components/dom_distiller/core/distillable_page_detector.cc

   1 // Copyright 2015 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 "components/dom_distiller/core/distillable_page_detector.h"
   6
   7 #include "base/logging.h"
   8 #include "grit/components_resources.h"
   9 #include "ui/base/resource/resource_bundle.h"
  10
  11 namespace dom_distiller {
  12
  13 const DistillablePageDetector* DistillablePageDetector::GetDefault() {
  14   static DistillablePageDetector* detector = nullptr;
  15   if (!detector) {
  16     std::string serialized_proto =
  17         ResourceBundle::GetSharedInstance()
  18             .GetRawDataResource(IDR_DISTILLABLE_PAGE_SERIALIZED_MODEL)
  19             .as_string();
  20     scoped_ptr<AdaBoostProto> proto(new AdaBoostProto);
  21     CHECK(proto->ParseFromString(serialized_proto));
  22     detector = new DistillablePageDetector(proto.Pass());
  23   }
  24   return detector;
  25 }
  26
  27 DistillablePageDetector::DistillablePageDetector(
  28     scoped_ptr<AdaBoostProto> proto)
  29     : proto_(proto.Pass()), threshold_(0.0) {
  30   CHECK(proto_->num_stumps() == proto_->stump_size());
  31   for (int i = 0; i < proto_->num_stumps(); ++i) {
  32     const StumpProto& stump = proto_->stump(i);
  33     CHECK(stump.feature_number() >= 0);
  34     CHECK(stump.feature_number() < proto_->num_features());
  35     threshold_ += stump.weight() / 2.0;
  36   }
  37 }
  38
  39 DistillablePageDetector::~DistillablePageDetector() {
  40 }
  41
  42 bool DistillablePageDetector::Classify(
  43     const std::vector<double>& features) const {
  44   return Score(features) > threshold_;
  45 }
  46
  47 double DistillablePageDetector::Score(
  48     const std::vector<double>& features) const {
  49   if (features.size() != size_t(proto_->num_features())) {
  50     return 0.0;
  51   }
  52   double score = 0.0;
  53   for (int i = 0; i < proto_->num_stumps(); ++i) {
  54     const StumpProto& stump = proto_->stump(i);
  55     if (features[stump.feature_number()] > stump.split()) {
  56       score += stump.weight();
  57     }
  58   }
  59   return score;
  60 }
  61
  62 double DistillablePageDetector::GetThreshold() const {
  63   return threshold_;
  64 }
  65
  66 }  // namespace dom_distiller