Fix for keeping lists structure

javatests/org/chromium/distiller/webdocument/filters/RelevantElementsTest.java

 // Copyright 2015 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 package org.chromium.distiller.webdocument.filters;
 
 import org.chromium.distiller.DomDistillerJsTestCase;
-import org.chromium.distiller.webdocument.TestWebTextBuilder;
+import org.chromium.distiller.webdocument.PlaceHolder;
 import org.chromium.distiller.webdocument.TestWebDocumentBuilder;
 import org.chromium.distiller.webdocument.WebDocument;
 import org.chromium.distiller.webdocument.WebElement;
 import org.chromium.distiller.webdocument.WebImage;
-import org.chromium.distiller.webdocument.WebText;
 import org.chromium.distiller.webdocument.WebTable;
 
 
 public class RelevantElementsTest extends DomDistillerJsTestCase {
     public void testEmptyDocument() {
         WebDocument document = new WebDocument();
         assertFalse(RelevantElements.process(document));
         assertTrue(document.getElements().isEmpty());
     }
 
@@ skipping 48 matching lines @@
 
     public void testImageAfterNonContent() {
         TestWebDocumentBuilder builder = new TestWebDocumentBuilder();
         builder.addText("text 1").setIsContent(true);
         builder.addText("text 2").setIsContent(false);
         WebImage wi = builder.addImage();
         WebDocument document = builder.build();
         assertFalse(RelevantElements.process(document));
         assertFalse(wi.getIsContent());
     }
+
+    public void testRelevantOrderedList() {
+        TestWebDocumentBuilder builder = new TestWebDocumentBuilder();
+        PlaceHolder placeHolder = builder.addListStart();

[wychen, 2015/08/01 01:00:20]

nitpick: rename placeHolder to olStart?

+        PlaceHolder liStart = builder.addListStart();
+        builder.addText("text 1").setIsContent(true);
+        PlaceHolder liEnd = builder.addListEnd();
+        PlaceHolder liStart2 = builder.addListStart();
+        builder.addText("text 2").setIsContent(true);
+        PlaceHolder liEnd2 = builder.addListEnd();
+        PlaceHolder liStart3 = builder.addListStart();
+        builder.addText("text 3").setIsContent(true);
+        PlaceHolder liEnd3 = builder.addListEnd();
+        PlaceHolder olEnd = builder.addListEnd();
+        WebDocument document = builder.build();
+        assertTrue(RelevantElements.process(document));
+        assertTrue(placeHolder.getIsContent());
+        assertTrue(liStart.getIsContent());
+        assertTrue(liEnd.getIsContent());
+        assertTrue(liStart2.getIsContent());
+        assertTrue(liEnd2.getIsContent());
+        assertTrue(liStart3.getIsContent());
+        assertTrue(liEnd3.getIsContent());
+        assertTrue(olEnd.getIsContent());
+    }
+
+    public void testRelevantUnorderedList() {
+        TestWebDocumentBuilder builder = new TestWebDocumentBuilder();
+        PlaceHolder ulStart = builder.addListStart();
+        PlaceHolder liStart = builder.addListStart();
+        builder.addText("text 1").setIsContent(true);
+        PlaceHolder liEnd = builder.addListEnd();
+        PlaceHolder liStart2 = builder.addListStart();
+        builder.addText("text 2").setIsContent(true);

[wychen, 2015/08/01 01:00:20]

We should make some items non-content.

+        PlaceHolder liEnd2 = builder.addListEnd();
+        PlaceHolder liStart3 = builder.addListStart();
+        builder.addText("text 3").setIsContent(true);
+        PlaceHolder liEnd3 = builder.addListEnd();
+        PlaceHolder ulEnd = builder.addListEnd();
+        WebDocument document = builder.build();
+        assertTrue(RelevantElements.process(document));
+        assertTrue(ulStart.getIsContent());
+        assertTrue(liStart.getIsContent());
+        assertTrue(liEnd.getIsContent());
+        assertTrue(liStart2.getIsContent());
+        assertTrue(liEnd2.getIsContent());
+        assertTrue(liStart3.getIsContent());
+        assertTrue(liEnd3.getIsContent());
+        assertTrue(ulEnd.getIsContent());
+    }
+
+    public void testRelevantNestedList() {

[wychen, 2015/08/01 01:00:20]

We should make some of the items non-content, and complicated enough to test
handlePlaceHolderElements().

+        TestWebDocumentBuilder builder = new TestWebDocumentBuilder();
+        PlaceHolder ulStart = builder.addListStart();
+        PlaceHolder liStart = builder.addListStart();
+        builder.addText("text 1").setIsContent(true);
+        PlaceHolder liEnd = builder.addListEnd();
+        PlaceHolder olStart = builder.addListStart();
+        PlaceHolder liOLStart = builder.addListStart();
+        builder.addText("text 2").setIsContent(true);
+        PlaceHolder liOLEnd = builder.addListEnd();
+        PlaceHolder olEnd = builder.addListEnd();
+        PlaceHolder ulEnd = builder.addListEnd();
+        WebDocument document = builder.build();
+        assertTrue(RelevantElements.process(document));
+        assertTrue(ulStart.getIsContent());
+        assertTrue(liStart.getIsContent());
+        assertTrue(olStart.getIsContent());
+        assertTrue(liOLStart.getIsContent());
+        assertTrue(liOLEnd.getIsContent());
+        assertTrue(olEnd.getIsContent());
+        assertTrue(liEnd.getIsContent());
+        assertTrue(ulEnd.getIsContent());
+    }
 }