Open Source alternative to Algolia + Pinecone and an Easier-to-Use alternative to ElasticSearch ⚡ 🔍 ✨ Fast, typo tolerant, in-memory fuzzy Search Engine for building delightful search experiences
C++ algoliadatastoreelasticsearchenterprise-searchfacetingfull-text-searchfuzzy-searchgeosearchin-memoryinstantsearchmerchandisingpineconesearchsearch-enginesemantic-searchsimilarity-searchsite-searchsynonymstypo-tolerancevector-search
Fix exact matching regression.
Kishore Nallan13d4c14889ed
diff --git a/src/index.cpp b/src/index.cppindex 9e05a46e..9d2af0f4 100644--- a/src/index.cpp+++ b/src/index.cpp@@ -638,8 +638,8 @@ void Index::index_field_in_memory(const field& afield, std::vector<index_record>art_tree *t = tree_it->second;for(auto& token_to_doc: token_to_doc_offsets) {- std::vector<art_document>& documents = token_to_doc.second;const std::string& token = token_to_doc.first;+ std::vector<art_document>& documents = token_to_doc.second;const auto *key = (const unsigned char *) token.c_str();int key_len = (int) token.length() + 1; // for the terminating \0 char@@ -1483,9 +1483,8 @@ void Index::do_filtering(uint32_t*& filter_ids, uint32_t& filter_ids_length,strt_ids_size = result_id_vec.size();}- if((a_filter.comparators[0] == EQUALS || a_filter.comparators[0] == NOT_EQUALS) && f.is_facet()) {- // need to do exact match (unlike CONTAINS) by using the facet index- // field being a facet is already enforced upstream+ if(a_filter.comparators[0] == EQUALS || a_filter.comparators[0] == NOT_EQUALS) {+ // need to do exact match (unlike CONTAINS)uint32_t* exact_strt_ids = new uint32_t[strt_ids_size];size_t exact_strt_size = 0;@@ -1497,7 +1496,7 @@ void Index::do_filtering(uint32_t*& filter_ids, uint32_t& filter_ids_length,strt_ids_size = exact_strt_size;}- if(a_filter.comparators[0] == NOT_EQUALS && f.is_facet()) {+ if(a_filter.comparators[0] == NOT_EQUALS) {// exclude records from existing IDs (from previous filters or ALL records)// upstream will guarantee that NOT_EQUALS is placed right at the end of filters listif(ids == nullptr) {diff --git a/src/posting_list.cpp b/src/posting_list.cppindex 83787abd..526848ec 100644--- a/src/posting_list.cpp+++ b/src/posting_list.cpp@@ -987,6 +987,7 @@ void posting_list_t::get_exact_matches(std::vector<iterator_t>& its, const boolif(its.size() == 1) {for(size_t i = 0; i < num_ids; i++) {uint32_t id = ids[i];+ its[0].skip_to(id);if(is_single_token_verbatim_match(its[0], field_is_array)) {exact_ids[exact_id_index++] = id;}diff --git a/test/collection_filtering_test.cpp b/test/collection_filtering_test.cppindex 02866b76..216fb3ec 100644--- a/test/collection_filtering_test.cpp+++ b/test/collection_filtering_test.cpp@@ -1963,8 +1963,54 @@ TEST_F(CollectionFilteringTest, FilteringWithTokenSeparators) {collectionManager.drop_collection("coll2");}+TEST_F(CollectionFilteringTest, ExactFilteringSingleQueryTerm) {+ std::vector<field> fields = {field("name", field_types::STRING, false),+ field("tags", field_types::STRING_ARRAY, false)};++ Collection* coll1 = collectionManager.create_collection(+ "coll1", 1, fields, "", 0, "", {}, {"."}+ ).get();++ nlohmann::json doc1;+ doc1["id"] = "0";+ doc1["name"] = "AT&T GoPhone";+ doc1["tags"] = {"AT&T GoPhone"};++ nlohmann::json doc2;+ doc2["id"] = "1";+ doc2["name"] = "AT&T";+ doc2["tags"] = {"AT&T"};++ ASSERT_TRUE(coll1->add(doc1.dump()).ok());+ ASSERT_TRUE(coll1->add(doc2.dump()).ok());++ auto results = coll1->search("*", {},"name:=AT&T", {}, {}, {0}, 10,+ 1, FREQUENCY, {false}).get();+ ASSERT_EQ(1, results["hits"].size());+ ASSERT_EQ("1", results["hits"][0]["document"]["id"].get<std::string>());++ results = coll1->search("*", {},"tags:=AT&T", {}, {}, {0}, 10,+ 1, FREQUENCY, {false}).get();+ ASSERT_EQ(1, results["hits"].size());+ ASSERT_EQ("1", results["hits"][0]["document"]["id"].get<std::string>());++ nlohmann::json doc3;+ doc3["id"] = "2";+ doc3["name"] = "Phone";+ doc3["tags"] = {"Samsung Phone", "Phone"};++ ASSERT_TRUE(coll1->add(doc3.dump()).ok());++ results = coll1->search("*", {},"tags:=Phone", {}, {}, {0}, 10,+ 1, FREQUENCY, {false}).get();+ ASSERT_EQ(1, results["hits"].size());+ ASSERT_EQ("2", results["hits"][0]["document"]["id"].get<std::string>());++ collectionManager.drop_collection("coll1");+}+TEST_F(CollectionFilteringTest, ExactFilteringRepeatingTokensSingularField) {- std::vector<field> fields = {field("name", field_types::STRING, true)};+ std::vector<field> fields = {field("name", field_types::STRING, false)};Collection* coll1 = collectionManager.create_collection("coll1", 1, fields, "", 0, "", {}, {"."}@@ -2013,7 +2059,7 @@ TEST_F(CollectionFilteringTest, ExactFilteringRepeatingTokensSingularField) {}TEST_F(CollectionFilteringTest, ExactFilteringRepeatingTokensArrayField) {- std::vector<field> fields = {field("name", field_types::STRING_ARRAY, true)};+ std::vector<field> fields = {field("name", field_types::STRING_ARRAY, false)};Collection* coll1 = collectionManager.create_collection("coll1", 1, fields, "", 0, "", {}, {"."}