predictors: Add origin learning.

chrome/browser/predictors/resource_prefetch_predictor.h

Index: chrome/browser/predictors/resource_prefetch_predictor.h
diff --git a/chrome/browser/predictors/resource_prefetch_predictor.h b/chrome/browser/predictors/resource_prefetch_predictor.h
index cec2207d48f5d4d8e62642162e41a75cbbbdb7e7..e2bc592928d42e90ddc3d9c55a1ca77ed0324ecf 100644
--- a/chrome/browser/predictors/resource_prefetch_predictor.h
+++ b/chrome/browser/predictors/resource_prefetch_predictor.h
@@ -104,6 +104,17 @@ class ResourcePrefetchPredictor
       public base::SupportsWeakPtr<ResourcePrefetchPredictor>,
       public precache::PrecacheManager::Delegate {
  public:
+  struct OriginRequestSummary {

[alexilin, 2017/04/10 14:58:28]

Do we really need this struct?
Can't we derive the summary from a vector of URLRequestSummary?

[Benoit L, 2017/04/11 09:04:05]

You're right, this is not strictly necessary.
This is mostly here for readability, and also because we do not store all the
matching URLRequestSummary. For instance, redirects and no-store resources.

I would prefer to keep it, if you don't mind.
wdyt?

[alexilin, 2017/04/11 09:42:19]

Sgtm. Anyway, we would have to reconstruct this map in LearnOrigins().

+    OriginRequestSummary();
+    OriginRequestSummary(const OriginRequestSummary& other);
+    ~OriginRequestSummary();
+
+    GURL origin;
+    bool always_access_network;
+    bool accessed_network;
+    int first_occurrence;
+  };
+
   // Stores the data that we need to get from the URLRequest.
   struct URLRequestSummary {
     URLRequestSummary();
@@ -122,6 +133,8 @@ class ResourcePrefetchPredictor
 
     bool has_validators;
     bool always_revalidate;
+    bool is_no_store;
+    bool network_accessed;
 
     // Initializes a |URLRequestSummary| from a |URLRequest| response.
     // Returns true for success. Note: NavigationID is NOT initialized
@@ -142,6 +155,9 @@ class ResourcePrefetchPredictor
     // Stores all subresource requests within a single navigation, from initial
     // main frame request to navigation completion.
     std::vector<URLRequestSummary> subresource_requests;
+    // Map of origin -> OriginRequestSummary. Only one instance of each origin
+    // is kept per navigation, but the summary is updated several times.
+    std::map<GURL, OriginRequestSummary> origins;
   };
 
   // Stores a result of prediction. Essentially, |subresource_urls| is main
@@ -280,6 +296,7 @@ class ResourcePrefetchPredictor
   typedef ResourcePrefetchPredictorTables::PrefetchDataMap PrefetchDataMap;
   typedef ResourcePrefetchPredictorTables::RedirectDataMap RedirectDataMap;
   typedef ResourcePrefetchPredictorTables::ManifestDataMap ManifestDataMap;
+  typedef ResourcePrefetchPredictorTables::OriginDataMap OriginDataMap;
 
   typedef std::map<NavigationID, std::unique_ptr<PageRequestSummary>>
       NavigationMap;
@@ -296,7 +313,7 @@ class ResourcePrefetchPredictor
                                     const std::string& mime_type);
 
   // Returns true if the request (should have a response in it) is "no-store".
-  static bool IsNoStore(const net::URLRequest* request);
+  static bool IsNoStore(const net::URLRequest& request);
 
   // Returns true iff |redirect_data_map| contains confident redirect endpoint
   // for |entry_point| and assigns it to the |redirect_endpoint|.
@@ -315,6 +332,7 @@ class ResourcePrefetchPredictor
   void OnMainFrameResponse(const URLRequestSummary& response);
   void OnMainFrameRedirect(const URLRequestSummary& response);
   void OnSubresourceResponse(const URLRequestSummary& response);
+  void OnSubresourceRedirect(const URLRequestSummary& response);
 
   // Called when onload completes for a navigation. We treat this point as the
   // "completion" of the navigation. The resources requested by the page up to
@@ -341,7 +359,8 @@ class ResourcePrefetchPredictor
                     std::unique_ptr<PrefetchDataMap> host_data_map,
                     std::unique_ptr<RedirectDataMap> url_redirect_data_map,
                     std::unique_ptr<RedirectDataMap> host_redirect_data_map,
-                    std::unique_ptr<ManifestDataMap> manifest_data_map);
+                    std::unique_ptr<ManifestDataMap> manifest_data_map,
+                    std::unique_ptr<OriginDataMap> origin_data_map);
 
   // Called during initialization when history is read and the predictor
   // database has been read.
@@ -371,6 +390,7 @@ class ResourcePrefetchPredictor
                                           RedirectDataMap* data_map);
 
   void RemoveOldestEntryInManifestDataMap(ManifestDataMap* data_map);
+  void RemoveOldestEntryInOriginDataMap(OriginDataMap* data_map);
 
   // Merges resources in |new_resources| into the |data_map| and correspondingly
   // updates the predictor database. Also calls LearnRedirect if relevant.
@@ -390,6 +410,11 @@ class ResourcePrefetchPredictor
                      size_t max_redirect_map_size,
                      RedirectDataMap* redirect_map);
 
+  void LearnOrigins(const std::string& host,
+                    const std::map<GURL, OriginRequestSummary>& summaries,
+                    size_t max_data_map_size,
+                    OriginDataMap* data_map);
+
   // Reports database readiness metric defined as percentage of navigated hosts
   // found in DB for last X entries in history.
   void ReportDatabaseReadiness(const history::TopHostsList& top_hosts) const;
@@ -426,6 +451,7 @@ class ResourcePrefetchPredictor
   std::unique_ptr<RedirectDataMap> url_redirect_table_cache_;
   std::unique_ptr<RedirectDataMap> host_redirect_table_cache_;
   std::unique_ptr<ManifestDataMap> manifest_table_cache_;
+  std::unique_ptr<OriginDataMap> origin_table_cache_;
 
   std::map<GURL, base::TimeTicks> inflight_prefetches_;
   NavigationMap inflight_navigations_;