1 // Copyright (c) 2012 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 CHROME_BROWSER_HISTORY_URL_INDEX_PRIVATE_DATA_H_ 6 #define CHROME_BROWSER_HISTORY_URL_INDEX_PRIVATE_DATA_H_ 7 8 #include <set> 9 #include <string> 10 11 #include "base/files/file_path.h" 12 #include "base/gtest_prod_util.h" 13 #include "base/memory/ref_counted.h" 14 #include "chrome/browser/common/cancelable_request.h" 15 #include "chrome/browser/history/history_service.h" 16 #include "chrome/browser/history/in_memory_url_index_cache.pb.h" 17 #include "chrome/browser/history/in_memory_url_index_types.h" 18 #include "chrome/browser/history/scored_history_match.h" 19 #include "content/public/browser/notification_details.h" 20 21 class BookmarkService; 22 class HistoryQuickProviderTest; 23 24 namespace in_memory_url_index { 25 class InMemoryURLIndexCacheItem; 26 } 27 28 namespace history { 29 30 namespace imui = in_memory_url_index; 31 32 class HistoryDatabase; 33 class InMemoryURLIndex; 34 class RefCountedBool; 35 36 // Current version of the cache file. 37 static const int kCurrentCacheFileVersion = 3; 38 39 // A structure private to InMemoryURLIndex describing its internal data and 40 // providing for restoring, rebuilding and updating that internal data. As 41 // this class is for exclusive use by the InMemoryURLIndex class there should 42 // be no calls from any other class. 43 // 44 // All public member functions are called on the main thread unless otherwise 45 // annotated. 46 class URLIndexPrivateData 47 : public base::RefCountedThreadSafe<URLIndexPrivateData> { 48 public: 49 URLIndexPrivateData(); 50 51 // Given a string16 in |term_string|, scans the history index and returns a 52 // vector with all scored, matching history items. The |term_string| is 53 // broken down into individual terms (words), each of which must occur in the 54 // candidate history item's URL or page title for the item to qualify; 55 // however, the terms do not necessarily have to be adjacent. We 56 // also allow breaking |term_string| at |cursor_position| (if 57 // set). Once we have a set of candidates, they are filtered to ensure 58 // that all |term_string| terms, as separated by whitespace and the 59 // cursor (if set), occur within the candidate's URL or page title. 60 // Scores are then calculated on no more than |kItemsToScoreLimit| 61 // candidates, as the scoring of such a large number of candidates may 62 // cause perceptible typing response delays in the omnibox. This is 63 // likely to occur for short omnibox terms such as 'h' and 'w' which 64 // will be found in nearly all history candidates. Results are sorted by 65 // descending score. The full results set (i.e. beyond the 66 // |kItemsToScoreLimit| limit) will be retained and used for subsequent calls 67 // to this function. |bookmark_service| is used to boost a result's score if 68 // its URL is referenced by one or more of the user's bookmarks. |languages| 69 // is used to help parse/format the URLs in the history index. 70 ScoredHistoryMatches HistoryItemsForTerms(string16 term_string, 71 size_t cursor_position, 72 const std::string& languages, 73 BookmarkService* bookmark_service); 74 75 // Adds the history item in |row| to the index if it does not already already 76 // exist and it meets the minimum 'quick' criteria. If the row already exists 77 // in the index then the index will be updated if the row still meets the 78 // criteria, otherwise the row will be removed from the index. Returns true 79 // if the index was actually updated. |languages| gives a list of language 80 // encodings by which the URLs and page titles are broken down into words and 81 // characters. |scheme_whitelist| is used to filter non-qualifying schemes. 82 // |history_service| is used to schedule an update to the recent visits 83 // component of this URL's entry in the index. 84 bool UpdateURL(HistoryService* history_service, 85 const URLRow& row, 86 const std::string& languages, 87 const std::set<std::string>& scheme_whitelist); 88 89 // Updates the entry for |url_id| in the index, replacing its 90 // recent visits information with |recent_visits|. If |url_id| 91 // is not in the index, does nothing. 92 void UpdateRecentVisits(URLID url_id, 93 const VisitVector& recent_visits); 94 95 // Using |history_service| schedules an update (using the historyDB 96 // thread) for the recent visits information for |url_id|. Unless 97 // something unexpectedly goes wrong, UdpateRecentVisits() should 98 // eventually be called from a callback. 99 void ScheduleUpdateRecentVisits(HistoryService* history_service, 100 URLID url_id); 101 102 // Deletes index data for the history item with the given |url|. 103 // The item may not have actually been indexed, which is the case if it did 104 // not previously meet minimum 'quick' criteria. Returns true if the index 105 // was actually updated. 106 bool DeleteURL(const GURL& url); 107 108 // Constructs a new object by restoring its contents from the cache file 109 // at |path|. Returns the new URLIndexPrivateData which on success will 110 // contain the restored data but upon failure will be empty. |languages| 111 // is used to break URLs and page titles into words. This function 112 // should be run on the the file thread. 113 static scoped_refptr<URLIndexPrivateData> RestoreFromFile( 114 const base::FilePath& path, 115 const std::string& languages); 116 117 // Constructs a new object by rebuilding its contents from the history 118 // database in |history_db|. Returns the new URLIndexPrivateData which on 119 // success will contain the rebuilt data but upon failure will be empty. 120 // |languages| gives a list of language encodings by which the URLs and page 121 // titles are broken down into words and characters. 122 static scoped_refptr<URLIndexPrivateData> RebuildFromHistory( 123 HistoryDatabase* history_db, 124 const std::string& languages, 125 const std::set<std::string>& scheme_whitelist); 126 127 // Writes |private_data| as a cache file to |file_path| and returns success. 128 static bool WritePrivateDataToCacheFileTask( 129 scoped_refptr<URLIndexPrivateData> private_data, 130 const base::FilePath& file_path); 131 132 // Stops all pending updates to recent visits fields. This should be 133 // called during shutdown. 134 void CancelPendingUpdates(); 135 136 // Creates a copy of ourself. 137 scoped_refptr<URLIndexPrivateData> Duplicate() const; 138 139 // Returns true if there is no data in the index. 140 bool Empty() const; 141 142 // Initializes all index data members in preparation for restoring the index 143 // from the cache or a complete rebuild from the history database. 144 void Clear(); 145 146 private: 147 friend class base::RefCountedThreadSafe<URLIndexPrivateData>; 148 ~URLIndexPrivateData(); 149 150 friend class AddHistoryMatch; 151 friend class ::HistoryQuickProviderTest; 152 friend class InMemoryURLIndexTest; 153 FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, CacheSaveRestore); 154 FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, HugeResultSet); 155 FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, ReadVisitsFromHistory); 156 FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, RebuildFromHistoryIfCacheOld); 157 FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, Scoring); 158 FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, TitleSearch); 159 FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, TypedCharacterCaching); 160 FRIEND_TEST_ALL_PREFIXES(InMemoryURLIndexTest, WhitelistedURLs); 161 FRIEND_TEST_ALL_PREFIXES(LimitedInMemoryURLIndexTest, Initialization); 162 163 // Support caching of term results so that we can optimize searches which 164 // build upon a previous search. Each entry in this map represents one 165 // search term from the most recent search. For example, if the user had 166 // typed "google blog trans" and then typed an additional 'l' (at the end, 167 // of course) then there would be four items in the cache: 'blog', 'google', 168 // 'trans', and 'transl'. All would be marked as being in use except for the 169 // 'trans' item; its cached data would have been used when optimizing the 170 // construction of the search results candidates for 'transl' but then would 171 // no longer needed. 172 // 173 // Items stored in the search term cache. If a search term exactly matches one 174 // in the cache then we can quickly supply the proper |history_id_set_| (and 175 // marking the cache item as being |used_|. If we find a prefix for a search 176 // term in the cache (which is very likely to occur as the user types each 177 // term into the omnibox) then we can short-circuit the index search for those 178 // characters in the prefix by returning the |word_id_set|. In that case we do 179 // not mark the item as being |used_|. 180 struct SearchTermCacheItem { 181 SearchTermCacheItem(const WordIDSet& word_id_set, 182 const HistoryIDSet& history_id_set); 183 // Creates a cache item for a term which has no results. 184 SearchTermCacheItem(); 185 186 ~SearchTermCacheItem(); 187 188 WordIDSet word_id_set_; 189 HistoryIDSet history_id_set_; 190 bool used_; // True if this item has been used for the current term search. 191 }; 192 typedef std::map<string16, SearchTermCacheItem> SearchTermCacheMap; 193 194 // A helper class which performs the final filter on each candidate 195 // history URL match, inserting accepted matches into |scored_matches_|. 196 class AddHistoryMatch : public std::unary_function<HistoryID, void> { 197 public: 198 AddHistoryMatch(const URLIndexPrivateData& private_data, 199 const std::string& languages, 200 BookmarkService* bookmark_service, 201 const string16& lower_string, 202 const String16Vector& lower_terms, 203 const base::Time now); 204 ~AddHistoryMatch(); 205 206 void operator()(const HistoryID history_id); 207 208 ScoredHistoryMatches ScoredMatches() const { return scored_matches_; } 209 210 private: 211 const URLIndexPrivateData& private_data_; 212 const std::string& languages_; 213 BookmarkService* bookmark_service_; 214 ScoredHistoryMatches scored_matches_; 215 const string16& lower_string_; 216 const String16Vector& lower_terms_; 217 const base::Time now_; 218 }; 219 220 // A helper predicate class used to filter excess history items when the 221 // candidate results set is too large. 222 class HistoryItemFactorGreater 223 : public std::binary_function<HistoryID, HistoryID, void> { 224 public: 225 explicit HistoryItemFactorGreater(const HistoryInfoMap& history_info_map); 226 ~HistoryItemFactorGreater(); 227 228 bool operator()(const HistoryID h1, const HistoryID h2); 229 230 private: 231 const history::HistoryInfoMap& history_info_map_; 232 }; 233 234 // URL History indexing support functions. 235 236 // Composes a set of history item IDs by intersecting the set for each word 237 // in |unsorted_words|. 238 HistoryIDSet HistoryIDSetFromWords(const String16Vector& unsorted_words); 239 240 // Helper function to HistoryIDSetFromWords which composes a set of history 241 // ids for the given term given in |term|. 242 HistoryIDSet HistoryIDsForTerm(const string16& term); 243 244 // Given a set of Char16s, finds words containing those characters. 245 WordIDSet WordIDSetForTermChars(const Char16Set& term_chars); 246 247 // Indexes one URL history item as described by |row|. Returns true if the 248 // row was actually indexed. |languages| gives a list of language encodings by 249 // which the URLs and page titles are broken down into words and characters. 250 // |scheme_whitelist| is used to filter non-qualifying schemes. If 251 // |history_db| is not NULL then this function uses the history database 252 // synchronously to get the URL's recent visits information. This mode should 253 // only be used on the historyDB thread. If |history_db| is NULL, then 254 // this function uses |history_service| to schedule a task on the 255 // historyDB thread to fetch and update the recent visits 256 // information. 257 bool IndexRow(HistoryDatabase* history_db, 258 HistoryService* history_service, 259 const URLRow& row, 260 const std::string& languages, 261 const std::set<std::string>& scheme_whitelist); 262 263 // Parses and indexes the words in the URL and page title of |row| and 264 // calculate the word starts in each, saving the starts in |word_starts|. 265 // |languages| gives a list of language encodings by which the URLs and page 266 // titles are broken down into words and characters. 267 void AddRowWordsToIndex(const URLRow& row, 268 RowWordStarts* word_starts, 269 const std::string& languages); 270 271 // Given a single word in |uni_word|, adds a reference for the containing 272 // history item identified by |history_id| to the index. 273 void AddWordToIndex(const string16& uni_word, HistoryID history_id); 274 275 // Creates a new entry in the word/history map for |word_id| and add 276 // |history_id| as the initial element of the word's set. 277 void AddWordHistory(const string16& uni_word, HistoryID history_id); 278 279 // Updates an existing entry in the word/history index by adding the 280 // |history_id| to set for |word_id| in the word_id_history_map_. 281 void UpdateWordHistory(WordID word_id, HistoryID history_id); 282 283 // Adds |word_id| to |history_id|'s entry in the history/word map, 284 // creating a new entry if one does not already exist. 285 void AddToHistoryIDWordMap(HistoryID history_id, WordID word_id); 286 287 // Removes |row| and all associated words and characters from the index. 288 void RemoveRowFromIndex(const URLRow& row); 289 290 // Removes all words and characters associated with |row| from the index. 291 void RemoveRowWordsFromIndex(const URLRow& row); 292 293 // Clears |used_| for each item in the search term cache. 294 void ResetSearchTermCache(); 295 296 // Caches the index private data and writes the cache file to the profile 297 // directory. Called by WritePrivateDataToCacheFileTask. 298 bool SaveToFile(const base::FilePath& file_path); 299 300 // Encode a data structure into the protobuf |cache|. 301 void SavePrivateData(imui::InMemoryURLIndexCacheItem* cache) const; 302 void SaveWordList(imui::InMemoryURLIndexCacheItem* cache) const; 303 void SaveWordMap(imui::InMemoryURLIndexCacheItem* cache) const; 304 void SaveCharWordMap(imui::InMemoryURLIndexCacheItem* cache) const; 305 void SaveWordIDHistoryMap(imui::InMemoryURLIndexCacheItem* cache) const; 306 void SaveHistoryInfoMap(imui::InMemoryURLIndexCacheItem* cache) const; 307 void SaveWordStartsMap(imui::InMemoryURLIndexCacheItem* cache) const; 308 309 // Decode a data structure from the protobuf |cache|. Return false if there 310 // is any kind of failure. |languages| will be used to break URLs and page 311 // titles into words 312 bool RestorePrivateData(const imui::InMemoryURLIndexCacheItem& cache, 313 const std::string& languages); 314 bool RestoreWordList(const imui::InMemoryURLIndexCacheItem& cache); 315 bool RestoreWordMap(const imui::InMemoryURLIndexCacheItem& cache); 316 bool RestoreCharWordMap(const imui::InMemoryURLIndexCacheItem& cache); 317 bool RestoreWordIDHistoryMap(const imui::InMemoryURLIndexCacheItem& cache); 318 bool RestoreHistoryInfoMap(const imui::InMemoryURLIndexCacheItem& cache); 319 bool RestoreWordStartsMap(const imui::InMemoryURLIndexCacheItem& cache, 320 const std::string& languages); 321 322 // Determines if |gurl| has a whitelisted scheme and returns true if so. 323 static bool URLSchemeIsWhitelisted(const GURL& gurl, 324 const std::set<std::string>& whitelist); 325 326 // Cache of search terms. 327 SearchTermCacheMap search_term_cache_; 328 329 // Allows canceling pending requests to update recent visits information. 330 CancelableRequestConsumer recent_visits_consumer_; 331 332 // Start of data members that are cached ------------------------------------- 333 334 // The version of the cache file most recently used to restore this instance 335 // of the private data. If the private data was rebuilt from the history 336 // database this will be 0. 337 int restored_cache_version_; 338 339 // The last time the data was rebuilt from the history database. 340 base::Time last_time_rebuilt_from_history_; 341 342 // A list of all of indexed words. The index of a word in this list is the 343 // ID of the word in the word_map_. It reduces the memory overhead by 344 // replacing a potentially long and repeated string with a simple index. 345 String16Vector word_list_; 346 347 // A list of available words slots in |word_list_|. An available word slot 348 // is the index of a unused word in word_list_ vector, also referred to as 349 // a WordID. As URL visits are added or modified new words may be added to 350 // the index, in which case any available words are used, if any, and then 351 // words are added to the end of the word_list_. When URL visits are 352 // modified or deleted old words may be removed from the index, in which 353 // case the slots for those words are added to available_words_ for resuse 354 // by future URL updates. 355 WordIDSet available_words_; 356 357 // A one-to-one mapping from the a word string to its slot number (i.e. 358 // WordID) in the |word_list_|. 359 WordMap word_map_; 360 361 // A one-to-many mapping from a single character to all WordIDs of words 362 // containing that character. 363 CharWordIDMap char_word_map_; 364 365 // A one-to-many mapping from a WordID to all HistoryIDs (the row_id as 366 // used in the history database) of history items in which the word occurs. 367 WordIDHistoryMap word_id_history_map_; 368 369 // A one-to-many mapping from a HistoryID to all WordIDs of words that occur 370 // in the URL and/or page title of the history item referenced by that 371 // HistoryID. 372 HistoryIDWordMap history_id_word_map_; 373 374 // A one-to-one mapping from HistoryID to the history item data governing 375 // index inclusion and relevance scoring. 376 HistoryInfoMap history_info_map_; 377 378 // A one-to-one mapping from HistoryID to the word starts detected in each 379 // item's URL and page title. 380 WordStartsMap word_starts_map_; 381 382 // End of data members that are cached --------------------------------------- 383 384 // For unit testing only. Specifies the version of the cache file to be saved. 385 // Used only for testing upgrading of an older version of the cache upon 386 // restore. 387 int saved_cache_version_; 388 389 // Used for unit testing only. Records the number of candidate history items 390 // at three stages in the index searching process. 391 size_t pre_filter_item_count_; // After word index is queried. 392 size_t post_filter_item_count_; // After trimming large result set. 393 size_t post_scoring_item_count_; // After performing final filter/scoring. 394 }; 395 396 } // namespace history 397 398 #endif // CHROME_BROWSER_HISTORY_URL_INDEX_PRIVATE_DATA_H_ 399