Extracting page shingle hashes for similarity detection.

chrome/renderer/safe_browsing/phishing_classifier_browsertest.cc

 // 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.
 
 #include "chrome/renderer/safe_browsing/phishing_classifier.h"
 
 #include <string>
 
 #include "base/bind.h"
 #include "base/command_line.h"
@@ skipping 75 matching lines @@
     rule = model.add_rule();
     rule->add_feature(0);
     rule->add_feature(1);
     rule->add_feature(2);
     rule->set_weight(1.0);
 
     model.add_page_term(3);
     model.set_murmur_hash_seed(2777808611U);
     model.add_page_word(MurmurHash3String("login", model.murmur_hash_seed()));
     model.set_max_words_per_term(1);
+    model.set_max_shingles_per_page(200);

[noelutz, 2014/05/06 21:40:18]

Is this necessary since you're setting these as default values?

[zysxqn, 2014/05/07 19:29:19]

Prefer to set a non-default value than removing it since this will catch some
minor errors such as name inconsistency.

+    model.set_shingle_size(4);
 
     clock_ = new MockFeatureExtractorClock;
     scorer_.reset(Scorer::Create(model.SerializeAsString()));
     ASSERT_TRUE(scorer_.get());
 
     classifier_.reset(new PhishingClassifier(
         content::RenderView::FromRoutingID(kRenderViewRoutingId),
         clock_));
   }
 
@@ skipping 182 matching lines @@
   // Now set the scorer.
   classifier_->set_phishing_scorer(scorer_.get());
   EXPECT_TRUE(classifier_->is_ready());
 
   // Set a NULL scorer, which turns detection back off.
   classifier_->set_phishing_scorer(NULL);
   EXPECT_FALSE(classifier_->is_ready());
 }
 
 }  // namespace safe_browsing