Use stricter comment-detecting heuristics

java/org/chromium/distiller/webdocument/ElementAction.java

 // Copyright 2014 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;
 
 import org.chromium.distiller.DomUtil;
 import org.chromium.distiller.labels.DefaultLabels;
 
 import com.google.gwt.core.client.JavaScriptObject;
 import com.google.gwt.core.client.JsArrayString;
 import com.google.gwt.dom.client.Element;
 import com.google.gwt.dom.client.Style;
 import com.google.gwt.regexp.shared.RegExp;
 
 public class ElementAction {
     public boolean changesTagLevel = false;
     public boolean flush = false;
     public boolean isAnchor = false;
     public JsArrayString labels = JavaScriptObject.createArray().<JsArrayString>cast();
 
     private static final RegExp REG_COMMENT = RegExp.compile("\\bcomments?\\b");
-    private static final int MAX_CLASS_COUNT = 5;
+    private static final int MAX_CLASS_COUNT = 2;

[mdjones, 2017/01/04 14:29:21]

nit: the way this is used seems to be as a minimum number, can we change this to
MIN_CLASS_COUNT?

[wychen, 2017/01/04 14:37:51]

What do you mean?
This constant is the largest allowed number of class names for this heuristic to
be enabled, so sounds like max more than min, right?

[mdjones, 2017/01/04 15:14:33]

Ah, I was looking at it from: "if class count doesn't meet this number, it's not
content".

 
     public static ElementAction getForElement(Element element) {
         Style style = DomUtil.getComputedStyle(element);
         ElementAction action = new ElementAction();
         String tagName = element.getTagName();
         switch (style.getDisplay()) {
             case "inline":
                 break;
             case "inline-block":
             case "inline-flex":
@@ skipping 21 matching lines @@
             // table-column-group
             // table-cell
             // table-caption
             // flex
             default:
                 action.flush = true;
                 action.changesTagLevel = true;
                 break;
         }
 
-        if (!"HTML".equals(tagName) && !"BODY".equals(tagName)) {
+        if (!"HTML".equals(tagName) && !"BODY".equals(tagName) && !"ARTICLE".equals(tagName)) {
             String className = element.getAttribute("class");
             int classCount = DomUtil.getClassList(element).length();
             String id = element.getAttribute("id");
             if ((REG_COMMENT.test(className) || REG_COMMENT.test(id)) &&
                     classCount <= MAX_CLASS_COUNT) {
                 action.labels.push(DefaultLabels.STRICTLY_NOT_CONTENT);
             }
 
             switch (tagName) {
                 case "ASIDE":
@@ skipping 28 matching lines @@
                         action.isAnchor = true;
                     }
                     break;
             }
         }
         return action;
     }
 
     private ElementAction() {}
 }