Support <picture> in image extraction

java/org/chromium/distiller/extractors/embeds/ImageExtractor.java

Index: java/org/chromium/distiller/extractors/embeds/ImageExtractor.java
diff --git a/java/org/chromium/distiller/extractors/embeds/ImageExtractor.java b/java/org/chromium/distiller/extractors/embeds/ImageExtractor.java
index 0064b9579eb32a3cf70f0907238d9ca1e18cee35..e1ad898b91f00955eb1cea48451233b4dcc50eaa 100644
--- a/java/org/chromium/distiller/extractors/embeds/ImageExtractor.java
+++ b/java/org/chromium/distiller/extractors/embeds/ImageExtractor.java
@@ -29,6 +29,7 @@ public class ImageExtractor implements EmbedExtractor {
     static {
         // TODO(mdjones): Add "DIV" to this list for css images and possibly captions.
         relevantTags.add("IMG");
+        relevantTags.add("PICTURE");
         relevantTags.add("FIGURE");
     }
 
@@ -50,10 +51,17 @@ public class ImageExtractor implements EmbedExtractor {
         if ("IMG".equals(e.getTagName())) {
             extractImageAttributes(ImageElement.as(e));
             return new WebImage(e, width, height, imgSrc);
+        } else if ("PICTURE".equals(e.getTagName())) {

[mdjones, 2017/01/17 17:41:46]

How would you feel about having some sort of utility function in this class
called "getImageElement"? That way we could eliminate the duplicate
"img"/"picture" tag checks here and below. It's not that hard to read now, but
adding another tag (maybe div) would further complicate this. wdyt? 

[wychen, 2017/01/18 17:29:54]

Done.

+                return new WebImage(e, width, height, imgSrc);

[mdjones, 2017/01/17 17:41:47]

nit: indented too far.

[wychen, 2017/01/18 17:29:54]

Done.

         } else if ("FIGURE".equals(e.getTagName())) {
-            Element img = getFirstElementByTagName(e, "IMG");
+            Element img = getFirstElementByTagName(e, "PICTURE");
+            if (img == null) {
+                img = getFirstElementByTagName(e, "IMG");
+            }
             if (img != null) {

[mdjones, 2017/01/17 17:41:46]

nit: early return instead of nesting?

[wychen, 2017/01/18 17:29:54]

Done.

-                extractImageAttributes(ImageElement.as(img));
+                if ("IMG".equals(img.getTagName())) {
+                    extractImageAttributes(ImageElement.as(img));
+                }
                 Element figcaption;
                 Element cap = getFirstElementByTagName(e, "FIGCAPTION");
                 if (cap != null) {