JavaScript that returns whether or not a page is an article

Has a JavaScript file that contains a function to tell whether or not a
page is an article and related functions to get a string of that JavaScript
so that it can be injected into the page.

BUG=383630

Review URL: https://codereview.chromium.org/464973002

Cr-Commit-Position: refs/heads/master@{#289192}
git-svn-id: svn://svn.chromium.org/chrome/trunk/src@289192 0039d316-1c4b-4281-b951-d872f2087c98

components/dom_distiller/android/java/src/org/chromium/components/dom_distiller/core/DomDistillerUrlUtils.java

@@ -49,7 +49,14 @@ public final class DomDistillerUrlUtils {
         return nativeIsUrlDistillable(url);
     }
 
+    // TODO(yfriedman): Change method so that it takes in a WebContents and a
+    // callback.
+    public static String getIsDistillableJs() {
+        return nativeGetIsDistillableJs();
+    }
+
     private static native String nativeGetDistillerViewUrlFromUrl(String scheme, String url);
+    private static native String nativeGetIsDistillableJs();
     private static native String nativeGetOriginalUrlFromDistillerUrl(String viewerUrl);
     private static native boolean nativeIsDistilledPage(String url);
     private static native boolean nativeIsUrlDistillable(String url);

components/dom_distiller/core/javascript/is_distillable_trigger.js

+// 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.
+
+(function() {
+  var elems = document.querySelectorAll(
+      'meta[property="og:type"],meta[name=\\"og:type\\"]');
+  for (var i in elems) {
+    if (elems[i].content && elems[i].content.toUpperCase() == 'ARTICLE') {
+      return true;
+    }
+  }
+  var elems = document.querySelectorAll(
+      '*[itemtype="http://schema.org/Article"]');
+  for (var i in elems) {
+    if (elems[i].itemscope) {
+      return true;
+    }
+  }
+  return false;
+})()

components/dom_distiller/core/url_utils.cc

@@ -8,7 +8,9 @@
 
 #include "base/guid.h"
 #include "components/dom_distiller/core/url_constants.h"
+#include "grit/component_resources.h"
 #include "net/base/url_util.h"
+#include "ui/base/resource/resource_bundle.h"
 #include "url/gurl.h"
 
 namespace dom_distiller {
@@ -51,6 +53,11 @@ bool IsDistilledPage(const GURL& url) {
   return url.is_valid() && url.scheme() == kDomDistillerScheme;
 }
 
+std::string GetIsDistillableJs() {
+  return ResourceBundle::GetSharedInstance()
+      .GetRawDataResource(IDR_IS_DISTILLABLE_JS).as_string();
+}
+
 }  // namespace url_utils
 
 }  // namespace dom_distiller

components/dom_distiller/core/url_utils.h

@@ -31,6 +31,10 @@ bool IsUrlDistillable(const GURL& url);
 // Returns whether the given |url| is for a distilled page.
 bool IsDistilledPage(const GURL& url);
 
+// Returns a JavaScript snippet that returns whether or not a page should be
+// used with DomDistillerService and can be executed in a live page.
+std::string GetIsDistillableJs();
+
 }  // namespace url_utils
 
 }  // namespace dom_distiller

components/dom_distiller/core/url_utils_android.cc

@@ -63,6 +63,11 @@ jboolean IsUrlDistillable(JNIEnv* env, jclass clazz, jstring j_url) {
   return dom_distiller::url_utils::IsUrlDistillable(url);
 }
 
+jstring GetIsDistillableJs(JNIEnv* env, jclass clazz) {
+  return base::android::ConvertUTF8ToJavaString(
+      env, dom_distiller::url_utils::GetIsDistillableJs()).Release();
+}
+
 bool RegisterUrlUtils(JNIEnv* env) { return RegisterNativesImpl(env); }
 
 }  // namespace android

components/resources/dom_distiller_resources.grdp

@@ -7,4 +7,5 @@
   <include name="IDR_DOM_DISTILLER_VIEWER_JS" file="../dom_distiller/content/resources/dom_distiller_viewer.js" type="BINDATA" />
   <include name="IDR_DISTILLER_JS" file="../dom_distiller/core/javascript/domdistiller.js" flattenhtml="true" type="BINDATA" />
   <include name="IDR_DISTILLER_CSS" file="../dom_distiller/core/css/distilledpage.css" type="BINDATA" />
+  <include name="IDR_IS_DISTILLABLE_JS" file="../dom_distiller/core/javascript/is_distillable_trigger.js" type="BINDATA" />
 </grit-part>