Add support for providing an external file for extracting content.

components/dom_distiller/standalone/content_extractor.cc

 // 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.
 
 #include <sstream>
 
 #include "base/command_line.h"
+#include "base/files/file_path.h"
+#include "base/files/file_util.h"
 #include "base/files/scoped_temp_dir.h"
 #include "base/id_map.h"
 #include "base/message_loop/message_loop.h"
 #include "base/path_service.h"
 #include "base/run_loop.h"
 #include "base/strings/string_number_conversions.h"
 #include "base/strings/string_split.h"
 #include "components/dom_distiller/content/distiller_page_web_contents.h"
 #include "components/dom_distiller/core/article_entry.h"
 #include "components/dom_distiller/core/distilled_page_prefs.h"
@@ skipping 84 matching lines @@
 // The original domain of the page if |kUrlSwitch| is a file.
 const char* kOriginalDomain = "original-domain";
 
 // A semi-colon-separated (i.e. ';') list of original domains corresponding to
 // "kUrlsSwitch".
 const char* kOriginalDomains = "original-domains";
 
 // Maximum number of concurrent started extractor requests.
 const int kMaxExtractorTasks = 8;
 
+// A path to a script for extracting content (domdistiller.js). If this argument
+// is passed in, the script will be used instead of using the bundled version
+// of the script.
+const char* kExternalDomDistillerJs = "external-dom-distiller-js";
+
 scoped_ptr<DomDistillerService> CreateDomDistillerService(
     content::BrowserContext* context,
     const base::FilePath& db_path,
     const UrlToDomainMap& url_to_domain_map) {
   scoped_refptr<base::SequencedTaskRunner> background_task_runner =
       content::BrowserThread::GetBlockingPool()->GetSequencedTaskRunner(
           content::BrowserThread::GetBlockingPool()->GetSequenceToken());
 
   // TODO(cjhopman): use an in-memory database instead of an on-disk one with
   // temporary directory.
   scoped_ptr<leveldb_proto::ProtoDatabaseImpl<ArticleEntry> > db(
       new leveldb_proto::ProtoDatabaseImpl<ArticleEntry>(
           background_task_runner));
   scoped_ptr<DomDistillerStore> dom_distiller_store(
       new DomDistillerStore(db.Pass(), db_path));
 
-  scoped_ptr<DistillerPageFactory> distiller_page_factory(
-      new DistillerPageWebContentsFactory(context));
+  scoped_ptr<DistillerPageFactory> distiller_page_factory;
+  if (base::CommandLine::ForCurrentProcess()->HasSwitch(
+          kExternalDomDistillerJs)) {
+    std::string external_script_path =
+        base::CommandLine::ForCurrentProcess()->GetSwitchValueASCII(
+            kExternalDomDistillerJs);
+    std::string script_content;
+    if (!base::ReadFileToString(base::FilePath(external_script_path),
+                                &script_content)) {
+      ADD_FAILURE() << "Failed to read external script for distillation.";
+      return nullptr;
+    }
+    distiller_page_factory.reset(
+        new DistillerPageWebContentsFactory(context, script_content));
+  } else {
+    distiller_page_factory.reset(new DistillerPageWebContentsFactory(context));
+  }
+
   scoped_ptr<DistillerURLFetcherFactory> distiller_url_fetcher_factory(
       new DistillerURLFetcherFactory(context->GetRequestContext()));
 
   dom_distiller::proto::DomDistillerOptions options;
   if (base::CommandLine::ForCurrentProcess()->HasSwitch(kExtractTextOnly)) {
     options.set_extract_text_only(true);
   }
   int debug_level = 0;
   if (base::CommandLine::ForCurrentProcess()->HasSwitch(kDebugLevel) &&
       base::StringToInt(
@@ skipping 269 matching lines @@
   std::string output_data_;
   scoped_ptr<google::protobuf::io::StringOutputStream> protobuf_output_stream_;
 };
 
 IN_PROC_BROWSER_TEST_F(ContentExtractor, MANUAL_ExtractUrl) {
   Start();
   base::RunLoop().Run();
 }
 
 }  // namespace dom_distiller