Add "AllArticles" mode to Reader Mode heuristics

components/dom_distiller/content/renderer/distillability_agent.cc

 // Copyright 2015 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "components/dom_distiller/content/renderer/distillability_agent.h"
 
 #include "base/metrics/histogram_macros.h"
 #include "base/strings/string_util.h"
 #include "components/dom_distiller/content/common/distillability_service.mojom.h"
 #include "components/dom_distiller/core/distillable_page_detector.h"
@@ skipping 23 matching lines @@
   BLACKLISTED,
   TOO_SHORT,
   NOT_REJECTED,
   REJECTION_BUCKET_BOUNDARY
 };
 
 // Returns whether it is necessary to send updates back to the browser.
 // The number of updates can be from 0 to 2. See the tests in
 // "distillable_page_utils_browsertest.cc".
 // Most heuristics types only require one update after parsing.
-// Adaboost is the only one doing the second update, which is after loading.
+// Adaboost-based heuristics are the only ones doing the second update,
+// which is after loading.
 bool NeedToUpdate(bool is_loaded) {
   switch (GetDistillerHeuristicsType()) {
     case DistillerHeuristicsType::ALWAYS_TRUE:
       return !is_loaded;
     case DistillerHeuristicsType::OG_ARTICLE:
       return !is_loaded;
     case DistillerHeuristicsType::ADABOOST_MODEL:
+    case DistillerHeuristicsType::ALL_ARTICLES:
       return true;
     case DistillerHeuristicsType::NONE:
     default:
       return false;
   }
 }
 
 // Returns whether this update is the last one for the page.
 bool IsLast(bool is_loaded) {
-  if (GetDistillerHeuristicsType() == DistillerHeuristicsType::ADABOOST_MODEL)
+  if (GetDistillerHeuristicsType() == DistillerHeuristicsType::ADABOOST_MODEL ||
+      GetDistillerHeuristicsType() == DistillerHeuristicsType::ALL_ARTICLES)
     return is_loaded;
 
   return true;
 }
 
 bool IsBlacklisted(const GURL& url) {
   for (size_t i = 0; i < arraysize(kBlacklist); ++i) {
     if (base::LowerCaseEqualsASCII(url.host(), kBlacklist[i])) {
       return true;
     }
   }
   return false;
 }
 
 bool IsDistillablePageAdaboost(WebDocument& doc,
                                const DistillablePageDetector* detector,
                                const DistillablePageDetector* long_page,
-                               bool is_last) {
+                               bool is_last,
+                               bool exclude_mobile) {
   WebDistillabilityFeatures features = doc.DistillabilityFeatures();
   GURL parsed_url(doc.Url());
   if (!parsed_url.is_valid()) {
     return false;
   }
   std::vector<double> derived = CalculateDerivedFeatures(
       features.open_graph, parsed_url, features.element_count,
       features.anchor_count, features.form_count, features.moz_score,
       features.moz_score_all_sqrt, features.moz_score_all_linear);
   double score = detector->Score(derived) - detector->GetThreshold();
@@ skipping 48 matching lines @@
           TOO_SHORT, REJECTION_BUCKET_BOUNDARY);
     } else {
       UMA_HISTOGRAM_ENUMERATION("DomDistiller.DistillabilityRejection",
           NOT_REJECTED, REJECTION_BUCKET_BOUNDARY);
     }
   }
 
   if (blacklisted) {
     return false;
   }
-  if (features.is_mobile_friendly) {
+  if (exclude_mobile && features.is_mobile_friendly) {
     return false;
   }
   return distillable && long_article;
 }
 
 bool IsDistillablePage(WebDocument& doc, bool is_last) {
   switch (GetDistillerHeuristicsType()) {
     case DistillerHeuristicsType::ALWAYS_TRUE:
       return true;
     case DistillerHeuristicsType::OG_ARTICLE:
       return doc.DistillabilityFeatures().open_graph;
     case DistillerHeuristicsType::ADABOOST_MODEL:
-      return IsDistillablePageAdaboost(doc,
-          DistillablePageDetector::GetNewModel(),
-          DistillablePageDetector::GetLongPageModel(), is_last);
+      return IsDistillablePageAdaboost(
+          doc, DistillablePageDetector::GetNewModel(),
+          DistillablePageDetector::GetLongPageModel(), is_last, true);
+    case DistillerHeuristicsType::ALL_ARTICLES:
+      return IsDistillablePageAdaboost(
+          doc, DistillablePageDetector::GetNewModel(),
+          DistillablePageDetector::GetLongPageModel(), is_last, false);
     case DistillerHeuristicsType::NONE:
     default:
       return false;
   }
 }
 
 }  // namespace
 
 DistillabilityAgent::DistillabilityAgent(
     content::RenderFrame* render_frame)
@@ skipping 30 matching lines @@
       IsDistillablePage(doc, is_last), is_last);
 }
 
 DistillabilityAgent::~DistillabilityAgent() {}
 
 void DistillabilityAgent::OnDestruct() {
   delete this;
 }
 
 }  // namespace dom_distiller