Track fraction of visits to top URLs.

chrome/browser/history/history_backend.cc

 // Copyright (c) 2012 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 "chrome/browser/history/history_backend.h"
 
 #include <algorithm>
 #include <functional>
 #include <list>
 #include <map>
@@ skipping 36 matching lines @@
 #include "sql/error_delegate_util.h"
 
 #if defined(OS_ANDROID)
 #include "chrome/browser/history/android/android_provider_backend.h"
 #endif
 
 using base::Time;
 using base::TimeDelta;
 using base::TimeTicks;
 
+#if defined(OS_ANDROID)
+namespace {
+
+// The maximum number of top sites to track when recording top page visit stats.
+const size_t kPageVisitStatsMaxTopSites = 50;
+
+// Populate a map from a |MostVisitedURLList|. The map assigns a rank to each
+// top URL and its redirects. This should only be done once at backend
+// initialization.
+void PopulateMostVisitedURLMap(const history::MostVisitedURLList& top_urls,
+                               std::map<GURL, int>& top_urls_map) {

[jar (doing other things), 2013/06/19 00:49:58]

nit: Pass in the pointer to modify the map, rather than a mutable reference.

... or even better... as mentioned... use methods, so you don't need to pass the
slot around to functions.

[bengr, 2013/06/19 17:44:24]

Done.

+  DCHECK_LE(top_urls.size(), kPageVisitStatsMaxTopSites);
+  for (size_t i = 0; i < top_urls.size(); ++i) {
+    top_urls_map[top_urls[i].url] = i;
+    for (size_t j = 0; j < top_urls[i].redirects.size(); ++j)
+      top_urls_map[top_urls[i].redirects[j]] = i;
+  }
+}
+
+// Record counts of page visits by rank. If a url is not ranked, record the
+// page visit in a slot corresponding to |max_top_url_count|, which should
+// be one greater than the largest rank of any url in |top_urls|.
+// This can be removed for M31. (See issue 248761.)
+void RecordTopPageVisitStats(const std::map<GURL, int>& top_urls_map,
+                             const GURL& url) {
+  int rank = kPageVisitStatsMaxTopSites;
+  std::map<GURL, int>::const_iterator it = top_urls_map.find(url);
+  if (it != top_urls_map.end())
+    rank = (*it).second;
+  LOG(WARNING) << "XXX rank " << rank << " url: " << url.spec();

[jar (doing other things), 2013/06/19 00:49:58]

Generally we use DLOG, so that we don't swell up the binary with a pile of text
segments and logging that is never used by folks in the field.

[bengr, 2013/06/19 17:44:24]

Thanks for catching this. It was some logging leftover from debugging. Removed.

On 2013/06/19 00:49:58, jar wrote:

+  UMA_HISTOGRAM_ENUMERATION("History.TopSitesVisitsByRank",
+                            rank, kPageVisitStatsMaxTopSites + 1);
+}
+
+}  // namespace
+#endif
+
 /* The HistoryBackend consists of a number of components:
 
     HistoryDatabase (stores past 3 months of history)
       URLDatabase (stores a list of URLs)
       DownloadDatabase (stores a list of downloads)
       VisitDatabase (stores a list of visits for the URLs)
       VisitSegmentDatabase (stores groups of URLs for the most visited view).
 
     ArchivedDatabase (stores history older than 3 months)
       URLDatabase (stores a list of URLs)
@@ skipping 29 matching lines @@
 static const int kSessionCloseTimeWindowSecs = 10;
 
 // The maximum number of items we'll allow in the redirect list before
 // deleting some.
 static const int kMaxRedirectCount = 32;
 
 // The number of days old a history entry can be before it is considered "old"
 // and is archived.
 static const int kArchiveDaysThreshold = 90;
 
+
 // Converts from PageUsageData to MostVisitedURL. |redirects| is a
 // list of redirects for this URL. Empty list means no redirects.
 MostVisitedURL MakeMostVisitedURL(const PageUsageData& page_data,
                                   const RedirectList& redirects) {
   MostVisitedURL mv;
   mv.url = page_data.GetURL();
   mv.title = page_data.GetTitle();
   if (redirects.empty()) {
     // Redirects must contain at least the target url.
     mv.redirects.push_back(mv.url);
@@ skipping 133 matching lines @@
 #if defined(OS_ANDROID)
   file_util::Delete(GetAndroidCacheFileName(), false);
 #endif
 }
 
 void HistoryBackend::Init(const std::string& languages, bool force_fail) {
   if (!force_fail)
     InitImpl(languages);
   delegate_->DBLoaded(id_);
   typed_url_syncable_service_.reset(new TypedUrlSyncableService(this));
+#if defined(OS_ANDROID)
+  MostVisitedURLList most_visited_urls;
+  QueryMostVisitedURLsImpl(kPageVisitStatsMaxTopSites, kSegmentDataRetention,
+                           &most_visited_urls);
+  PopulateMostVisitedURLMap(most_visited_urls, most_visited_urls_map_);
+#endif
 }
 
 void HistoryBackend::SetOnBackendDestroyTask(base::MessageLoop* message_loop,
                                              const base::Closure& task) {
   if (!backend_destroy_task_.is_null())
     DLOG(WARNING) << "Setting more than one destroy task, overriding";
   backend_destroy_message_loop_ = message_loop;
   backend_destroy_task_ = task;
 }
 
@@ skipping 524 matching lines @@
   // TODO(pkasting): http://b/1148304 We shouldn't be marking so many URLs as
   // typed, which would eliminate the need for this code.
   int typed_increment = 0;
   content::PageTransition transition_type =
       content::PageTransitionStripQualifier(transition);
   if ((transition_type == content::PAGE_TRANSITION_TYPED &&
       !content::PageTransitionIsRedirect(transition)) ||
       transition_type == content::PAGE_TRANSITION_KEYWORD_GENERATED)
     typed_increment = 1;
 
+#if defined(OS_ANDROID)
+  // Only count the page visit if it came from user browsing and only count it
+  // once when cycling through a redirect chain.
+  if (visit_source == SOURCE_BROWSED &&
+      (transition & content::PAGE_TRANSITION_CHAIN_END) != 0) {
+    RecordTopPageVisitStats(most_visited_urls_map_, url);
+  }
+#endif
+
   // See if this URL is already in the DB.
   URLRow url_info(url);
   URLID url_id = db_->GetRowForURL(url, &url_info);
   if (url_id) {
     // Update of an existing row.
     if (content::PageTransitionStripQualifier(transition) !=
         content::PAGE_TRANSITION_RELOAD)
       url_info.set_visit_count(url_info.visit_count() + 1);
     if (typed_increment)
       url_info.set_typed_count(url_info.typed_count() + typed_increment);
@@ skipping 2255 matching lines @@
   info.url_id = visit.url_id;
   info.time = visit.visit_time;
   info.transition = visit.transition;
   // If we don't have a delegate yet during setup or shutdown, we will drop
   // these notifications.
   if (delegate_)
     delegate_->NotifyVisitDBObserversOnAddVisit(info);
 }
 
 }  // namespace history