summaryrefslogtreecommitdiffstats
path: root/chrome/renderer/safe_browsing/phishing_classifier.cc
blob: 32fa48ed73ca19acd6765edea9d164c312404d28 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
// Copyright (c) 2010 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/callback.h"
#include "base/compiler_specific.h"
#include "base/logging.h"
#include "base/sha2.h"
#include "chrome/renderer/render_view.h"
#include "chrome/renderer/safe_browsing/feature_extractor_clock.h"
#include "chrome/renderer/safe_browsing/features.h"
#include "chrome/renderer/safe_browsing/phishing_dom_feature_extractor.h"
#include "chrome/renderer/safe_browsing/phishing_term_feature_extractor.h"
#include "chrome/renderer/safe_browsing/phishing_url_feature_extractor.h"
#include "chrome/renderer/safe_browsing/scorer.h"
#include "googleurl/src/gurl.h"
#include "third_party/WebKit/WebKit/chromium/public/WebFrame.h"
#include "third_party/WebKit/WebKit/chromium/public/WebURL.h"
#include "third_party/WebKit/WebKit/chromium/public/WebView.h"

namespace safe_browsing {

const double PhishingClassifier::kInvalidScore = -1.0;
const double PhishingClassifier::kPhishyThreshold = 0.5;

PhishingClassifier::PhishingClassifier(RenderView* render_view,
                                       const Scorer* scorer,
                                       FeatureExtractorClock* clock)
    : render_view_(render_view),
      scorer_(scorer),
      clock_(clock),
      ALLOW_THIS_IN_INITIALIZER_LIST(method_factory_(this)) {
  url_extractor_.reset(new PhishingUrlFeatureExtractor);
  dom_extractor_.reset(
      new PhishingDOMFeatureExtractor(render_view_, clock_.get()));
  term_extractor_.reset(new PhishingTermFeatureExtractor(
      &scorer_->page_terms(),
      &scorer_->page_words(),
      scorer_->max_words_per_term(),
      clock_.get()));

  Clear();
}

PhishingClassifier::~PhishingClassifier() {
  // The RenderView should have called CancelPendingClassification() before
  // we are destroyed.
  CheckNoPendingClassification();
}

void PhishingClassifier::BeginClassification(const string16* page_text,
                                             DoneCallback* done_callback) {
  // The RenderView should have called CancelPendingClassification() before
  // starting a new classification, so DCHECK this.
  CheckNoPendingClassification();
  // However, in an opt build, we will go ahead and clean up the pending
  // classification so that we can start in a known state.
  CancelPendingClassification();

  page_text_ = page_text;
  done_callback_.reset(done_callback);

  // For consistency, we always want to invoke the DoneCallback
  // asynchronously, rather than directly from this method.  To ensure that
  // this is the case, post a task to begin feature extraction on the next
  // iteration of the message loop.
  MessageLoop::current()->PostTask(
      FROM_HERE,
      method_factory_.NewRunnableMethod(
          &PhishingClassifier::BeginFeatureExtraction));
}

void PhishingClassifier::BeginFeatureExtraction() {
  WebKit::WebView* web_view = render_view_->webview();
  if (!web_view) {
    RunFailureCallback();
    return;
  }

  WebKit::WebFrame* frame = web_view->mainFrame();
  if (!frame) {
    RunFailureCallback();
    return;
  }

  features_.reset(new FeatureMap);
  if (!url_extractor_->ExtractFeatures(GURL(frame->url()), features_.get())) {
    RunFailureCallback();
    return;
  }

  // DOM feature extraction can take awhile, so it runs asynchronously
  // in several chunks of work and invokes the callback when finished.
  dom_extractor_->ExtractFeatures(
      features_.get(),
      NewCallback(this, &PhishingClassifier::DOMExtractionFinished));
}

void PhishingClassifier::CancelPendingClassification() {
  // Note that cancelling the feature extractors is simply a no-op if they
  // were not running.
  dom_extractor_->CancelPendingExtraction();
  term_extractor_->CancelPendingExtraction();
  method_factory_.RevokeAll();
  Clear();
}

void PhishingClassifier::DOMExtractionFinished(bool success) {
  if (success) {
    // Term feature extraction can take awhile, so it runs asynchronously
    // in several chunks of work and invokes the callback when finished.
    term_extractor_->ExtractFeatures(
        page_text_,
        features_.get(),
        NewCallback(this, &PhishingClassifier::TermExtractionFinished));
  } else {
    RunFailureCallback();
  }
}

void PhishingClassifier::TermExtractionFinished(bool success) {
  if (success) {
    // Hash all of the features so that they match the model, then compute
    // the score.
    FeatureMap hashed_features;
    for (base::hash_map<std::string, double>::const_iterator it =
             features_->features().begin();
         it != features_->features().end(); ++it) {
      DCHECK(hashed_features.AddRealFeature(base::SHA256HashString(it->first),
                                            it->second));
    }

    double score = scorer_->ComputeScore(hashed_features);
    RunCallback(score >= kPhishyThreshold, score);
  } else {
    RunFailureCallback();
  }
}

void PhishingClassifier::CheckNoPendingClassification() {
  DCHECK(!done_callback_.get());
  DCHECK(!page_text_);
  if (done_callback_.get() || page_text_) {
    LOG(ERROR) << "Classification in progress, missing call to "
               << "CancelPendingClassification";
  }
}

void PhishingClassifier::RunCallback(bool phishy, double phishy_score) {
  done_callback_->Run(phishy, phishy_score);
  Clear();
}

void PhishingClassifier::RunFailureCallback() {
  RunCallback(false /* not phishy */, kInvalidScore);
}

void PhishingClassifier::Clear() {
  page_text_ = NULL;
  done_callback_.reset(NULL);
  features_.reset(NULL);
}

}  // namespace safe_browsing