Omnibox - HistoryQuick Provider - Don't Run Duplicate Searches

components/omnibox/browser/url_index_private_data.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 "components/omnibox/browser/url_index_private_data.h"
 
 #include <stdint.h>
 
 #include <functional>
 #include <iterator>
 #include <limits>
 #include <numeric>
+#include <set>
 #include <string>
 #include <vector>
 
 #include "base/files/file_util.h"
 #include "base/i18n/break_iterator.h"
 #include "base/i18n/case_conversion.h"
 #include "base/macros.h"
 #include "base/metrics/histogram_macros.h"
 #include "base/strings/string_split.h"
 #include "base/strings/string_util.h"
@@ skipping 125 matching lines @@
     bookmarks::BookmarkModel* bookmark_model,
     TemplateURLService* template_url_service) {
   // This list will contain the original search string and any other string
   // transformations.
   String16Vector search_strings;
   search_strings.push_back(original_search_string);
   if ((cursor_position != base::string16::npos) &&
       (cursor_position < original_search_string.length()) &&
       (cursor_position > 0)) {
     // The original search_string broken at cursor position. This is one type of
-    // transformation.
+    // transformation.  It's possible this transformation doesn't actually
+    // break any words.  There's no harm in adding the transformation in this
+    // case because the searching code below prevents running duplicate
+    // searches.
     base::string16 transformed_search_string(original_search_string);
     transformed_search_string.insert(cursor_position, base::ASCIIToUTF16(" "));
     search_strings.push_back(transformed_search_string);
   }
 
   ScoredHistoryMatches scored_items;
   // Invalidate the term cache and return if we have indexed no words (probably
   // because we've not been initialized yet).
   if (word_list_.empty()) {
     search_term_cache_.clear();
     return scored_items;
   }
   // Reset used_ flags for search_term_cache_. We use a basic mark-and-sweep
   // approach.
   ResetSearchTermCache();
 
   bool history_ids_were_trimmed = false;
+  // A set containing the list of words extracted from each search string,
+  // used to prevent running duplicate searches.
+  std::set<String16Vector> search_string_words;
   for (const base::string16& search_string : search_strings) {
     // The search string we receive may contain escaped characters. For reducing
     // the index we need individual, lower-cased words, ignoring escapings. For
     // the final filtering we need whitespace separated substrings possibly
     // containing escaped characters.
     base::string16 lower_raw_string(base::i18n::ToLower(search_string));
     base::string16 lower_unescaped_string = net::UnescapeURLComponent(
         lower_raw_string,
         net::UnescapeRule::SPACES | net::UnescapeRule::PATH_SEPARATORS |
             net::UnescapeRule::URL_SPECIAL_CHARS_EXCEPT_PATH_SEPARATORS);
 
     // Extract individual 'words' (as opposed to 'terms'; see comment in
     // HistoryIdsToScoredMatches()) from the search string. When the user types
     // "colspec=ID%20Mstone Release" we get four 'words': "colspec", "id",
     // "mstone" and "release".
     String16Vector lower_words(
         String16VectorFromString16(lower_unescaped_string, false, nullptr));
     if (lower_words.empty())
       continue;
+    // If we've already searched for this list of words, don't do it again.
+    if (search_string_words.find(lower_words) != search_string_words.end())
+      continue;
+    search_string_words.insert(lower_words);
 
     HistoryIDVector history_ids = HistoryIDsFromWords(lower_words);
     history_ids_were_trimmed |= TrimHistoryIdsPool(&history_ids);
 
     HistoryIdsToScoredMatches(std::move(history_ids), lower_raw_string,
                               template_url_service, bookmark_model,
                               &scored_items);
   }
   // Select and sort only the top |max_matches| results.
   if (scored_items.size() > max_matches) {
@@ skipping 1061 matching lines @@
   // First cut: typed count, visit count, recency.
   // TODO(mrossetti): This is too simplistic. Consider an approach which ranks
   // recently visited (within the last 12/24 hours) as highly important. Get
   // input from mpearson.
   if (r1.typed_count() != r2.typed_count())
     return (r1.typed_count() > r2.typed_count());
   if (r1.visit_count() != r2.visit_count())
     return (r1.visit_count() > r2.visit_count());
   return (r1.last_visit() > r2.last_visit());
 }