Extracting page shingle hashes for similarity detection.

chrome/renderer/safe_browsing/phishing_classifier.h

 // Copyright (c) 2012 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.
 //
 // This class handles the process of extracting all of the features from a
 // page and computing a phishyness score.  The basic steps are:
 //  - Run each feature extractor over the page, building up a FeatureMap of
 //    feature -> value.
 //  - SHA-256 hash all of the feature names in the map so that they match the
 //    supplied model.
 //  - Hand the hashed map off to a Scorer, which computes the probability that
 //    the page is phishy.
 //  - If the page is phishy, run the supplied callback.
 //
 // For more details, see phishing_*_feature_extractor.h, scorer.h, and
 // client_model.proto.
 
 #ifndef CHROME_RENDERER_SAFE_BROWSING_PHISHING_CLASSIFIER_H_
 #define CHROME_RENDERER_SAFE_BROWSING_PHISHING_CLASSIFIER_H_
 
+#include <set>
+
 #include "base/basictypes.h"
 #include "base/callback.h"
 #include "base/memory/scoped_ptr.h"
 #include "base/memory/weak_ptr.h"
 #include "base/strings/string16.h"
 
 namespace content {
 class RenderView;
 }
 
@@ skipping 97 matching lines @@
 
   content::RenderView* render_view_;  // owns us
   const Scorer* scorer_;  // owned by the caller
   scoped_ptr<FeatureExtractorClock> clock_;
   scoped_ptr<PhishingUrlFeatureExtractor> url_extractor_;
   scoped_ptr<PhishingDOMFeatureExtractor> dom_extractor_;
   scoped_ptr<PhishingTermFeatureExtractor> term_extractor_;
 
   // State for any in-progress extraction.
   scoped_ptr<FeatureMap> features_;
+  scoped_ptr<std::set<uint32> > shingle_hashes_;
   const base::string16* page_text_;  // owned by the caller
   DoneCallback done_callback_;
 
   // Used in scheduling BeginFeatureExtraction tasks.
   // These pointers are invalidated if classification is cancelled.
   base::WeakPtrFactory<PhishingClassifier> weak_factory_;
 
   DISALLOW_COPY_AND_ASSIGN(PhishingClassifier);
 };
 
 }  // namespace safe_browsing
 
 #endif  // CHROME_RENDERER_SAFE_BROWSING_PHISHING_CLASSIFIER_H_