OLD | NEW |
---|---|
(Empty) | |
1 package org.chromium.distiller.webdocument.filters; | |
2 | |
3 import org.chromium.distiller.webdocument.WebDocument; | |
4 import org.chromium.distiller.webdocument.WebElement; | |
5 import org.chromium.distiller.webdocument.WebTag; | |
6 import org.chromium.distiller.webdocument.WebText; | |
7 | |
8 import java.util.Stack; | |
9 | |
10 public class WebTagStructureKeeper { | |
mdjones
2015/08/03 23:29:45
How about NestedElement{Builder|Organizer|Retainer
| |
11 public static void process(WebDocument document) { | |
12 boolean isContent = false; | |
13 int stackMark = -1; | |
14 Stack<WebTag> stack = new Stack<>(); | |
15 | |
16 for (WebElement e : document.getElements()) { | |
17 if (e instanceof WebText) { | |
mdjones
2015/08/03 23:29:45
Though I'm not sure it is a common case, this does
| |
18 if (!isContent) { | |
19 isContent = e.getIsContent(); | |
20 } | |
21 } else if (e instanceof WebTag) { | |
22 WebTag webTag = (WebTag) e; | |
23 if (webTag.isStartTag()) { | |
24 webTag.setIsContent(isContent); | |
25 stack.push(webTag); | |
26 isContent = false; | |
27 } else { | |
28 WebTag startWebTag = stack.pop(); | |
29 boolean content = isContent || stackMark >= stack.size(); | |
mdjones
2015/08/03 23:29:45
isContent |= stackMark >= stackSize();
Then just
| |
30 if (content) { | |
31 stackMark = stack.size() - 1; | |
32 } | |
33 startWebTag.setIsContent(content); | |
34 webTag.setIsContent(content); | |
35 isContent = startWebTag.getIsContent(); | |
wychen
2015/08/04 02:37:01
Does this pass the test? Moving this line 2 lines
| |
36 } | |
37 } | |
38 } | |
39 } | |
40 } | |
OLD | NEW |