feat: urls in css now replaced by the base64 content (#215)

* feat: urls in css now replaced by the base64 content

Refs: SchweizerischeBundesbahnen/weasyprint-service#61

pom.xml

@@ -55,6 +55,7 @@
         <exec-maven-plugin.version>3.3.0</exec-maven-plugin.version>
         <weasyprint.version>62.3</weasyprint.version>
         <awaitility.version>4.2.2</awaitility.version>
+        <tika.version>1.28.5</tika.version>
 
         <maven-jar-plugin.Extension-Context>pdf-exporter</maven-jar-plugin.Extension-Context>
         <maven-jar-plugin.Automatic-Module-Name>ch.sbb.polarion.extension.pdf_exporter</maven-jar-plugin.Automatic-Module-Name>
@@ -94,6 +95,13 @@
             <scope>provided</scope>
         </dependency>
 
+        <dependency>
+            <groupId>org.apache.tika</groupId>
+            <artifactId>tika-core</artifactId>
+            <version>${tika.version}</version>
+            <scope>provided</scope>
+        </dependency>
+
         <!--  3rd-party libraries to be packaged into resulting jar  -->
         <dependency>
             <groupId>org.apache.pdfbox</groupId>

src/main/java/ch/sbb/polarion/extension/pdf_exporter/converter/HtmlToPdfConverter.java

@@ -74,7 +74,7 @@ String preprocessHtml(String origHtml, Orientation orientation, PaperSize paperS
         } else {
             html = replaceTagContent(origHtml, "head", head);
         }
-        html = htmlProcessor.replaceImagesAsBase64Encoded(html);
+        html = htmlProcessor.replaceResourcesAsBase64Encoded(html);
         html = htmlProcessor.internalizeLinks(html);
 
         return html;

src/main/java/ch/sbb/polarion/extension/pdf_exporter/converter/PdfConverter.java

@@ -284,7 +284,7 @@ String getCssContent(
         String processed = velocityEvaluator.evaluateVelocityExpressions(documentData, content);
 
         String cssContent = (exportParams.getDocumentType() != DocumentType.LIVE_DOC) ? appendWikiCss(processed) : processed;
-        return htmlProcessor.replaceImagesAsBase64Encoded(cssContent);
+        return htmlProcessor.replaceResourcesAsBase64Encoded(cssContent);
     }
 
     @VisibleForTesting
@@ -310,7 +310,7 @@ String getHeaderFooterContent(
                 .toList();
 
         String headerFooterContent = String.format(ScopeUtils.getFileContent("webapp/pdf-exporter/html/headerAndFooter.html"), nonNullHeaderFooterContents.toArray());
-        return htmlProcessor.replaceImagesAsBase64Encoded(headerFooterContent);
+        return htmlProcessor.replaceResourcesAsBase64Encoded(headerFooterContent);
     }
 
     private String appendWikiCss(String css) {

src/main/java/ch/sbb/polarion/extension/pdf_exporter/util/CustomResourceUrlResolver.java

@@ -0,0 +1,59 @@
+package ch.sbb.polarion.extension.pdf_exporter.util;
+
+import com.polarion.alm.tracker.internal.url.IUrlResolver;
+import com.polarion.core.util.logging.Logger;
+import lombok.SneakyThrows;
+import org.jetbrains.annotations.NotNull;
+import org.jetbrains.annotations.VisibleForTesting;
+
+import java.io.InputStream;
+import java.net.HttpURLConnection;
+import java.net.URL;
+
+import static java.net.HttpURLConnection.*;
+
+/**
+ * Custom version of {@link com.polarion.alm.tracker.internal.url.GenericUrlResolver} with the redirect support.
+ */
+public class CustomResourceUrlResolver implements IUrlResolver {
+
+    private static final int CONNECTION_TIMEOUT_MS = 3_000;
+    private static final int READ_TIMEOUT_MS = 3_000;
+
+    public boolean canResolve(@NotNull String url) {
+        return url.startsWith("http://") || url.startsWith("https://");
+    }
+
+    public InputStream resolve(@NotNull String urlStr) {
+        try {
+            URL url = new URL(normalizeUrl(urlStr));
+            return resolveImpl(url);
+        } catch (Exception e) {
+            Logger.getLogger(this).warn("Failed to load resource: " + urlStr, e);
+        }
+        return null;
+    }
+
+    @SneakyThrows
+    @VisibleForTesting
+    public InputStream resolveImpl(@NotNull URL url) {
+        HttpURLConnection connection = (HttpURLConnection) url.openConnection();
+        connection.setConnectTimeout(CONNECTION_TIMEOUT_MS);
+        connection.setReadTimeout(READ_TIMEOUT_MS);
+        connection.connect();
+        int responseCode = connection.getResponseCode();
+        if (responseCode == HTTP_OK) {
+            return connection.getInputStream();
+        } else if (responseCode == HTTP_MOVED_PERM || responseCode == HTTP_MOVED_TEMP) {
+            String location = connection.getHeaderField("Location");
+            if (location != null && canResolve(location)) {
+                return resolve(location);
+            }
+        }
+        return null;
+    }
+
+    private String normalizeUrl(String urlStr) {
+        return urlStr.replace(" ", "%20").replace("%5F", "_");
+    }
+}

src/main/java/ch/sbb/polarion/extension/pdf_exporter/util/FileResourceProvider.java

@@ -1,7 +1,13 @@
 package ch.sbb.polarion.extension.pdf_exporter.util;
 
+import org.jetbrains.annotations.NotNull;
+import org.jetbrains.annotations.Nullable;
+
 public interface FileResourceProvider {
 
-    byte[] getResourceAsBytes(String resource);
+    @Nullable
+    String getResourceAsBase64String(@NotNull String resource);
+
+    byte[] getResourceAsBytes(@NotNull String resource);
 
 }

src/main/java/ch/sbb/polarion/extension/pdf_exporter/util/HtmlProcessor.java

@@ -19,13 +19,7 @@
 import org.jetbrains.annotations.Nullable;
 import org.jetbrains.annotations.VisibleForTesting;
 
-import java.io.BufferedInputStream;
-import java.io.ByteArrayInputStream;
-import java.io.InputStream;
-import java.net.URLConnection;
-import java.nio.charset.StandardCharsets;
 import java.util.Arrays;
-import java.util.Base64;
 import java.util.HashMap;
 import java.util.LinkedHashMap;
 import java.util.LinkedList;
@@ -192,8 +186,8 @@ public String processHtmlForPDF(@NotNull String html, @NotNull ExportParams expo
                 yield removeFloatLeftFromReports(processingHtml);
             }
         };
-        html = replaceImagesAsBase64Encoded(html);
-        html = removeSvgUnsupportedFeatureHint(html); //note that there is one more replacement attempt before replacing images with base64 representation
+        html = replaceResourcesAsBase64Encoded(html);
+        html = MediaUtils.removeSvgUnsupportedFeatureHint(html); //note that there is one more replacement attempt before replacing images with base64 representation
         html = properTableHeads(html);
         html = cleanExtraTableContent(html);
         html = switch (exportParams.getDocumentType()) {
@@ -980,58 +974,14 @@ private int findTableEnd(String html, int tableStart) {
 
     @SneakyThrows
     @SuppressWarnings({"java:S5852", "java:S5857"}) //need by design
-    public String replaceImagesAsBase64Encoded(String html) {
-        StringBuilder result = new StringBuilder();
-
-        //Retrieves data of 'src' attribute from 'img' tags
-        IRegexEngine imageRegexEngine = RegexMatcher.get("<img[^<>]*src=(\"|')(?<url>[^(\"|')]*)(\"|')").createEngine(html);
-        while (imageRegexEngine.find()) {
-            String url = imageRegexEngine.group("url");
-            if (url.startsWith("data:")) {
-                continue;
-            }
-            byte[] imgBytes = fileResourceProvider.getResourceAsBytes(url.replace("%5F", "_")); // Replace encoded underscore symbol in 'src' attribute of images
-            if (imgBytes != null && imgBytes.length != 0) { // Don't make any manipulations if image wasn't resolved
-                try (InputStream is = new BufferedInputStream(new ByteArrayInputStream(imgBytes))) {
-                    String mimeType = URLConnection.guessContentTypeFromStream(is);
-
-                    // looks like sometimes mime type for svg isn't recognized
-                    if (url.contains(".svg") && (mimeType == null || mimeType.equals(MIME_TYPE_SVG))) {
-                        imgBytes = processPossibleSvgImage(imgBytes);
-                    }
-
-                    String encodedImage = String.format("data:%s;base64, %s", mimeType, Base64.getEncoder().encodeToString(imgBytes));
-
-                    imageRegexEngine.appendReplacement(result, imageRegexEngine.group().replace(url, encodedImage));
-                }
-            }
-        }
-        imageRegexEngine.appendTail(result);
-
-        return result.toString();
+    public String replaceResourcesAsBase64Encoded(String html) {
+        return MediaUtils.inlineBase64Resources(html, fileResourceProvider);
     }
 
     public String internalizeLinks(String html) {
         return httpLinksHelper.internalizeLinks(html);
     }
 
-    @VisibleForTesting
-    @SuppressWarnings("squid:S1166") // no need to log or rethrow exception by design
-    public byte[] processPossibleSvgImage(byte[] possibleSvgImageBytes) {
-        try {
-            String svgContent = new String(possibleSvgImageBytes, StandardCharsets.UTF_8);
-            return removeSvgUnsupportedFeatureHint(svgContent).getBytes(StandardCharsets.UTF_8);
-        } catch (Exception e) {
-            // not a valid string, just nvm
-        }
-        return possibleSvgImageBytes;
-    }
-
-    @VisibleForTesting
-    public String removeSvgUnsupportedFeatureHint(String html) {
-        return html.replaceAll("(?s)<switch>[^<]*?<g requiredFeatures=\"[^\"]+?\"/>.*?</switch>", "");
-    }
-
     private boolean hasCustomPageBreaks(String html) {
         return html.contains(PAGE_BREAK_MARK);
     }