chrome/renderer/safe_browsing/features.cc

   1 // Copyright (c) 2010 The Chromium Authors. All rights reserved.
   2 // Use of this source code is governed by a BSD-style license that can be
   3 // found in the LICENSE file.
   4
   5 #include "chrome/renderer/safe_browsing/features.h"
   6
   7 #include "base/logging.h"
   8 #include "base/metrics/histogram.h"
   9
  10 namespace safe_browsing {
  11
  12 const size_t FeatureMap::kMaxFeatureMapSize = 10000;
  13
  14 FeatureMap::FeatureMap() {}
  15 FeatureMap::~FeatureMap() {}
  16
  17 bool FeatureMap::AddBooleanFeature(const std::string& name) {
  18   return AddRealFeature(name, 1.0);
  19 }
  20
  21 bool FeatureMap::AddRealFeature(const std::string& name, double value) {
  22   if (features_.size() >= kMaxFeatureMapSize) {
  23     // If we hit this case, it indicates that either kMaxFeatureMapSize is
  24     // too small, or there is a bug causing too many features to be added.
  25     // In this case, we'll log to a histogram so we can see that this is
  26     // happening, and make phishing classification fail silently.
  27     LOG(ERROR) << "Not adding feature: " << name << " because the "
  28                << "feature map is too large.";
  29     UMA_HISTOGRAM_COUNTS("SBClientPhishing.TooManyFeatures", 1);
  30     return false;
  31   }
  32   // We only expect features in the range [0.0, 1.0], so fail if the feature is
  33   // outside this range.
  34   if (value < 0.0 || value > 1.0) {
  35     LOG(ERROR) << "Not adding feature: " << name << " because the value "
  36                << value << " is not in the range [0.0, 1.0].";
  37     UMA_HISTOGRAM_COUNTS("SBClientPhishing.IllegalFeatureValue", 1);
  38     return false;
  39   }
  40
  41   features_[name] = value;
  42   return true;
  43 }
  44
  45 void FeatureMap::Clear() {
  46   features_.clear();
  47 }
  48
  49 namespace features {
  50 // URL host features
  51 const char kUrlHostIsIpAddress[] = "UrlHostIsIpAddress";
  52 const char kUrlTldToken[] = "UrlTld=";
  53 const char kUrlDomainToken[] = "UrlDomain=";
  54 const char kUrlOtherHostToken[] = "UrlOtherHostToken=";
  55
  56 // URL host aggregate features
  57 const char kUrlNumOtherHostTokensGTOne[] = "UrlNumOtherHostTokens>1";
  58 const char kUrlNumOtherHostTokensGTThree[] = "UrlNumOtherHostTokens>3";
  59
  60 // URL path features
  61 const char kUrlPathToken[] = "UrlPathToken=";
  62
  63 // DOM HTML form features
  64 const char kPageHasForms[] = "PageHasForms";
  65 const char kPageActionOtherDomainFreq[] = "PageActionOtherDomainFreq";
  66 const char kPageHasTextInputs[] = "PageHasTextInputs";
  67 const char kPageHasPswdInputs[] = "PageHasPswdInputs";
  68 const char kPageHasRadioInputs[] = "PageHasRadioInputs";
  69 const char kPageHasCheckInputs[] = "PageHasCheckInputs";
  70
  71 // DOM HTML link features
  72 const char kPageExternalLinksFreq[] = "PageExternalLinksFreq";
  73 const char kPageLinkDomain[] = "PageLinkDomain=";
  74 const char kPageSecureLinksFreq[] = "PageSecureLinksFreq";
  75
  76 // DOM HTML script features
  77 const char kPageNumScriptTagsGTOne[] = "PageNumScriptTags>1";
  78 const char kPageNumScriptTagsGTSix[] = "PageNumScriptTags>6";
  79
  80 // Other DOM HTML features
  81 const char kPageImgOtherDomainFreq[] = "PageImgOtherDomainFreq";
  82
  83 // Page term features
  84 const char kPageTerm[] = "PageTerm=";
  85
  86 }  // namespace features
  87 }  // namespace safe_browsing