)
         {
             if (!IsValidFile(filePath) || Parser.Parser.IsPdf(filePath)) return null;