recognize and parse Schema.org Markup

src/com/dom_distiller/client/SchemaOrgParserAccessor.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 com.dom_distiller.client;
+
+import java.util.ArrayList;
+import java.util.List;
+
+import com.google.gwt.dom.client.Element;
+
+/**
+ * This class instantiates SchemaOrgParser and implements MarkupParser.Parser interface to provide
+ * access to properties that SchemaOrgParser has parsed.
+ */
+public class SchemaOrgParserAccessor implements MarkupParser.Parser {
+    private final SchemaOrgParser parser;
+
+    /**
+     * The object that instantiates SchemaOrgParser and implements its MarkupParser.Parser
+     * interface.
+     */
+    public SchemaOrgParserAccessor(Element root) {
+        parser = new SchemaOrgParser(root);
+    }
+
+    @Override
+    public String getTitle() {
+        String title = "";
+        List<SchemaOrgParser.ThingItem> itemScopes = parser.getItemScopes();
+
+        // Get the "headline" property of the first article that has it.
+        for (int i = 0; i < itemScopes.size() && title.isEmpty(); i++) {
+            SchemaOrgParser.ThingItem item = itemScopes.get(i);
+            if (item.getType() == SchemaOrgParser.Type.ARTICLE) {

[cjhopman, 2014/04/29 17:04:19]

Iterating through the articles seems pretty common. We could add a
getArticleItems() to the parser (then we could do the cast in there instead).
You could probably drop the findFirstArticle then and just get the first item in
the list returned by getArticleItems.

Probably could do the same for images even though I think we only iterate
through those once.

[kuan, 2014/04/29 23:26:43]

Done.

+                title = item.getStringProperty(SchemaOrgParser.HEADLINE_PROP);
+            }
+        }
+
+        // If there's no "headline" property, use "name" property.
+        for (int i = 0; i < itemScopes.size() && title.isEmpty(); i++) {
+            SchemaOrgParser.ThingItem item = itemScopes.get(i);
+            if (item.getType() == SchemaOrgParser.Type.ARTICLE) {
+                title = item.getStringProperty(SchemaOrgParser.NAME_PROP);
+            }
+        }
+
+        return title;
+    }
+
+    @Override
+    public String getType() {
+        // TODO(kuan): consolidate/standardize types returned from all 3 parsers in MarkupParser.
+        // Returns ARTICLe if there's an article.
+        return parser.findFirstArticle() != null ? SchemaOrgParser.Type.ARTICLE.toString() : "";
+    }
+
+    @Override
+    public String getUrl() {
+        SchemaOrgParser.ArticleItem item = parser.findFirstArticle();
+        return item != null ? item.getStringProperty(SchemaOrgParser.URL_PROP) : "";
+    }
+
+    @Override
+    public MarkupParser.Image[] getImages() {
+        List<SchemaOrgParser.ThingItem> itemScopes = parser.getItemScopes();
+        if (itemScopes.isEmpty()) return null;
+
+        List<MarkupParser.Image> images = new ArrayList<MarkupParser.Image>();
+        boolean hasRepresentativeImage = false;
+        MarkupParser.Image imageOfArticle = null;
+        SchemaOrgParser.ImageItem associatedImageOfArticle = null;
+
+        for (int i = 0; i < itemScopes.size(); i++) {
+            SchemaOrgParser.ThingItem item = itemScopes.get(i);
+            MarkupParser.Image image = null;
+            if (item.getType() == SchemaOrgParser.Type.ARTICLE) {
+                SchemaOrgParser.ArticleItem articleItem = (SchemaOrgParser.ArticleItem) item;
+                // If article has an associated image or the "image" property, remember them for
+                // now; they'll be added to the list later when the position in the list can be
+                // determined.
+                if (associatedImageOfArticle == null) {
+                    associatedImageOfArticle = articleItem.getRepresentativeImageItem();
+                    if (associatedImageOfArticle != null) continue;
+                }
+                image = articleItem.getImage();
+                if (image == null) continue;
+                if (imageOfArticle == null) {
+                    imageOfArticle = image;
+                } else {
+                    images.add(image);
+                }
+            } else if (item.getType() == SchemaOrgParser.Type.IMAGE) {
+                SchemaOrgParser.ImageItem imageItem = (SchemaOrgParser.ImageItem) item;
+                image = imageItem.getImage();
+                // Insert |image| at beginning of list if it's the first image that's
+                // representative of page or it's the associated image of the first article.
+                if (!hasRepresentativeImage && (imageItem == associatedImageOfArticle ||
+                                                imageItem.isRepresentativeOfPage())) {
+                    hasRepresentativeImage = true;
+                    images.add(0, image);
+                } else {
+                    images.add(image);
+                }
+            }
+        }
+
+        // Prepend |imageOfArticle| to list if there's no image representative of page; append it
+        // otherwise.
+        if (imageOfArticle != null) {
+            if (!hasRepresentativeImage) images.add(0, imageOfArticle);
+            else images.add(imageOfArticle);
+        }
+
+        if (images.isEmpty()) return null;
+
+        return images.toArray(new MarkupParser.Image[images.size()]);
+    }
+
+    @Override
+    public String getDescription() {
+        SchemaOrgParser.ArticleItem item = parser.findFirstArticle();
+        return item != null ? item.getStringProperty(SchemaOrgParser.DESCRIPTION_PROP) : "";
+    }
+
+    @Override
+    public String getPublisher() {
+        // Returns either the "publisher" or "copyrightHolder" property of the first article.
+        String publisher = "";
+        SchemaOrgParser.ArticleItem article = parser.findFirstArticle();
+        if (article != null) {
+            publisher = article.getPersonOrOrganizationName(SchemaOrgParser.PUBLISHER_PROP);
+            if (publisher.isEmpty()) {
+                publisher = article.getPersonOrOrganizationName(
+                        SchemaOrgParser.COPYRIGHT_HOLDER_PROP);
+            }
+        }
+        return publisher;
+    }
+
+    @Override
+    public String getCopyright() {
+        // Returns a concatenated string of copyright year and copyright holder of the first article
+        // that has these properties, delimited by a whitespace.
+        SchemaOrgParser.ArticleItem item = parser.findFirstArticle();
+        if (item == null) return "";
+        String copyright = SchemaOrgParser.concat(
+                item.getStringProperty(SchemaOrgParser.COPYRIGHT_YEAR_PROP),
+                item.getPersonOrOrganizationName(SchemaOrgParser.COPYRIGHT_HOLDER_PROP));
+        return copyright.isEmpty() ? copyright : "Copyright " + copyright;
+    }
+
+    @Override
+    public String getAuthor() {
+        String author = "";
+        SchemaOrgParser.ArticleItem item = parser.findFirstArticle();
+        if (item != null) {
+             author = item.getPersonOrOrganizationName(SchemaOrgParser.AUTHOR_PROP);
+             // If there's no "author" property, use "creator" property.
+             if (author.isEmpty()) {
+                 author = item.getPersonOrOrganizationName(SchemaOrgParser.CREATOR_PROP);
+             }
+        }
+        // Otherwise, use "rel=author" tag.
+        return author.isEmpty() ? parser.getAuthorFromRel() : author;
+    }
+
+    @Override
+    public MarkupParser.Article getArticle() {
+        SchemaOrgParser.ArticleItem item = parser.findFirstArticle();
+        return item != null ? item.getArticle() : null;
+    }
+
+    @Override
+    public boolean optOut() {
+        return false;
+    }
+}