src/chrome/renderer/safe_browsing/phishing_url_feature_extractor.h

   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 // PhishingUrlFeatureExtractor handles computing URL-based features for
   6 // the client-side phishing detection model.  These include tokens in the
   7 // host and path, features pertaining to host length, and IP addresses.
   8
   9 #ifndef CHROME_RENDERER_SAFE_BROWSING_PHISHING_URL_FEATURE_EXTRACTOR_H_
  10 #define CHROME_RENDERER_SAFE_BROWSING_PHISHING_URL_FEATURE_EXTRACTOR_H_
  11
  12 #include <string>
  13 #include <vector>
  14
  15 #include "base/basictypes.h"
  16
  17 class GURL;
  18
  19 namespace safe_browsing {
  20 class FeatureMap;
  21
  22 class PhishingUrlFeatureExtractor {
  23  public:
  24   PhishingUrlFeatureExtractor();
  25   ~PhishingUrlFeatureExtractor();
  26
  27   // Extracts features for |url| into the given feature map.
  28   // Returns true on success.
  29   bool ExtractFeatures(const GURL& url, FeatureMap* features);
  30
  31  private:
  32   friend class PhishingUrlFeatureExtractorTest;
  33
  34   static const size_t kMinPathComponentLength = 3;
  35
  36   // Given a string, finds all substrings of consecutive alphanumeric
  37   // characters of length >= kMinPathComponentLength and inserts them into
  38   // tokens.
  39   static void SplitStringIntoLongAlphanumTokens(
  40       const std::string& full,
  41       std::vector<std::string>* tokens);
  42
  43   DISALLOW_COPY_AND_ASSIGN(PhishingUrlFeatureExtractor);
  44 };
  45
  46 }  // namespace safe_browsing
  47
  48 #endif  // CHROME_RENDERER_SAFE_BROWSING_PHISHING_URL_FEATURE_EXTRACTOR_H_