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
Add `CollectionJoinTest, FilterByReferenceAlias`.
Harpreet Sangare13adbb0c28e
diff --git a/test/collection_join_test.cpp b/test/collection_join_test.cppindex e3e15354..577bd9d7 100644--- a/test/collection_join_test.cpp+++ b/test/collection_join_test.cpp@@ -4056,3 +4056,181 @@ TEST_F(CollectionJoinTest, SortByReference) {ASSERT_FALSE(search_op.ok());ASSERT_EQ("Multiple references found to sort by on `Customers.product_price`.", search_op.error());}++TEST_F(CollectionJoinTest, FilterByReferenceAlias) {+ auto schema_json =+ R"({+ "name": "Products",+ "fields": [+ {"name": "product_id", "type": "string"},+ {"name": "product_name", "type": "string", "infix": true},+ {"name": "product_description", "type": "string"},+ {"name": "embedding", "type":"float[]", "embed":{"from": ["product_description"], "model_config": {"model_name": "ts/e5-small"}}},+ {"name": "rating", "type": "int32"}+ ]+ })"_json;+ std::vector<nlohmann::json> documents = {+ R"({+ "product_id": "product_a",+ "product_name": "shampoo",+ "product_description": "Our new moisturizing shampoo is perfect for those with dry or damaged hair.",+ "rating": "2"+ })"_json,+ R"({+ "product_id": "product_b",+ "product_name": "soap",+ "product_description": "Introducing our all-natural, organic soap bar made with essential oils and botanical ingredients.",+ "rating": "4"+ })"_json+ };++ TextEmbedderManager::set_model_dir("/tmp/typesense_test/models");++ auto collection_create_op = collectionManager.create_collection(schema_json);+ ASSERT_TRUE(collection_create_op.ok());+ for (auto const &json: documents) {+ auto add_op = collection_create_op.get()->add(json.dump());+ if (!add_op.ok()) {+ LOG(INFO) << add_op.error();+ }+ ASSERT_TRUE(add_op.ok());+ }++ schema_json =+ R"({+ "name": "Customers",+ "fields": [+ {"name": "customer_id", "type": "string"},+ {"name": "customer_name", "type": "string"},+ {"name": "product_price", "type": "float"},+ {"name": "product_id", "type": "string", "reference": "Products.product_id"}+ ]+ })"_json;+ documents = {+ R"({+ "customer_id": "customer_a",+ "customer_name": "Joe",+ "product_price": 143,+ "product_id": "product_a"+ })"_json,+ R"({+ "customer_id": "customer_a",+ "customer_name": "Joe",+ "product_price": 73.5,+ "product_id": "product_b"+ })"_json,+ R"({+ "customer_id": "customer_b",+ "customer_name": "Dan",+ "product_price": 75,+ "product_id": "product_a"+ })"_json,+ R"({+ "customer_id": "customer_b",+ "customer_name": "Dan",+ "product_price": 140,+ "product_id": "product_b"+ })"_json+ };+ collection_create_op = collectionManager.create_collection(schema_json);+ ASSERT_TRUE(collection_create_op.ok());+ for (auto const &json: documents) {+ auto add_op = collection_create_op.get()->add(json.dump());+ if (!add_op.ok()) {+ LOG(INFO) << add_op.error();+ }+ ASSERT_TRUE(add_op.ok());+ }++ auto symlink_op = collectionManager.upsert_symlink("Products_alias", "Products");+ ASSERT_TRUE(symlink_op.ok());++ symlink_op = collectionManager.upsert_symlink("$Customers_alias", "$Customers");+ ASSERT_TRUE(symlink_op.ok());++ std::map<std::string, std::string> req_params = {+ {"collection", "Products_alias"},+ {"q", "*"},+ {"query_by", "product_name"},+ {"filter_by", "$Customers(customer_id:=customer_a && product_price:<100)"},+ };+ nlohmann::json embedded_params;+ std::string json_res;+ auto now_ts = std::chrono::duration_cast<std::chrono::microseconds>(+ std::chrono::system_clock::now().time_since_epoch()).count();++ auto search_op = collectionManager.do_search(req_params, embedded_params, json_res, now_ts);++ nlohmann::json res_obj = nlohmann::json::parse(json_res);+ ASSERT_EQ(1, res_obj["found"].get<size_t>());+ ASSERT_EQ(1, res_obj["hits"].size());+ // No fields are mentioned in `include_fields`, should include all fields of Products and Customers by default.+ ASSERT_EQ(7, res_obj["hits"][0]["document"].size());+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_name"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_description"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("embedding"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("rating"));+ // Default strategy of reference includes is nest. No alias was provided, collection name becomes the field name.+ ASSERT_EQ(5, res_obj["hits"][0]["document"]["Customers"].size());+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("customer_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("customer_name"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("product_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("product_price"));++ req_params = {+ {"collection", "Products"},+ {"q", "*"},+ {"query_by", "product_name"},+ {"filter_by", "$Customers_alias(customer_id:=customer_a && product_price:<100)"},+ };+ search_op = collectionManager.do_search(req_params, embedded_params, json_res, now_ts);++ res_obj = nlohmann::json::parse(json_res);+ ASSERT_EQ(1, res_obj["found"].get<size_t>());+ ASSERT_EQ(1, res_obj["hits"].size());+ // No fields are mentioned in `include_fields`, should include all fields of Products and Customers by default.+ ASSERT_EQ(7, res_obj["hits"][0]["document"].size());+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_name"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_description"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("embedding"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("rating"));+ // Default strategy of reference includes is nest. No alias was provided, collection name becomes the field name.+ ASSERT_EQ(5, res_obj["hits"][0]["document"]["Customers"].size());+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("customer_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("customer_name"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("product_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("product_price"));++ req_params = {+ {"collection", "Products_alias"},+ {"q", "*"},+ {"query_by", "product_name"},+ {"filter_by", "$Customers_alias(customer_id:=customer_a && product_price:<100)"},+ };+ search_op = collectionManager.do_search(req_params, embedded_params, json_res, now_ts);++ res_obj = nlohmann::json::parse(json_res);+ ASSERT_EQ(1, res_obj["found"].get<size_t>());+ ASSERT_EQ(1, res_obj["hits"].size());+ // No fields are mentioned in `include_fields`, should include all fields of Products and Customers by default.+ ASSERT_EQ(7, res_obj["hits"][0]["document"].size());+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_name"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("product_description"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("embedding"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"].count("rating"));+ // Default strategy of reference includes is nest. No alias was provided, collection name becomes the field name.+ ASSERT_EQ(5, res_obj["hits"][0]["document"]["Customers"].size());+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("customer_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("customer_name"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("product_id"));+ ASSERT_EQ(1, res_obj["hits"][0]["document"]["Customers"].count("product_price"));+}\ No newline at end of file