bookmark_provider_unittest.cc revision 5f1c94371a64b3196d4be9466099bb892df9b88e
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#include "chrome/browser/autocomplete/bookmark_provider.h" 6 7#include <algorithm> 8#include <string> 9#include <vector> 10 11#include "base/memory/ref_counted.h" 12#include "base/memory/scoped_ptr.h" 13#include "base/strings/string16.h" 14#include "base/strings/string_number_conversions.h" 15#include "base/strings/string_split.h" 16#include "base/strings/utf_string_conversions.h" 17#include "chrome/browser/autocomplete/chrome_autocomplete_scheme_classifier.h" 18#include "chrome/test/base/testing_profile.h" 19#include "components/bookmarks/browser/bookmark_match.h" 20#include "components/bookmarks/browser/bookmark_model.h" 21#include "components/bookmarks/test/test_bookmark_client.h" 22#include "components/metrics/proto/omnibox_event.pb.h" 23#include "components/omnibox/autocomplete_provider.h" 24#include "testing/gtest/include/gtest/gtest.h" 25 26using bookmarks::BookmarkMatch; 27 28// The bookmark corpus against which we will simulate searches. 29struct BookmarksTestInfo { 30 std::string title; 31 std::string url; 32} bookmark_provider_test_data[] = { 33 { "abc def", "http://www.catsanddogs.com/a" }, 34 { "abcde", "http://www.catsanddogs.com/b" }, 35 { "abcdef", "http://www.catsanddogs.com/c" }, 36 { "a definition", "http://www.catsanddogs.com/d" }, 37 { "carry carbon carefully", "http://www.catsanddogs.com/e" }, 38 { "ghi jkl", "http://www.catsanddogs.com/f" }, 39 { "jkl ghi", "http://www.catsanddogs.com/g" }, 40 { "frankly frankly frank", "http://www.catsanddogs.com/h" }, 41 { "foobar foobar", "http://www.foobar.com/" }, 42 { "domain", "http://www.domain.com/http/" }, 43 { "repeat", "http://www.repeat.com/1/repeat/2/" }, 44 // For testing inline_autocompletion. 45 { "http://blah.com/", "http://blah.com/" }, 46 { "http://fiddle.com/", "http://fiddle.com/" }, 47 { "http://www.www.com/", "http://www.www.com/" }, 48 { "chrome://version", "chrome://version" }, 49 { "chrome://omnibox", "chrome://omnibox" }, 50 // For testing ranking with different URLs. 51 {"achlorhydric featherheads resuscitates mockingbirds", 52 "http://www.featherheads.com/a" }, 53 {"achlorhydric mockingbirds resuscitates featherhead", 54 "http://www.featherheads.com/b" }, 55 {"featherhead resuscitates achlorhydric mockingbirds", 56 "http://www.featherheads.com/c" }, 57 {"mockingbirds resuscitates featherheads achlorhydric", 58 "http://www.featherheads.com/d" }, 59 // For testing URL boosting. 60 {"burning worms #1", "http://www.burned.com/" }, 61 {"burning worms #2", "http://www.worms.com/" }, 62 {"worming burns #10", "http://www.burned.com/" }, 63 {"worming burns #20", "http://www.worms.com/" }, 64 {"jive music", "http://www.worms.com/" }, 65}; 66 67class BookmarkProviderTest : public testing::Test { 68 public: 69 BookmarkProviderTest(); 70 71 protected: 72 virtual void SetUp() OVERRIDE; 73 74 test::TestBookmarkClient client_; 75 scoped_ptr<TestingProfile> profile_; 76 scoped_ptr<BookmarkModel> model_; 77 scoped_refptr<BookmarkProvider> provider_; 78 79 private: 80 DISALLOW_COPY_AND_ASSIGN(BookmarkProviderTest); 81}; 82 83BookmarkProviderTest::BookmarkProviderTest() { 84 model_ = client_.CreateModel(false); 85} 86 87void BookmarkProviderTest::SetUp() { 88 profile_.reset(new TestingProfile()); 89 DCHECK(profile_.get()); 90 provider_ = new BookmarkProvider(profile_.get()); 91 DCHECK(provider_.get()); 92 provider_->set_bookmark_model_for_testing(model_.get()); 93 94 const BookmarkNode* other_node = model_->other_node(); 95 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(bookmark_provider_test_data); ++i) { 96 const BookmarksTestInfo& cur(bookmark_provider_test_data[i]); 97 const GURL url(cur.url); 98 model_->AddURL(other_node, other_node->child_count(), 99 base::ASCIIToUTF16(cur.title), url); 100 } 101} 102 103// Structures and functions supporting the BookmarkProviderTest.Positions 104// unit test. 105 106struct TestBookmarkPosition { 107 TestBookmarkPosition(size_t begin, size_t end) 108 : begin(begin), end(end) {} 109 110 size_t begin; 111 size_t end; 112}; 113typedef std::vector<TestBookmarkPosition> TestBookmarkPositions; 114 115// Return |positions| as a formatted string for unit test diagnostic output. 116std::string TestBookmarkPositionsAsString( 117 const TestBookmarkPositions& positions) { 118 std::string position_string("{"); 119 for (TestBookmarkPositions::const_iterator i = positions.begin(); 120 i != positions.end(); ++i) { 121 if (i != positions.begin()) 122 position_string += ", "; 123 position_string += "{" + base::IntToString(i->begin) + ", " + 124 base::IntToString(i->end) + "}"; 125 } 126 position_string += "}\n"; 127 return position_string; 128} 129 130// Return the positions in |matches| as a formatted string for unit test 131// diagnostic output. 132base::string16 MatchesAsString16(const ACMatches& matches) { 133 base::string16 matches_string; 134 for (ACMatches::const_iterator i = matches.begin(); i != matches.end(); ++i) { 135 matches_string.append(base::ASCIIToUTF16(" '")); 136 matches_string.append(i->description); 137 matches_string.append(base::ASCIIToUTF16("'\n")); 138 } 139 return matches_string; 140} 141 142// Comparison function for sorting search terms by descending length. 143bool TestBookmarkPositionsEqual(const TestBookmarkPosition& pos_a, 144 const TestBookmarkPosition& pos_b) { 145 return pos_a.begin == pos_b.begin && pos_a.end == pos_b.end; 146} 147 148// Convience function to make comparing ACMatchClassifications against the 149// test expectations structure easier. 150TestBookmarkPositions PositionsFromAutocompleteMatch( 151 const AutocompleteMatch& match) { 152 TestBookmarkPositions positions; 153 bool started = false; 154 size_t start = 0; 155 for (AutocompleteMatch::ACMatchClassifications::const_iterator 156 i = match.description_class.begin(); 157 i != match.description_class.end(); ++i) { 158 if (i->style & AutocompleteMatch::ACMatchClassification::MATCH) { 159 // We have found the start of a match. 160 EXPECT_FALSE(started); 161 started = true; 162 start = i->offset; 163 } else if (started) { 164 // We have found the end of a match. 165 started = false; 166 positions.push_back(TestBookmarkPosition(start, i->offset)); 167 start = 0; 168 } 169 } 170 // Record the final position if the last match goes to the end of the 171 // candidate string. 172 if (started) 173 positions.push_back(TestBookmarkPosition(start, match.description.size())); 174 return positions; 175} 176 177// Convience function to make comparing test expectations structure against the 178// actual ACMatchClassifications easier. 179TestBookmarkPositions PositionsFromExpectations( 180 const size_t expectations[9][2]) { 181 TestBookmarkPositions positions; 182 size_t i = 0; 183 // The array is zero-terminated in the [1]th element. 184 while (expectations[i][1]) { 185 positions.push_back( 186 TestBookmarkPosition(expectations[i][0], expectations[i][1])); 187 ++i; 188 } 189 return positions; 190} 191 192TEST_F(BookmarkProviderTest, Positions) { 193 // Simulate searches. 194 // Description of |positions|: 195 // The first index represents the collection of positions for each expected 196 // match. The count of the actual subarrays in each instance of |query_data| 197 // must equal |match_count|. The second index represents each expected 198 // match position. The third index represents the |start| and |end| of the 199 // expected match's position within the |test_data|. This array must be 200 // terminated by an entry with a value of '0' for |end|. 201 // Example: 202 // Consider the line for 'def' below: 203 // {"def", 2, {{{4, 7}, {XXX, 0}}, {{2, 5}, {11, 14}, {XXX, 0}}}}, 204 // There are two expected matches: 205 // 0. {{4, 7}, {XXX, 0}} 206 // 1. {{2, 5}, {11 ,14}, {XXX, 0}} 207 // For the first match, [0], there is one match within the bookmark's title 208 // expected, {4, 7}, which maps to the 'def' within "abc def". The 'XXX' 209 // value is ignored. The second match, [1], indicates that two matches are 210 // expected within the bookmark title "a definite definition". In each case, 211 // the {XXX, 0} indicates the end of the subarray. Or: 212 // Match #1 Match #2 213 // ------------------ ---------------------------- 214 // Pos1 Term Pos1 Pos2 Term 215 // ------ -------- ------ -------- -------- 216 // {"def", 2, {{{4, 7}, {999, 0}}, {{2, 5}, {11, 14}, {999, 0}}}}, 217 // 218 struct QueryData { 219 const std::string query; 220 const size_t match_count; // This count must match the number of major 221 // elements in the following |positions| array. 222 const size_t positions[99][9][2]; 223 } query_data[] = { 224 // This first set is primarily for position detection validation. 225 {"abc", 3, {{{0, 3}, {0, 0}}, 226 {{0, 3}, {0, 0}}, 227 {{0, 3}, {0, 0}}}}, 228 {"abcde", 2, {{{0, 5}, {0, 0}}, 229 {{0, 5}, {0, 0}}}}, 230 {"foo bar", 0, {{{0, 0}}}}, 231 {"fooey bark", 0, {{{0, 0}}}}, 232 {"def", 2, {{{2, 5}, {0, 0}}, 233 {{4, 7}, {0, 0}}}}, 234 {"ghi jkl", 2, {{{0, 3}, {4, 7}, {0, 0}}, 235 {{0, 3}, {4, 7}, {0, 0}}}}, 236 // NB: GetBookmarksWithTitlesMatching(...) uses exact match for "a". 237 {"a", 1, {{{0, 1}, {0, 0}}}}, 238 {"a d", 0, {{{0, 0}}}}, 239 {"carry carbon", 1, {{{0, 5}, {6, 12}, {0, 0}}}}, 240 // NB: GetBookmarksWithTitlesMatching(...) sorts the match positions. 241 {"carbon carry", 1, {{{0, 5}, {6, 12}, {0, 0}}}}, 242 {"arbon", 0, {{{0, 0}}}}, 243 {"ar", 0, {{{0, 0}}}}, 244 {"arry", 0, {{{0, 0}}}}, 245 // Quoted terms are single terms. 246 {"\"carry carbon\"", 1, {{{0, 12}, {0, 0}}}}, 247 {"\"carry carbon\" care", 1, {{{0, 12}, {13, 17}, {0, 0}}}}, 248 // Quoted terms require complete word matches. 249 {"\"carry carbo\"", 0, {{{0, 0}}}}, 250 // This set uses duplicated and/or overlaps search terms in the title. 251 {"frank", 1, {{{0, 5}, {8, 13}, {16, 21}, {0, 0}}}}, 252 {"frankly", 1, {{{0, 7}, {8, 15}, {0, 0}}}}, 253 {"frankly frankly", 1, {{{0, 7}, {8, 15}, {0, 0}}}}, 254 {"foobar foo", 1, {{{0, 6}, {7, 13}, {0, 0}}}}, 255 {"foo foobar", 1, {{{0, 6}, {7, 13}, {0, 0}}}}, 256 }; 257 258 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(query_data); ++i) { 259 AutocompleteInput input(base::ASCIIToUTF16(query_data[i].query), 260 base::string16::npos, base::string16(), GURL(), 261 metrics::OmniboxEventProto::INVALID_SPEC, false, 262 false, false, true, 263 ChromeAutocompleteSchemeClassifier(profile_.get())); 264 provider_->Start(input, false); 265 const ACMatches& matches(provider_->matches()); 266 // Validate number of results is as expected. 267 EXPECT_LE(matches.size(), query_data[i].match_count) 268 << "One or more of the following matches were unexpected:\n" 269 << MatchesAsString16(matches) 270 << "For query '" << query_data[i].query << "'."; 271 EXPECT_GE(matches.size(), query_data[i].match_count) 272 << "One or more expected matches are missing. Matches found:\n" 273 << MatchesAsString16(matches) 274 << "for query '" << query_data[i].query << "'."; 275 // Validate positions within each match is as expected. 276 for (size_t j = 0; j < matches.size(); ++j) { 277 // Collect the expected positions as a vector, collect the match's 278 // classifications for match positions as a vector, then compare. 279 TestBookmarkPositions expected_positions( 280 PositionsFromExpectations(query_data[i].positions[j])); 281 TestBookmarkPositions actual_positions( 282 PositionsFromAutocompleteMatch(matches[j])); 283 EXPECT_TRUE(std::equal(expected_positions.begin(), 284 expected_positions.end(), 285 actual_positions.begin(), 286 TestBookmarkPositionsEqual)) 287 << "EXPECTED: " << TestBookmarkPositionsAsString(expected_positions) 288 << "ACTUAL: " << TestBookmarkPositionsAsString(actual_positions) 289 << " for query: '" << query_data[i].query << "'."; 290 } 291 } 292} 293 294TEST_F(BookmarkProviderTest, Rankings) { 295 // Simulate searches. 296 struct QueryData { 297 const std::string query; 298 // |match_count| must match the number of elements in the following 299 // |matches| array. 300 const size_t match_count; 301 // |matches| specifies the titles for all bookmarks expected to be matched 302 // by the |query| 303 const std::string matches[3]; 304 } query_data[] = { 305 // Basic ranking test. 306 {"abc", 3, {"abcde", // Most complete match. 307 "abcdef", 308 "abc def"}}, // Least complete match. 309 {"ghi", 2, {"ghi jkl", // Matched earlier. 310 "jkl ghi", // Matched later. 311 ""}}, 312 // Rankings of exact-word matches with different URLs. 313 {"achlorhydric", 314 3, {"achlorhydric mockingbirds resuscitates featherhead", 315 "achlorhydric featherheads resuscitates mockingbirds", 316 "featherhead resuscitates achlorhydric mockingbirds"}}, 317 {"achlorhydric featherheads", 318 2, {"achlorhydric featherheads resuscitates mockingbirds", 319 "mockingbirds resuscitates featherheads achlorhydric", 320 ""}}, 321 {"mockingbirds resuscitates", 322 3, {"mockingbirds resuscitates featherheads achlorhydric", 323 "achlorhydric mockingbirds resuscitates featherhead", 324 "featherhead resuscitates achlorhydric mockingbirds"}}, 325 // Ranking of exact-word matches with URL boost. 326 {"worms", 2, {"burning worms #2", // boosted 327 "burning worms #1", // not boosted 328 ""}}, 329 // Ranking of prefix matches with URL boost. Note that a query of 330 // "worm burn" will have the same results. 331 {"burn worm", 3, {"burning worms #2", // boosted 332 "worming burns #20", // boosted 333 "burning worms #1"}}, // not boosted but shorter 334 }; 335 336 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(query_data); ++i) { 337 AutocompleteInput input(base::ASCIIToUTF16(query_data[i].query), 338 base::string16::npos, base::string16(), GURL(), 339 metrics::OmniboxEventProto::INVALID_SPEC, false, 340 false, false, true, 341 ChromeAutocompleteSchemeClassifier(profile_.get())); 342 provider_->Start(input, false); 343 const ACMatches& matches(provider_->matches()); 344 // Validate number and content of results is as expected. 345 for (size_t j = 0; j < std::max(query_data[i].match_count, matches.size()); 346 ++j) { 347 EXPECT_LT(j, query_data[i].match_count) << " Unexpected match '" 348 << base::UTF16ToUTF8(matches[j].description) << "' for query: '" 349 << query_data[i].query << "'."; 350 if (j >= query_data[i].match_count) 351 continue; 352 EXPECT_LT(j, matches.size()) << " Missing match '" 353 << query_data[i].matches[j] << "' for query: '" 354 << query_data[i].query << "'."; 355 if (j >= matches.size()) 356 continue; 357 EXPECT_EQ(query_data[i].matches[j], 358 base::UTF16ToUTF8(matches[j].description)) 359 << " Mismatch at [" << base::IntToString(j) << "] for query '" 360 << query_data[i].query << "'."; 361 } 362 } 363} 364 365TEST_F(BookmarkProviderTest, InlineAutocompletion) { 366 // Simulate searches. 367 struct QueryData { 368 const std::string query; 369 const std::string url; 370 const bool allowed_to_be_default_match; 371 const std::string inline_autocompletion; 372 } query_data[] = { 373 { "bla", "http://blah.com/", true, "h.com" }, 374 { "blah ", "http://blah.com/", false, ".com" }, 375 { "http://bl", "http://blah.com/", true, "ah.com" }, 376 { "fiddle.c", "http://fiddle.com/", true, "om" }, 377 { "www", "http://www.www.com/", true, ".com" }, 378 { "chro", "chrome://version", true, "me://version" }, 379 { "chrome://ve", "chrome://version", true, "rsion" }, 380 { "chrome ver", "chrome://version", false, "" }, 381 { "versi", "chrome://version", false, "" }, 382 { "abou", "chrome://omnibox", false, "" }, 383 { "about:om", "chrome://omnibox", true, "nibox" } 384 // Note: when adding a new URL to this test, be sure to add it to the list 385 // of bookmarks at the top of the file as well. All items in this list 386 // need to be in the bookmarks list because BookmarkProvider's 387 // TitleMatchToACMatch() has an assertion that verifies the URL is 388 // actually bookmarked. 389 }; 390 391 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(query_data); ++i) { 392 const std::string description = "for query=" + query_data[i].query + 393 " and url=" + query_data[i].url; 394 AutocompleteInput input(base::ASCIIToUTF16(query_data[i].query), 395 base::string16::npos, base::string16(), GURL(), 396 metrics::OmniboxEventProto::INVALID_SPEC, false, 397 false, false, true, 398 ChromeAutocompleteSchemeClassifier(profile_.get())); 399 const base::string16 fixed_up_input( 400 provider_->FixupUserInput(input).second); 401 BookmarkNode node(GURL(query_data[i].url)); 402 node.SetTitle(base::ASCIIToUTF16(query_data[i].url)); 403 BookmarkMatch bookmark_match; 404 bookmark_match.node = &node; 405 const AutocompleteMatch& ac_match = provider_->BookmarkMatchToACMatch( 406 input, fixed_up_input, bookmark_match); 407 EXPECT_EQ(query_data[i].allowed_to_be_default_match, 408 ac_match.allowed_to_be_default_match) << description; 409 EXPECT_EQ(base::ASCIIToUTF16(query_data[i].inline_autocompletion), 410 ac_match.inline_autocompletion) << description; 411 } 412} 413 414TEST_F(BookmarkProviderTest, StripHttpAndAdjustOffsets) { 415 // Simulate searches. 416 struct QueryData { 417 const std::string query; 418 const std::string expected_contents; 419 // |expected_contents_class| is in format offset:style,offset:style,... 420 const std::string expected_contents_class; 421 } query_data[] = { 422 { "foo", "www.foobar.com", "0:1,4:3,7:1" }, 423 { "www foo", "www.foobar.com", "0:3,3:1,4:3,7:1" }, 424 { "foo www", "www.foobar.com", "0:3,3:1,4:3,7:1" }, 425 { "foo http", "http://www.foobar.com", "0:3,4:1,11:3,14:1" }, 426 { "blah", "blah.com", "0:3,4:1" }, 427 { "http blah", "http://blah.com", "0:3,4:1,7:3,11:1" }, 428 { "dom", "www.domain.com/http/", "0:1,4:3,7:1" }, 429 { "dom http", "http://www.domain.com/http/", 430 "0:3,4:1,11:3,14:1,22:3,26:1" }, 431 { "rep", "www.repeat.com/1/repeat/2/", "0:1,4:3,7:1,17:3,20:1" }, 432 { "versi", "chrome://version", "0:1,9:3,14:1" } 433 }; 434 435 // Reload the bookmarks index with |index_urls| == true. 436 model_ = client_.CreateModel(true); 437 SetUp(); 438 439 for (size_t i = 0; i < ARRAYSIZE_UNSAFE(query_data); ++i) { 440 std::string description = "for query=" + query_data[i].query; 441 AutocompleteInput input(base::ASCIIToUTF16(query_data[i].query), 442 base::string16::npos, base::string16(), GURL(), 443 metrics::OmniboxEventProto::INVALID_SPEC, false, 444 false, false, true, 445 ChromeAutocompleteSchemeClassifier(profile_.get())); 446 provider_->Start(input, false); 447 const ACMatches& matches(provider_->matches()); 448 ASSERT_EQ(1U, matches.size()) << description; 449 const AutocompleteMatch& match = matches[0]; 450 EXPECT_EQ(base::ASCIIToUTF16(query_data[i].expected_contents), 451 match.contents) << description; 452 std::vector<std::string> class_strings; 453 base::SplitString( 454 query_data[i].expected_contents_class, ',', &class_strings); 455 ASSERT_EQ(class_strings.size(), match.contents_class.size()) 456 << description; 457 for (size_t i = 0; i < class_strings.size(); ++i) { 458 std::vector<std::string> chunks; 459 base::SplitString(class_strings[i], ':', &chunks); 460 ASSERT_EQ(2U, chunks.size()) << description; 461 size_t offset; 462 EXPECT_TRUE(base::StringToSizeT(chunks[0], &offset)) << description; 463 EXPECT_EQ(offset, match.contents_class[i].offset) << description; 464 int style; 465 EXPECT_TRUE(base::StringToInt(chunks[1], &style)) << description; 466 EXPECT_EQ(style, match.contents_class[i].style) << description; 467 } 468 } 469} 470