components/dom_distiller/core/page_features.h

   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 #ifndef COMPONENTS_DOM_DISTILLER_CORE_PAGE_FEATURES_H_
   6 #define COMPONENTS_DOM_DISTILLER_CORE_PAGE_FEATURES_H_
   7
   8 #include <vector>
   9
  10 #include "base/values.h"
  11 #include "url/gurl.h"
  12
  13 namespace dom_distiller {
  14
  15 // The length of the derived features vector.
  16 extern int kDerivedFeaturesCount;
  17
  18 // The distillable page detector is a model trained on a list of numeric
  19 // features derived from core more complex features of a webpage (like the
  20 // body's .textContent). This derives the numeric features for a set of core
  21 // features.
  22 //
  23 // Note: It is crucial that these features are derived in the same way and are
  24 // in the same order as in the training pipeline. See //heuristics/distillable
  25 // in the external DomDistillerJs repo.
  26 std::vector<double> CalculateDerivedFeatures(bool isOGArticle,
  27                                              const GURL& url,
  28                                              double numElements,
  29                                              double numAnchors,
  30                                              double numForms,
  31                                              const std::string& innerText,
  32                                              const std::string& textContent,
  33                                              const std::string& innerHTML);
  34
  35 // Calculates the derived features from the JSON value as returned by the
  36 // javascript core feature extraction.
  37 std::vector<double> CalculateDerivedFeaturesFromJSON(
  38     const base::Value* stringified_json);
  39
  40 }  // namespace dom_distiller
  41
  42 #endif  // COMPONENTS_DOM_DISTILLER_CORE_PAGE_FEATURES_H_