summaryrefslogtreecommitdiffstats
path: root/chrome/browser/history/query_parser.cc
blob: a5447f60818c351ad7d3c23e1986ff864062e629 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
// Copyright (c) 2010 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 "chrome/browser/history/query_parser.h"

#include <algorithm>

#include "base/i18n/break_iterator.h"
#include "base/logging.h"
#include "base/scoped_vector.h"
#include "base/string_util.h"
#include "base/utf_string_conversions.h"
#include "ui/base/l10n/l10n_util.h"
#include "unicode/uscript.h"

namespace {

// Returns true if |mp1.first| is less than |mp2.first|. This is used to
// sort match positions.
int CompareMatchPosition(const Snippet::MatchPosition& mp1,
                         const Snippet::MatchPosition& mp2) {
  return mp1.first < mp2.first;
}

// Returns true if |mp2| intersects |mp1|. This is intended for use by
// CoalesceMatchesFrom and isn't meant as a general intersectpion comparison
// function.
bool SnippetIntersects(const Snippet::MatchPosition& mp1,
                       const Snippet::MatchPosition& mp2) {
  return mp2.first >= mp1.first && mp2.first <= mp1.second;
}

// Coalesces match positions in |matches| after index that intersect the match
// position at |index|.
void CoalesceMatchesFrom(size_t index,
                         Snippet::MatchPositions* matches) {
  Snippet::MatchPosition& mp = (*matches)[index];
  for (Snippet::MatchPositions::iterator i = matches->begin() + index + 1;
       i != matches->end(); ) {
    if (SnippetIntersects(mp, *i)) {
      mp.second = i->second;
      i = matches->erase(i);
    } else {
      return;
    }
  }
}

// Sorts the match positions in |matches| by their first index, then coalesces
// any match positions that intersect each other.
void CoalseAndSortMatchPositions(Snippet::MatchPositions* matches) {
  std::sort(matches->begin(), matches->end(), &CompareMatchPosition);
  // WARNING: we don't use iterator here as CoalesceMatchesFrom may remove
  // from matches.
  for (size_t i = 0; i < matches->size(); ++i)
    CoalesceMatchesFrom(i, matches);
}

}  // namespace

// Inheritance structure:
// Queries are represented as trees of QueryNodes.
// QueryNodes are either a collection of subnodes (a QueryNodeList)
// or a single word (a QueryNodeWord).

// A QueryNodeWord is a single word in the query.
class QueryNodeWord : public QueryNode {
 public:
  explicit QueryNodeWord(const string16& word)
      : word_(word), literal_(false) {}
  virtual ~QueryNodeWord() {}
  virtual int AppendToSQLiteQuery(string16* query) const;
  virtual bool IsWord() const { return true; }

  const string16& word() const { return word_; }
  void set_literal(bool literal) { literal_ = literal; }

  virtual bool HasMatchIn(const std::vector<QueryWord>& words,
                          Snippet::MatchPositions* match_positions) const;

  virtual bool Matches(const string16& word, bool exact) const;
  virtual void AppendWords(std::vector<string16>* words) const;

 private:
  string16 word_;
  bool literal_;
};

bool QueryNodeWord::HasMatchIn(const std::vector<QueryWord>& words,
                               Snippet::MatchPositions* match_positions) const {
  for (size_t i = 0; i < words.size(); ++i) {
    if (Matches(words[i].word, false)) {
      size_t match_start = words[i].position;
      match_positions->push_back(
          Snippet::MatchPosition(match_start,
                                 match_start + static_cast<int>(word_.size())));
      return true;
    }
  }
  return false;
}

bool QueryNodeWord::Matches(const string16& word, bool exact) const {
  if (exact || !QueryParser::IsWordLongEnoughForPrefixSearch(word_))
    return word == word_;
  return word.size() >= word_.size() &&
         (word_.compare(0, word_.size(), word, 0, word_.size()) == 0);
}

void QueryNodeWord::AppendWords(std::vector<string16>* words) const {
  words->push_back(word_);
}

int QueryNodeWord::AppendToSQLiteQuery(string16* query) const {
  query->append(word_);

  // Use prefix search if we're not literal and long enough.
  if (!literal_ && QueryParser::IsWordLongEnoughForPrefixSearch(word_))
    *query += L'*';
  return 1;
}

// A QueryNodeList has a collection of child QueryNodes
// which it cleans up after.
class QueryNodeList : public QueryNode {
 public:
  virtual ~QueryNodeList();

  virtual int AppendToSQLiteQuery(string16* query) const {
    return AppendChildrenToString(query);
  }
  virtual bool IsWord() const { return false; }

  void AddChild(QueryNode* node) { children_.push_back(node); }

  typedef std::vector<QueryNode*> QueryNodeVector;
  QueryNodeVector* children() { return &children_; }

  // Remove empty subnodes left over from other parsing.
  void RemoveEmptySubnodes();

  // QueryNodeList is never used with Matches or HasMatchIn.
  virtual bool Matches(const string16& word, bool exact) const {
    NOTREACHED();
    return false;
  }
  virtual bool HasMatchIn(const std::vector<QueryWord>& words,
                          Snippet::MatchPositions* match_positions) const {
    NOTREACHED();
    return false;
  }
  virtual void AppendWords(std::vector<string16>* words) const;

 protected:
  int AppendChildrenToString(string16* query) const;

  QueryNodeVector children_;
};

QueryNodeList::~QueryNodeList() {
  for (QueryNodeVector::iterator node = children_.begin();
       node != children_.end(); ++node)
    delete *node;
}

void QueryNodeList::RemoveEmptySubnodes() {
  for (size_t i = 0; i < children_.size(); ++i) {
    if (children_[i]->IsWord())
      continue;

    QueryNodeList* list_node = static_cast<QueryNodeList*>(children_[i]);
    list_node->RemoveEmptySubnodes();
    if (list_node->children()->empty()) {
      children_.erase(children_.begin() + i);
      --i;
      delete list_node;
    }
  }
}

void QueryNodeList::AppendWords(std::vector<string16>* words) const {
  for (size_t i = 0; i < children_.size(); ++i)
    children_[i]->AppendWords(words);
}

int QueryNodeList::AppendChildrenToString(string16* query) const {
  int num_words = 0;
  for (QueryNodeVector::const_iterator node = children_.begin();
       node != children_.end(); ++node) {
    if (node != children_.begin())
      query->push_back(L' ');
    num_words += (*node)->AppendToSQLiteQuery(query);
  }
  return num_words;
}

// A QueryNodePhrase is a phrase query ("quoted").
class QueryNodePhrase : public QueryNodeList {
 public:
  virtual int AppendToSQLiteQuery(string16* query) const {
    query->push_back(L'"');
    int num_words = AppendChildrenToString(query);
    query->push_back(L'"');
    return num_words;
  }

  virtual bool Matches(const string16& word, bool exact) const;
  virtual bool HasMatchIn(const std::vector<QueryWord>& words,
                          Snippet::MatchPositions* match_positions) const;
};

bool QueryNodePhrase::Matches(const string16& word, bool exact) const {
  NOTREACHED();
  return false;
}

bool QueryNodePhrase::HasMatchIn(
    const std::vector<QueryWord>& words,
    Snippet::MatchPositions* match_positions) const {
  if (words.size() < children_.size())
    return false;

  for (size_t i = 0, max = words.size() - children_.size() + 1; i < max; ++i) {
    bool matched_all = true;
    for (size_t j = 0; j < children_.size(); ++j) {
      if (!children_[j]->Matches(words[i + j].word, true)) {
        matched_all = false;
        break;
      }
    }
    if (matched_all) {
      const QueryWord& last_word = words[i + children_.size() - 1];
      match_positions->push_back(
          Snippet::MatchPosition(words[i].position,
                                 last_word.position + last_word.word.length()));
      return true;
    }
  }
  return false;
}

QueryParser::QueryParser() {
}

// static
bool QueryParser::IsWordLongEnoughForPrefixSearch(const string16& word) {
  DCHECK(!word.empty());
  size_t minimum_length = 3;
  // We intentionally exclude Hangul Jamos (both Conjoining and compatibility)
  // because they 'behave like' Latin letters. Moreover, we should
  // normalize the former before reaching here.
  if (0xAC00 <= word[0] && word[0] <= 0xD7A3)
    minimum_length = 2;
  return word.size() >= minimum_length;
}

// Returns true if the character is considered a quote.
static bool IsQueryQuote(wchar_t ch) {
  return ch == '"' ||
         ch == 0xab ||    // left pointing double angle bracket
         ch == 0xbb ||    // right pointing double angle bracket
         ch == 0x201c ||  // left double quotation mark
         ch == 0x201d ||  // right double quotation mark
         ch == 0x201e;    // double low-9 quotation mark
}

int QueryParser::ParseQuery(const string16& query,
                            string16* sqlite_query) {
  QueryNodeList root;
  if (!ParseQueryImpl(query, &root))
    return 0;
  return root.AppendToSQLiteQuery(sqlite_query);
}

void QueryParser::ParseQuery(const string16& query,
                             std::vector<QueryNode*>* nodes) {
  QueryNodeList root;
  if (ParseQueryImpl(l10n_util::ToLower(query), &root))
    nodes->swap(*root.children());
}


void QueryParser::ExtractQueryWords(const string16& query,
                                    std::vector<string16>* words) {
  QueryNodeList root;
  if (!ParseQueryImpl(query, &root))
    return;
  root.AppendWords(words);
}

bool QueryParser::DoesQueryMatch(const string16& text,
                                 const std::vector<QueryNode*>& query_nodes,
                                 Snippet::MatchPositions* match_positions) {
  if (query_nodes.empty())
    return false;

  std::vector<QueryWord> query_words;
  string16 lower_text = l10n_util::ToLower(text);
  ExtractQueryWords(lower_text, &query_words);

  if (query_words.empty())
    return false;

  Snippet::MatchPositions matches;
  for (size_t i = 0; i < query_nodes.size(); ++i) {
    if (!query_nodes[i]->HasMatchIn(query_words, &matches))
      return false;
  }
  if (lower_text.length() != text.length()) {
    // The lower case string differs from the original string. The matches are
    // meaningless.
    // TODO(sky): we need a better way to align the positions so that we don't
    // completely punt here.
    match_positions->clear();
  } else {
    CoalseAndSortMatchPositions(&matches);
    match_positions->swap(matches);
  }
  return true;
}

bool QueryParser::ParseQueryImpl(const string16& query,
                                QueryNodeList* root) {
  base::BreakIterator iter(&query, base::BreakIterator::BREAK_WORD);
  // TODO(evanm): support a locale here
  if (!iter.Init())
    return false;

  // To handle nesting, we maintain a stack of QueryNodeLists.
  // The last element (back) of the stack contains the current, deepest node.
  std::vector<QueryNodeList*> query_stack;
  query_stack.push_back(root);

  bool in_quotes = false;  // whether we're currently in a quoted phrase
  while (iter.Advance()) {
    // Just found a span between 'prev' (inclusive) and 'pos' (exclusive). It
    // is not necessarily a word, but could also be a sequence of punctuation
    // or whitespace.
    if (iter.IsWord()) {
      string16 word = iter.GetString();

      QueryNodeWord* word_node = new QueryNodeWord(word);
      if (in_quotes)
        word_node->set_literal(true);
      query_stack.back()->AddChild(word_node);
    } else {  // Punctuation.
      if (IsQueryQuote(query[iter.prev()])) {
        if (!in_quotes) {
          QueryNodeList* quotes_node = new QueryNodePhrase;
          query_stack.back()->AddChild(quotes_node);
          query_stack.push_back(quotes_node);
          in_quotes = true;
        } else {
          query_stack.pop_back();  // Stop adding to the quoted phrase.
          in_quotes = false;
        }
      }
    }
  }

  root->RemoveEmptySubnodes();
  return true;
}

void QueryParser::ExtractQueryWords(const string16& text,
                                    std::vector<QueryWord>* words) {
  base::BreakIterator iter(&text, base::BreakIterator::BREAK_WORD);
  // TODO(evanm): support a locale here
  if (!iter.Init())
    return;

  while (iter.Advance()) {
    // Just found a span between 'prev' (inclusive) and 'pos' (exclusive). It
    // is not necessarily a word, but could also be a sequence of punctuation
    // or whitespace.
    if (iter.IsWord()) {
      string16 word = iter.GetString();
      if (!word.empty()) {
        words->push_back(QueryWord());
        words->back().word = word;
        words->back().position = iter.prev();
      }
    }
  }
}