NTPSnippetsService: Garbage collect orphaned images at startup.

components/ntp_snippets/remote/ntp_snippets_service.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/ntp_snippets/remote/ntp_snippets_service.h"
 
 #include <algorithm>
 #include <iterator>
 #include <utility>
 
 #include "base/command_line.h"
 #include "base/location.h"
+#include "base/memory/ptr_util.h"
 #include "base/metrics/histogram_macros.h"
 #include "base/metrics/sparse_histogram.h"
 #include "base/path_service.h"
 #include "base/stl_util.h"
 #include "base/strings/string_number_conversions.h"
 #include "base/strings/utf_string_conversions.h"
 #include "base/task_runner_util.h"
 #include "base/time/time.h"
 #include "base/values.h"
 #include "components/data_use_measurement/core/data_use_user_data.h"
@@ skipping 99 matching lines @@
 std::set<std::string> GetAllIDs(const NTPSnippet::PtrVector& snippets) {
   std::set<std::string> ids;
   for (const std::unique_ptr<NTPSnippet>& snippet : snippets) {
     ids.insert(snippet->id());
     for (const SnippetSource& source : snippet->sources())
       ids.insert(source.url.spec());
   }
   return ids;
 }
 
-std::set<std::string> GetMainIDs(const NTPSnippet::PtrVector& snippets) {
+std::set<std::string> GetSnippetIDSet(const NTPSnippet::PtrVector& snippets) {
   std::set<std::string> ids;
   for (const std::unique_ptr<NTPSnippet>& snippet : snippets)
     ids.insert(snippet->id());
   return ids;
 }
 
+std::unique_ptr<std::vector<std::string>> GetSnippetIDVector(
+    const NTPSnippet::PtrVector& snippets) {
+  auto result = base::MakeUnique<std::vector<std::string>>();
+  for (const auto& snippet : snippets) {
+    result->push_back(snippet->id());
+  }
+  return result;
+}
+
 bool IsSnippetInSet(const std::unique_ptr<NTPSnippet>& snippet,
                     const std::set<std::string>& ids,
                     bool match_all_ids) {
   if (ids.count(snippet->id()))
     return true;
   if (!match_all_ids)
     return false;
   for (const SnippetSource& source : snippet->sources()) {
     if (ids.count(source.url.spec()))
       return true;
@@ skipping 226 matching lines @@
   if (!initialized())
     return;
 
   if (categories_.find(category) == categories_.end())
     return;
   CategoryContent* content = &categories_[category];
   if (content->snippets.empty())
     return;
 
   if (category == articles_category_) {
-    database_->DeleteSnippets(content->snippets);
-    database_->DeleteImages(content->snippets);
+    database_->DeleteSnippets(GetSnippetIDVector(content->snippets));
+    database_->DeleteImages(GetSnippetIDVector(content->snippets));
   }
   content->snippets.clear();
 
   NotifyNewSuggestions();
 }
 
 void NTPSnippetsService::GetDismissedSuggestionsForDebugging(
     Category category,
     const DismissedSuggestionsCallback& callback) {
   DCHECK(categories_.find(category) != categories_.end());
@@ skipping 23 matching lines @@
 
   if (!initialized())
     return;
 
   CategoryContent* content = &categories_[category];
   if (content->dismissed.empty())
     return;
 
   if (category == articles_category_) {
     // The image got already deleted when the suggestion was dismissed.
-    database_->DeleteSnippets(content->dismissed);
+    database_->DeleteSnippets(GetSnippetIDVector(content->dismissed));
   }
   content->dismissed.clear();
 }
 
 std::set<std::string> NTPSnippetsService::GetSuggestionsHosts() const {
   // |suggestions_service_| can be null in tests.
   if (!suggestions_service_)
     return std::set<std::string>();
 
   // TODO(treib): This should just call GetSnippetHostsFromPrefs.
@@ skipping 59 matching lines @@
     else
       content->snippets.emplace_back(std::move(snippet));
   }
 
   std::sort(content->snippets.begin(), content->snippets.end(),
             [](const std::unique_ptr<NTPSnippet>& lhs,
                const std::unique_ptr<NTPSnippet>& rhs) {
               return lhs->score() > rhs->score();
             });
 
+  // TODO(tschumann): If I move ClearExpiredDismisedSnippets() to the beginning
+  // of the function, it essentially does nothing but tests are still green. Fix
+  // this!
   ClearExpiredDismissedSnippets();
   ClearOrphanedImages();
   FinishInitialization();
 }
 
 void NTPSnippetsService::OnDatabaseError() {
   EnterState(State::ERROR_OCCURRED);
   UpdateAllCategoryStatus(CategoryStatus::LOADING_ERROR);
 }
 
@@ skipping 66 matching lines @@
       categories_[category].provided_by_server = true;
 
       DCHECK_LE(snippets->size(), static_cast<size_t>(kMaxSnippetCount));
       // TODO(sfiera): histograms for server categories.
       // Sparse histogram used because the number of snippets is small (bound by
       // kMaxSnippetCount).
       if (category == articles_category_) {
         UMA_HISTOGRAM_SPARSE_SLOWLY("NewTabPage.Snippets.NumArticlesFetched",
                                     fetched_category.snippets.size());
       }
-
       ReplaceSnippets(category, std::move(fetched_category.snippets));
     }
   }
 
   for (const auto& item : categories_) {
     Category category = item.first;
     UpdateCategoryStatus(category, CategoryStatus::AVAILABLE);
   }
 
   // TODO(sfiera): equivalent metrics for non-articles.
@@ skipping 15 matching lines @@
   if (snippets)
     RescheduleFetching(true);
 }
 
 void NTPSnippetsService::ArchiveSnippets(Category category,
                                          NTPSnippet::PtrVector* to_archive) {
   CategoryContent* content = &categories_[category];
 
   // TODO(sfiera): handle DB for non-articles too.
   if (category == articles_category_) {
-    database_->DeleteSnippets(*to_archive);
+    database_->DeleteSnippets(GetSnippetIDVector(*to_archive));
     // Do not delete the thumbnail images as they are still handy on open NTPs.
   }
 
   // Archive previous snippets - move them at the beginning of the list.
   content->archived.insert(content->archived.begin(),
                            std::make_move_iterator(to_archive->begin()),
                            std::make_move_iterator(to_archive->end()));
   Compact(to_archive);
 
   // If there are more archived snippets than we want to keep, delete the
   // oldest ones by their fetch time (which are always in the back).
   if (content->archived.size() > kMaxArchivedSnippetCount) {
     NTPSnippet::PtrVector to_delete(
         std::make_move_iterator(content->archived.begin() +
                                 kMaxArchivedSnippetCount),
         std::make_move_iterator(content->archived.end()));
     content->archived.resize(kMaxArchivedSnippetCount);
     if (category == articles_category_)
-      database_->DeleteImages(to_delete);
+      database_->DeleteImages(GetSnippetIDVector(to_delete));
   }
 }
 
 void NTPSnippetsService::ReplaceSnippets(Category category,
                                          NTPSnippet::PtrVector new_snippets) {
   DCHECK(ready());
   CategoryContent* content = &categories_[category];
 
   // Remove new snippets that have been dismissed.
   EraseMatchingSnippets(&new_snippets, GetAllIDs(content->dismissed),
@@ skipping 29 matching lines @@
                                   num_snippets_dismissed);
     }
   }
 
   // Do not touch the current set of snippets if the newly fetched one is empty.
   if (new_snippets.empty())
     return;
 
   // Remove current snippets that have been fetched again. We do not need to
   // archive those as they will be in the new current set.
-  EraseMatchingSnippets(&content->snippets, GetMainIDs(new_snippets),
+  EraseMatchingSnippets(&content->snippets, GetSnippetIDSet(new_snippets),
                         /*match_all_ids=*/false);
 
   ArchiveSnippets(category, &content->snippets);
 
   // TODO(sfiera): handle DB for non-articles too.
   if (category == articles_category_) {
     // Save new articles to the DB.
     database_->SaveSnippets(new_snippets);
   }
 
@@ skipping 33 matching lines @@
     // Move expired dismissed snippets over into |to_delete|.
     for (std::unique_ptr<NTPSnippet>& snippet : content->dismissed) {
       if (snippet->expiry_date() <= now)
         to_delete.emplace_back(std::move(snippet));
     }
     Compact(&content->dismissed);
 
     // Delete the removed article suggestions from the DB.
     if (category == articles_category_) {
       // The image got already deleted when the suggestion was dismissed.
-      database_->DeleteSnippets(to_delete);
+      database_->DeleteSnippets(GetSnippetIDVector(to_delete));
     }
 
     if (content->snippets.empty() && content->dismissed.empty() &&
         category != articles_category_ && !content->provided_by_server) {
       categories_to_erase.push_back(category);
     }
   }
 
   for (Category category : categories_to_erase) {
     UpdateCategoryStatus(category, CategoryStatus::NOT_PROVIDED);
     categories_.erase(category);
   }
 }
 
 void NTPSnippetsService::ClearOrphanedImages() {
-  // TODO(jkrcal): Implement. crbug.com/649009
+  auto alive_snippets = base::MakeUnique<std::set<std::string>>();
+  for (const auto& snippet_ptr : categories_[articles_category_].snippets) {
+    alive_snippets->insert(snippet_ptr->id());
+  }
+  for (const auto& snippet_ptr : categories_[articles_category_].dismissed) {
+    alive_snippets->insert(snippet_ptr->id());
+  }
+  database_->GarbageCollectImages(std::move(alive_snippets));
 }
 
 void NTPSnippetsService::NukeAllSnippets() {
   std::vector<Category> categories_to_erase;
 
   // Empty the ARTICLES category and remove all others, since they may or may
   // not be personalized.
   for (const auto& item : categories_) {
     Category category = item.first;
 
@@ skipping 302 matching lines @@
 }
 
 NTPSnippetsService::CategoryContent::CategoryContent() = default;
 NTPSnippetsService::CategoryContent::CategoryContent(CategoryContent&&) =
     default;
 NTPSnippetsService::CategoryContent::~CategoryContent() = default;
 NTPSnippetsService::CategoryContent& NTPSnippetsService::CategoryContent::
 operator=(CategoryContent&&) = default;
 
 }  // namespace ntp_snippets