chrome/renderer/safe_browsing/scorer.h - Issue 2667343006: Componentize safe_browsing [X+1] : move the renderer part to component.

Side by Side Diff: chrome/renderer/safe_browsing/scorer.h

Issue 2667343006: Componentize safe_browsing [X+1] : move the renderer part to component.

Patch Set: Created 3 years, 10 months ago

Use n/p to move between diff chunks; N/P to move between comments. Draft comments are only viewable by you.

Jump to:

OLD	NEW
	(Empty)
1 // Copyright (c) 2011 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 // This class loads a client-side model and lets you compute a phishing score

6 // for a set of previously extracted features. The phishing score corresponds

7 // to the probability that the features are indicative of a phishing site.

8 //

9 // For more details on how the score is actually computed for a given model

10 // and a given set of features read the comments in client_model.proto file.

11 //

12 // See features.h for a list of features that are currently used.

13

14 #ifndef CHROME_RENDERER_SAFE_BROWSING_SCORER_H_

15 #define CHROME_RENDERER_SAFE_BROWSING_SCORER_H_

16

17 #include <stddef.h>

18 #include <stdint.h>

19

20 #include <string>

21

22 #include "base/containers/hash_tables.h"

23 #include "base/macros.h"

24 #include "base/strings/string_piece.h"

25 #include "chrome/common/safe_browsing/client_model.pb.h"

26

27 namespace safe_browsing {

28 class FeatureMap;

29

30 // Scorer methods are virtual to simplify mocking of this class.

31 class Scorer {

32 public:

33 virtual ~Scorer();

34

35 // Factory method which creates a new Scorer object by parsing the given

36 // model. If parsing fails this method returns NULL.

37 static Scorer* Create(const base::StringPiece& model_str);

38

39 // This method computes the probability that the given features are indicative

40 // of phishing. It returns a score value that falls in the range [0.0,1.0]

41 // (range is inclusive on both ends).

42 virtual double ComputeScore(const FeatureMap& features) const;

43

44 // Returns the version number of the loaded client model.

45 int model_version() const;

46

47 // -- Accessors used by the page feature extractor ---------------------------

48

49 // Returns a set of hashed page terms that appear in the model in binary

50 // format.

51 const base::hash_set<std::string>& page_terms() const;

52

53 // Returns a set of hashed page words that appear in the model in binary

54 // format.

55 const base::hash_set<uint32_t>& page_words() const;

56

57 // Return the maximum number of words per term for the loaded model.

58 size_t max_words_per_term() const;

59

60 // Returns the murmurhash3 seed for the loaded model.

61 uint32_t murmurhash3_seed() const;

62

63 // Return the maximum number of unique shingle hashes per page.

64 size_t max_shingles_per_page() const;

65

66 // Return the number of words in a shingle.

67 size_t shingle_size() const;

68

69 protected:

70 // Most clients should use the factory method. This constructor is public

71 // to allow for mock implementations.

72 Scorer();

73

74 private:

75 friend class PhishingScorerTest;

76

77 // Computes the score for a given rule and feature map. The score is computed

78 // by multiplying the rule weight with the product of feature weights for the

79 // given rule. The feature weights are stored in the feature map. If a

80 // particular feature does not exist in the feature map we set its weight to

81 // zero.

82 double ComputeRuleScore(const ClientSideModel::Rule& rule,

83 const FeatureMap& features) const;

84

85 ClientSideModel model_;

86 base::hash_set<std::string> page_terms_;

87 base::hash_set<uint32_t> page_words_;

88

89 DISALLOW_COPY_AND_ASSIGN(Scorer);

90 };

91 } // namespace safe_browsing

92

93 #endif // CHROME_RENDERER_SAFE_BROWSING_SCORER_H_

OLD	NEW