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 distillable page detector is a model trained on a list of numeric
  16 // features derived from core more complex features of a webpage (like the
  17 // body's .textContent). This derives the numeric features for a set of core
  18 // features.
  19 //
  20 // Note: It is crucial that these features are derived in the same way and are
  21 // in the same order as in the training pipeline. See //heuristics/distillable
  22 // in the external DomDistillerJs repo.
  23 std::vector<double> CalculateDerivedFeatures(bool isOGArticle,
  24                                              const GURL& url,
  25                                              double numElements,
  26                                              double numAnchors,
  27                                              double numForms,
  28                                              const std::string& innerText,
  29                                              const std::string& textContent,
  30                                              const std::string& innerHTML);
  31
  32 // Calculates the derived features from the JSON value as returned by the
  33 // javascript core feature extraction.
  34 std::vector<double> CalculateDerivedFeaturesFromJSON(const base::Value* json);
  35
  36 }  // namespace dom_distiller
  37
  38 #endif  // COMPONENTS_DOM_DISTILLER_CORE_PAGE_FEATURES_H_