OLD | NEW |
1 // Copyright (c) 2010 The Chromium Authors. All rights reserved. | 1 // Copyright (c) 2010 The Chromium Authors. All rights reserved. |
2 // Use of this source code is governed by a BSD-style license that can be | 2 // Use of this source code is governed by a BSD-style license that can be |
3 // found in the LICENSE file. | 3 // found in the LICENSE file. |
4 | 4 |
5 #include "chrome/browser/history/in_memory_url_index.h" | 5 #include "chrome/browser/history/in_memory_url_index.h" |
6 | 6 |
7 #include <algorithm> | |
8 #include <limits> | |
9 | |
10 #include "app/l10n_util.h" | |
11 #include "base/i18n/word_iterator.h" | |
12 #include "base/string_util.h" | |
13 #include "base/time.h" | |
14 #include "base/utf_string_conversions.h" | |
15 #include "chrome/browser/history/url_database.h" | 7 #include "chrome/browser/history/url_database.h" |
16 #include "net/base/escape.h" | |
17 #include "net/base/net_util.h" | |
18 | |
19 using base::Time; | |
20 using base::TimeDelta; | |
21 | 8 |
22 namespace history { | 9 namespace history { |
23 | 10 |
24 InMemoryURLIndex::InMemoryURLIndex() : history_item_count_(0) {} | |
25 | |
26 InMemoryURLIndex::~InMemoryURLIndex() {} | |
27 | |
28 // Indexing | 11 // Indexing |
29 | 12 |
30 bool InMemoryURLIndex::Init(history::URLDatabase* history_db, | 13 bool InMemoryURLIndex::Init(history::URLDatabase* history_db) { |
31 const string16& languages) { | 14 bool success = true; |
32 // TODO(mrossetti): Register for profile/language change notifications. | 15 // TODO(mrossetti): Implement. |
33 languages_ = languages; | 16 return success; |
34 // Reset our indexes. | |
35 char_word_map_.clear(); | |
36 word_id_history_map_.clear(); | |
37 if (!history_db) | |
38 return false; | |
39 URLDatabase::URLEnumerator history_enum; | |
40 if (history_db->InitURLEnumeratorForEverything(&history_enum)) { | |
41 URLRow row; | |
42 Time recent_threshold = InMemoryURLIndex::RecentThreshold(); | |
43 while (history_enum.GetNextURL(&row)) { | |
44 // Do some filtering so that we only get history items which could | |
45 // possibly pass the HistoryURLProvider::CullPoorMatches filter later. | |
46 if ((row.typed_count() > kLowQualityMatchTypedLimit) || | |
47 (row.visit_count() > kLowQualityMatchVisitLimit) || | |
48 (row.last_visit() >= recent_threshold)) { | |
49 if (!IndexRow(row)) | |
50 return false; | |
51 } | |
52 } | |
53 } | |
54 return true; | |
55 } | |
56 | |
57 bool InMemoryURLIndex::IndexRow(URLRow row) { | |
58 const GURL& gurl(row.url()); | |
59 string16 url(WideToUTF16(net::FormatUrl(gurl, UTF16ToWide(languages_), | |
60 net::kFormatUrlOmitUsernamePassword, | |
61 UnescapeRule::SPACES | UnescapeRule::URL_SPECIAL_CHARS, NULL, NULL, | |
62 NULL))); | |
63 | |
64 // TODO(mrossetti): Find or implement a ConvertPercentEncoding and use it | |
65 // on the url. | |
66 | |
67 // TODO(mrossetti): Detect row_id > std::numeric_limits<HistoryID>::max(). | |
68 HistoryID history_id = static_cast<HistoryID>(row.id()); | |
69 | |
70 // Add the row for quick lookup in the history info store. | |
71 url = l10n_util::ToLower(url); | |
72 URLRow new_row(GURL(url), row.id()); | |
73 new_row.set_visit_count(row.visit_count()); | |
74 new_row.set_typed_count(row.typed_count()); | |
75 new_row.set_last_visit(row.last_visit()); | |
76 new_row.set_title(row.title()); | |
77 history_info_map_.insert(std::make_pair(history_id, new_row)); | |
78 | |
79 // Split into individual, unique words. | |
80 String16Set words = WordsFromString16(url); | |
81 | |
82 // For each word, add a new entry into the word index referring to the | |
83 // associated history item. | |
84 for (String16Set::iterator iter = words.begin(); | |
85 iter != words.end(); ++iter) { | |
86 String16Set::value_type uni_word = *iter; | |
87 AddWordToIndex(uni_word, history_id); | |
88 } | |
89 ++history_item_count_; | |
90 return true; | |
91 } | |
92 | |
93 // Utility Functions | |
94 | |
95 InMemoryURLIndex::String16Set InMemoryURLIndex::WordsFromString16( | |
96 const string16& uni_string) { | |
97 String16Set words; | |
98 | |
99 // TODO(mrossetti): Replace all | and _'s with a space, all % quoted | |
100 // characters with real characters, and break into words, using | |
101 // appropriate string16 functions. | |
102 WordIterator iter(&uni_string, WordIterator::BREAK_WORD); | |
103 if (iter.Init()) { | |
104 while (iter.Advance()) { | |
105 if (iter.IsWord()) { | |
106 words.insert(iter.GetWord()); | |
107 } | |
108 } | |
109 } | |
110 return words; | |
111 } | |
112 | |
113 InMemoryURLIndex::Char16Set InMemoryURLIndex::CharactersFromString16( | |
114 const string16& uni_word) { | |
115 Char16Set characters; | |
116 for (string16::const_iterator iter = uni_word.begin(); | |
117 iter != uni_word.end(); ++iter) | |
118 characters.insert(*iter); | |
119 return characters; | |
120 } | |
121 | |
122 void InMemoryURLIndex::AddWordToIndex(const string16& uni_word, | |
123 HistoryID history_id) { | |
124 WordMap::iterator word_pos = word_map_.find(uni_word); | |
125 if (word_pos != word_map_.end()) | |
126 UpdateWordHistory(word_pos->second, history_id); | |
127 else | |
128 AddWordHistory(uni_word, history_id); | |
129 } | |
130 | |
131 void InMemoryURLIndex::UpdateWordHistory(WordID word_id, HistoryID history_id) { | |
132 WordIDHistoryMap::iterator history_pos = word_id_history_map_.find(word_id); | |
133 DCHECK(history_pos != word_id_history_map_.end()); | |
134 HistoryIDSet& history_id_set(history_pos->second); | |
135 history_id_set.insert(history_id); | |
136 } | |
137 | |
138 // Add a new word to the word list and the word map, and then create a | |
139 // new entry in the word/history map. | |
140 void InMemoryURLIndex::AddWordHistory(const string16& uni_word, | |
141 HistoryID history_id) { | |
142 word_list_.push_back(uni_word); | |
143 WordID word_id = word_list_.size() - 1; | |
144 DCHECK(word_map_.insert(std::make_pair(uni_word, word_id)).second); | |
145 HistoryIDSet history_id_set; | |
146 history_id_set.insert(history_id); | |
147 DCHECK(word_id_history_map_.insert( | |
148 std::make_pair(word_id, history_id_set)).second); | |
149 // For each character in the newly added word (i.e. a word that is not | |
150 // already in the word index), add the word to the character index. | |
151 Char16Set characters = CharactersFromString16(uni_word); | |
152 for (Char16Set::iterator uni_char_iter = characters.begin(); | |
153 uni_char_iter != characters.end(); ++uni_char_iter) { | |
154 Char16Set::value_type uni_string = *uni_char_iter; | |
155 CharWordIDMap::iterator char_iter = char_word_map_.find(uni_string); | |
156 if (char_iter != char_word_map_.end()) { | |
157 // Update existing entry in the char/word index. | |
158 WordIDSet& word_id_set(char_iter->second); | |
159 word_id_set.insert(word_id); | |
160 } else { | |
161 // Create a new entry in the char/word index. | |
162 WordIDSet word_id_set; | |
163 word_id_set.insert(word_id); | |
164 DCHECK(char_word_map_.insert(std::make_pair(uni_string, | |
165 word_id_set)).second); | |
166 } | |
167 } | |
168 } | |
169 | |
170 // static | |
171 Time InMemoryURLIndex::RecentThreshold() { | |
172 return Time::Now() - TimeDelta::FromDays(kLowQualityMatchAgeLimitInDays); | |
173 } | 17 } |
174 | 18 |
175 } // namespace history | 19 } // namespace history |
OLD | NEW |