Test distillability without JavaScript

components/dom_distiller/content/renderer/distillability_agent.cc

+// Copyright 2015 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.
+
+#include "components/dom_distiller/content/common/distiller_messages.h"
+#include "components/dom_distiller/content/renderer/distillability_agent.h"
+
+#include "content/public/renderer/render_frame.h"
+
+#include "third_party/WebKit/public/web/WebDocument.h"
+#include "third_party/WebKit/public/web/WebElement.h"
+#include "third_party/WebKit/public/web/WebLocalFrame.h"
+#include "third_party/WebKit/public/web/WebNode.h"
+#include "third_party/WebKit/public/web/WebNodeList.h"
+
+namespace dom_distiller {
+
+namespace {
+
+using namespace blink;
+
+class ExtractFeatureWalker {
+public:
+  ExtractFeatureWalker()
+  : numElements(0),
+    numAnchors(0),
+    numForms(0) {
+  }
+  void walk(WebElement root) {
+    blink::WebNodeList children = root.childNodes();
+    for (unsigned i = 0; i < children.length(); ++i) {
+      WebNode child = children.item(i);
+      if (!child.isElementNode())
+        continue;
+      WebElement elem = child.to<WebElement>();
+      numElements++;
+      if (elem.hasHTMLTagName("a")) {
+        numAnchors++;
+      }
+      if (elem.hasHTMLTagName("form")) {
+        numForms++;
+      }
+      walk(elem);
+    }
+  }
+  int numElements;
+  int numAnchors;
+  int numForms;
+};
+
+bool hasOGArticle(WebElement head) {
+  blink::WebNodeList children = head.childNodes();
+  for (unsigned i = 0; i < children.length(); ++i) {
+    WebNode child = children.item(i);
+    if (!child.isElementNode())
+      continue;
+    WebElement elem = child.to<WebElement>();
+    if (!elem.hasHTMLTagName("meta"))
+      continue;
+    if ((elem.getAttribute("name").equals("og:type")) ||
+        (elem.getAttribute("property").equals("og:type"))) {
+      std::string content = elem.getAttribute("content").utf8();
+      if (base::ToUpperASCII(content) == "ARTICLE") {
+        return true;
+      }
+    }
+  }
+  return false;
+};
+
+}  // namespace
+
+DistillabilityAgent::DistillabilityAgent(
+    content::RenderFrame* render_frame)
+    : RenderFrameObserver(render_frame) {
+}
+
+bool DistillabilityAgent::OnMessageReceived(const IPC::Message& msg) {
+  bool handled = true;
+  IPC_BEGIN_MESSAGE_MAP(DistillabilityAgent, msg)
+    IPC_MESSAGE_HANDLER(FrameMsg_ExtractFeatureRequest,
+                        OnExtractFeatureRequest)
+    IPC_MESSAGE_UNHANDLED(handled = false)
+  IPC_END_MESSAGE_MAP()
+  return handled;
+}
+
+void DistillabilityAgent::OnExtractFeatureRequest(int id) {
+  FrameHostMsg_ExtractFeatureResponse_Params result;
+  DCHECK(render_frame());
+  if (!render_frame() || !render_frame()->GetWebFrame()) {
+    Send(new FrameHostMsg_ExtractFeatureResponse(routing_id(), id, result));
+    return;
+  }
+  WebDocument doc = render_frame()->GetWebFrame()->document();
+  if (doc.isNull() || doc.body().isNull()) {
+    Send(new FrameHostMsg_ExtractFeatureResponse(routing_id(), id, result));
+    return;
+  }
+
+  result.isOGArticle = hasOGArticle(doc.head());
+  result.url = doc.url().string().utf8();
+
+  ExtractFeatureWalker walker;
+  walker.walk(doc.body());
+
+  result.numElements = walker.numElements;
+  result.numAnchors = walker.numAnchors;
+  result.numForms = walker.numForms;
+
+  result.innerText = doc.contentAsTextForTesting().utf8();
+  result.textContent = doc.body().textContent().utf8();
+  result.innerHTML = doc.body().innerHTML().utf8();
+
+  Send(new FrameHostMsg_ExtractFeatureResponse(routing_id(), id, result));
+}
+
+DistillabilityAgent::~DistillabilityAgent() {}
+
+}  // namespace dom_distiller