Headless demo

headless/app/headless_browser_main.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 "headless/app/headless_browser_main.h"
+
+#include "base/thread_task_runner_handle.h"
+#include "content/public/browser/render_frame_host.h"
+#include "content/public/browser/browser_main_runner.h"
+#include "content/public/browser/web_contents.h"
+#include "content/public/renderer/render_frame.h"
+#include "content/shell/browser/shell.h"
+#include "content/shell/browser/shell_browser_context.h"
+#include "content/shell/browser/shell_content_browser_client.h"
+#include "url/gurl.h"
+
+#include "base/json/json_writer.h"
+
+#include "third_party/WebKit/public/web/WebLocalFrame.h"
+#include "third_party/WebKit/public/web/WebNodeList.h"
+#include "third_party/WebKit/public/web/WebDocument.h"
+#include "base/threading/thread.h"
+#include "content/public/browser/browser_thread.h"
+#include "third_party/WebKit/public/platform/WebThread.h"
+#include "third_party/WebKit/public/platform/Platform.h"
+#include "third_party/WebKit/public/platform/WebTraceLocation.h"
+#include "third_party/WebKit/public/platform/WebTaskRunner.h"
+#include "third_party/WebKit/public/web/WebElement.h"
+#include "base/command_line.h"
+
+#include <iostream>
+#include <fstream>
+
+class Task : public blink::WebTaskRunner::Task {
+  public:
+    Task(std::function<void(void)> function_to_run): function_to_run_(function_to_run) { };
+    ~Task() { };
+
+    void run() override {
+      function_to_run_();
+    }
+  private:
+    std::function<void(void)> function_to_run_;
+};
+
+class Observer : public content::WebContentsObserver {
+  public:
+    Observer(content::WebContents* web_contents)
+      : WebContentsObserver(web_contents),
+        web_contents_(web_contents)
+    {
+    };
+
+    void DidStopLoading() override {
+      DumpDOMTree();
+    }
+
+    void DumpDOMTree() {
+      content::RenderFrame* render_frame =
+          content::RenderFrame::FromRoutingID(web_contents_->GetMainFrame()->GetRoutingID());
+      blink::WebFrame* web_frame = render_frame->GetWebFrame();
+
+      blink::Platform::current()->mainThread()->taskRunner()->postTask(BLINK_FROM_HERE,
+          new Task([this, web_frame]() {
+            auto dumped_tree = DumpNode(web_frame->document());
+
+            std::string dumped_string;
+            base::JSONWriter::WriteWithOptions(
+                *dumped_tree,
+                base::JSONWriter::OPTIONS_PRETTY_PRINT,
+                &dumped_string);
+
+            std::cerr << dumped_string << std::endl;
+
+            shutDownGraciously();
+          }));
+    }
+
+    static std::string NodeType(const blink::WebNode& node) {
+      if (node.isLink()) {
+        return "link";
+      } else if (node.isDocumentNode()) {
+        return "document";
+      } else if (node.isCommentNode()) {
+        return "comment";
+      } else if (node.isTextNode()) {
+        return "text";
+      } else if (node.isElementNode()) {
+        return "element";
+      } else {
+        return "unknown";
+      }
+    }
+
+    static void DumpElementNode(const blink::WebElement& element, base::DictionaryValue* output) {
+      output->SetString("tag", element.tagName());
+
+      output->Set("attributes", new base::DictionaryValue());
+      base::DictionaryValue* attributes;
+      output->GetDictionary("attributes", &attributes);
+
+      for (size_t i = 0; i < element.attributeCount(); ++i) {
+        attributes->SetString(
+            element.attributeLocalName(i).utf8(),
+            element.attributeValue(i).utf8());
+      }
+    }
+
+    void DumpBoundingBox(const blink::WebNode& node, base::DictionaryValue* output) {
+      output->Set("bounding_box", new base::DictionaryValue());
+      base::DictionaryValue* bounding_box;
+      output->GetDictionary("bounding_box", &bounding_box);
+
+      bounding_box->SetInteger("minX", node.boundingBox().minX);
+      bounding_box->SetInteger("maxX", node.boundingBox().maxX);
+      bounding_box->SetInteger("minY", node.boundingBox().minY);
+      bounding_box->SetInteger("maxY", node.boundingBox().maxY);
+    }
+
+    scoped_ptr<base::DictionaryValue> DumpNode(const blink::WebNode& node) {
+      scoped_ptr<base::DictionaryValue> value(new base::DictionaryValue());
+      value->SetString("type", NodeType(node));
+      value->SetString("name", node.nodeName());
+      value->SetString("value", node.nodeValue());
+
+      DumpBoundingBox(node, value.get());
+
+      if (node.isElementNode()) {
+        DumpElementNode(static_cast<const blink::WebElement&>(node), value.get());
+      }
+
+      if (node.hasChildNodes()) {
+        value->Set("children", new base::ListValue());
+
+        base::ListValue* children;
+        value->GetList("children", &children);
+
+        for (size_t i = 0; i < node.childNodes().length(); ++i) {
+          children->Append(DumpNode(node.childNodes().item(i)));
+        }
+      }
+
+      return value;
+    }
+
+
+    void shutDownGraciously() {
+      content::BrowserThread::PostTask(
+          content::BrowserThread::UI, FROM_HERE, base::MessageLoop::QuitWhenIdleClosure());
+    }
+  private:
+    content::WebContents* web_contents_;
+};
+
+int HeadlessBrowserMain(
+     const content::MainFunctionParams& parameters,
+     const scoped_ptr<content::BrowserMainRunner>& main_runner) {
+  int exit_code = main_runner->Initialize(parameters);
+
+  DCHECK_LT(exit_code, 0)
+      << "BrowserMainRunner::Initialize failed in HeadlessBrowserMain";
+
+  if (exit_code >= 0)
+    return exit_code;
+
+  content::Shell* main_window = content::Shell::CreateNewWindow(
+      content::ShellContentBrowserClient::Get()->browser_context(),
+      GURL(),
+      NULL,
+      gfx::Size(700, 700));
+
+  base::CommandLine::StringVector args =
+      base::CommandLine::ForCurrentProcess()->GetArgs();
+  if (args.empty()) {
+    std::cerr << "Provide url to open" << std::endl;
+    return 666;
+  }
+  GURL url(args[0]);
+
+  Observer observer(main_window->web_contents());
+
+  main_window->LoadURL(url);
+
+  main_runner->Run();
+
+  main_runner->Shutdown();
+
+  return exit_code;
+}
+
+