Merge pull request #367 from manikmagar/image-src-path-fix

Prevent image source path from breaking

build.gradle

@@ -46,6 +46,7 @@ ext {
     jsonSimpleVersion = '1.1.1'
     jade4jVersion = '1.2.5'
     mockitoVersion = '1.10.19'
+    jsoupVersion = '1.10.2'
 }
 
 dependencies {
@@ -68,6 +69,7 @@ dependencies {
     compile group: 'ch.qos.logback', name: 'logback-classic', version: logbackVersion
     compile group: 'ch.qos.logback', name: 'logback-core', version: logbackVersion
     compile group: 'de.neuland-bfi', name: 'jade4j', version: jade4jVersion
+	compile group: 'org.jsoup', name:'jsoup', version: jsoupVersion
     testCompile group: 'junit', name: 'junit', version: junitVersion
     testCompile group: 'org.assertj', name: 'assertj-core', version: assertjCoreVersion
     testCompile group: 'org.mockito', name: 'mockito-core', version: mockitoVersion

src/main/java/org/jbake/app/Crawler.java

@@ -1,23 +1,23 @@
 package org.jbake.app;
 
 
-import com.orientechnologies.orient.core.record.impl.ODocument;
+import java.io.File;
+import java.util.Arrays;
+import java.util.Date;
+import java.util.Map;
+
 import org.apache.commons.configuration.CompositeConfiguration;
 import org.apache.commons.io.FilenameUtils;
 import org.jbake.app.ConfigUtil.Keys;
 import org.jbake.app.Crawler.Attributes.Status;
 import org.jbake.model.DocumentAttributes;
 import org.jbake.model.DocumentStatus;
 import org.jbake.model.DocumentTypes;
+import org.jbake.util.HtmlUtil;
 import org.slf4j.Logger;
 import org.slf4j.LoggerFactory;
 
-import java.io.File;
-import java.util.Arrays;
-import java.util.Date;
-import java.util.Map;
-
-import static java.io.File.separator;
+import com.orientechnologies.orient.core.record.impl.ODocument;
 
 /**
  * Crawls a file system looking for content.
@@ -179,6 +179,10 @@ private void crawlSourceFile(final File sourceFile, final String sha1, final Str
             if (config.getBoolean(Keys.URI_NO_EXTENSION)) {
             	fileContents.put(Attributes.NO_EXTENSION_URI, uri.replace("/index.html", "/"));
             }
+
+
+            // Prevent image source url's from breaking
+            HtmlUtil.fixImageSourceUrls(fileContents,config);
 
             ODocument doc = new ODocument(documentType);
             doc.fields(fileContents);

src/main/java/org/jbake/util/HtmlUtil.java

@@ -0,0 +1,82 @@
+package org.jbake.util;
+
+import java.util.Map;
+
+import org.apache.commons.configuration.CompositeConfiguration;
+import org.jbake.app.ConfigUtil.Keys;
+import org.jbake.app.Crawler.Attributes;
+import org.jsoup.Jsoup;
+import org.jsoup.nodes.Document;
+import org.jsoup.nodes.Element;
+import org.jsoup.select.Elements;
+
+/**
+ * 
+ * @author Manik Magar
+ *
+ */
+public class HtmlUtil {
+
+	/**
+	 * Image paths are specified as w.r.t. assets folder. This function prefix site host to all img src except 
+	 * the ones that starts with http://, https://.
+	 * 
+	 * If image path starts with "./", i.e. relative to the source file, then it first replace that with output file directory and the add site host.
+	 * 
+	 * @param fileContents
+	 */
+    public static void fixImageSourceUrls(Map<String, Object> fileContents, CompositeConfiguration config){
+
+    	String htmlContent = fileContents.get(Attributes.BODY).toString();
+
+    	String siteHost = config.getString(Keys.SITE_HOST);
+
+    	String uri = fileContents.get(Attributes.URI).toString();
+
+    	if(fileContents.get(Attributes.NO_EXTENSION_URI) != null){
+    		uri = fileContents.get(Attributes.NO_EXTENSION_URI).toString();
+
+    		//remove trailing "/"
+    		if(uri.endsWith("/")) {
+    			uri = uri.substring(0, uri.length() - 1);
+    		}
+
+    	}
+
+    	if(uri.contains("/")){
+        	//strip that file name, leaving end "/"
+        		uri = uri.substring(0, uri.lastIndexOf("/") + 1);
+        }
+
+    	Document document = Jsoup.parseBodyFragment(htmlContent);
+
+    	Elements allImgs = document.getElementsByTag("img");
+
+    	for (Element img : allImgs) {
+			String source = img.attr("src");
+
+			if(source.startsWith("./")){
+				// image relative to current content is specified,
+				// lets add current url to it.
+				source = source.replaceFirst("./", uri);
+			}
+
+			// Now add the root path
+			if(!source.startsWith("http://") 
+					&& !source.startsWith("https://")){
+
+				if (!siteHost.endsWith("/") && !source.startsWith("/")) siteHost = siteHost.concat("/");
+
+				String fullUrl = siteHost + source;
+
+				img.attr("src", fullUrl);
+
+			}
+		}
+
+
+    	//Use body().html() to prevent adding <body></body> from parsed fragment.
+    	fileContents.put(Attributes.BODY, document.body().html());
+    }
+
+}

src/test/java/org/jbake/util/HtmlUtilTest.java

@@ -0,0 +1,149 @@
+package org.jbake.util;
+
+import static org.assertj.core.api.Assertions.*;
+
+import java.io.File;
+import java.util.HashMap;
+import java.util.Map;
+
+import org.apache.commons.configuration.CompositeConfiguration;
+import org.jbake.app.ConfigUtil;
+import org.jbake.app.Crawler.Attributes;
+import org.jbake.util.HtmlUtil;
+import org.junit.Before;
+import org.junit.Test;
+
+
+
+public class HtmlUtilTest {
+
+	private CompositeConfiguration config;
+
+	@Before
+	public void setUp() throws Exception{
+		config = ConfigUtil.load(new File(this.getClass().getResource("/fixture").getFile()));
+	}
+
+	@Test
+	public void shouldNotAddBodyHTMLElement(){
+		Map<String, Object> fileContent = new HashMap<String, Object>();
+		fileContent.put(Attributes.ROOTPATH, "../../../");
+		fileContent.put(Attributes.URI, "blog/2017/05/first_post.html");
+		fileContent.put(Attributes.BODY, "<div> Test <img src='blog/2017/05/first.jpg' /></div>");
+
+		HtmlUtil.fixImageSourceUrls(fileContent, config);
+
+		String body = fileContent.get(Attributes.BODY).toString();
+
+		assertThat(body).doesNotContain("<body>");
+		assertThat(body).doesNotContain("</body>");
+
+	}
+
+	@Test
+	public void shouldAddRootpath(){
+		Map<String, Object> fileContent = new HashMap<String, Object>();
+		fileContent.put(Attributes.ROOTPATH, "../../../");
+		fileContent.put(Attributes.URI, "blog/2017/05/first_post.html");
+		fileContent.put(Attributes.BODY, "<div> Test <img src='blog/2017/05/first.jpg' /></div>");
+
+		HtmlUtil.fixImageSourceUrls(fileContent, config);
+
+		String body = fileContent.get(Attributes.BODY).toString();
+
+		assertThat(body).contains("src=\"http://www.jbake.org/blog/2017/05/first.jpg\"");
+
+	}
+
+	@Test
+	public void shouldAddContentpath(){
+		Map<String, Object> fileContent = new HashMap<String, Object>();
+		fileContent.put(Attributes.ROOTPATH, "../../../");
+		fileContent.put(Attributes.URI, "blog/2017/05/first_post.html");
+		fileContent.put(Attributes.BODY, "<div> Test <img src='./first.jpg' /></div>");
+
+		HtmlUtil.fixImageSourceUrls(fileContent, config);
+
+		String body = fileContent.get(Attributes.BODY).toString();
+
+		assertThat(body).contains("src=\"http://www.jbake.org/blog/2017/05/first.jpg\"");
+
+	}
+
+	@Test
+	public void shouldNotAddRootPath(){
+		Map<String, Object> fileContent = new HashMap<String, Object>();
+		fileContent.put(Attributes.ROOTPATH, "../../../");
+		fileContent.put(Attributes.URI, "blog/2017/05/first_post.html");
+		fileContent.put(Attributes.BODY, "<div> Test <img src='/blog/2017/05/first.jpg' /></div>");
+
+		HtmlUtil.fixImageSourceUrls(fileContent,config);
+
+		String body = fileContent.get(Attributes.BODY).toString();
+
+		assertThat(body).contains("src=\"http://www.jbake.org/blog/2017/05/first.jpg\"");
+
+	}
+
+	@Test
+	public void shouldAddRootPathForNoExtension(){
+		Map<String, Object> fileContent = new HashMap<String, Object>();
+		fileContent.put(Attributes.ROOTPATH, "../../../");
+		fileContent.put(Attributes.URI, "blog/2017/05/first_post.html");
+		fileContent.put(Attributes.NO_EXTENSION_URI, "blog/2017/05/first_post/");
+		fileContent.put(Attributes.BODY, "<div> Test <img src='blog/2017/05/first.jpg' /></div>");
+
+		HtmlUtil.fixImageSourceUrls(fileContent, config);
+
+		String body = fileContent.get(Attributes.BODY).toString();
+
+		assertThat(body).contains("src=\"http://www.jbake.org/blog/2017/05/first.jpg\"");
+
+	}
+
+	@Test
+	public void shouldAddContentPathForNoExtension(){
+		Map<String, Object> fileContent = new HashMap<String, Object>();
+		fileContent.put(Attributes.ROOTPATH, "../../../");
+		fileContent.put(Attributes.URI, "blog/2017/05/first_post.html");
+		fileContent.put(Attributes.NO_EXTENSION_URI, "blog/2017/05/first_post/");
+		fileContent.put(Attributes.BODY, "<div> Test <img src='./first.jpg' /></div>");
+
+		HtmlUtil.fixImageSourceUrls(fileContent, config);
+
+		String body = fileContent.get(Attributes.BODY).toString();
+
+		assertThat(body).contains("src=\"http://www.jbake.org/blog/2017/05/first.jpg\"");
+	}
+
+	@Test
+	public void shouldNotChangeForHTTP(){
+		Map<String, Object> fileContent = new HashMap<String, Object>();
+		fileContent.put(Attributes.ROOTPATH, "../../../");
+		fileContent.put(Attributes.URI, "blog/2017/05/first_post.html");
+		fileContent.put(Attributes.NO_EXTENSION_URI, "blog/2017/05/first_post/");
+		fileContent.put(Attributes.BODY, "<div> Test <img src='http://example.com/first.jpg' /></div>");
+
+		HtmlUtil.fixImageSourceUrls(fileContent, config);
+
+		String body = fileContent.get(Attributes.BODY).toString();
+
+		assertThat(body).contains("src=\"http://example.com/first.jpg\"");
+
+	}
+
+	@Test
+	public void shouldNotChangeForHTTPS(){
+		Map<String, Object> fileContent = new HashMap<String, Object>();
+		fileContent.put(Attributes.ROOTPATH, "../../../");
+		fileContent.put(Attributes.URI, "blog/2017/05/first_post.html");
+		fileContent.put(Attributes.NO_EXTENSION_URI, "blog/2017/05/first_post/");
+		fileContent.put(Attributes.BODY, "<div> Test <img src='https://example.com/first.jpg' /></div>");
+
+		HtmlUtil.fixImageSourceUrls(fileContent, config);
+
+		String body = fileContent.get(Attributes.BODY).toString();
+
+		assertThat(body).contains("src=\"https://example.com/first.jpg\"");
+	}
+}