#pragma once #include #include #include #include #include #include #include #include class Collection { private: Store* store; std::string name; spp::sparse_hash_map schema; // Integer ID used internally for bitmaps - not exposed to the client uint32_t seq_id; spp::sparse_hash_map index_map; spp::sparse_hash_map doc_scores; uint32_t next_seq_id(); const std::string SEQ_ID_PREFIX = "SQ_"; const std::string ID_PREFIX = "ID_"; const std::string META_PREFIX = "MT_"; const std::string FIELDS_KEY = META_PREFIX + "_fields"; std::string get_seq_id_key(uint32_t seq_id); std::string get_id_key(std::string id); static inline std::vector next_suggestion(const std::vector> &token_leaves, long long int n); void log_leaves(const int cost, const std::string &token, const std::vector &leaves) const; void search_candidates(std::vector> & token_leaves, std::vector & results, spp::sparse_hash_set & dedup_seq_ids, size_t & total_results, const size_t & max_results); void index_string_field(const std::string &field_name, art_tree *t, const nlohmann::json &document, uint32_t seq_id) const; void index_int32_field(const std::string &field_name, art_tree *t, const nlohmann::json &document, uint32_t seq_id) const; public: Collection() = delete; Collection(const std::string & state_dir_path, const std::string & name, const std::vector & fields); ~Collection(); std::string add(std::string json_str); std::vector search(std::string query, const int num_typos, const size_t num_results, const token_ordering token_order = FREQUENCY, const bool prefix = false); void remove(std::string id); void score_results(Topster<100> &topster, const std::vector &query_suggestion, const uint32_t *result_ids, const size_t result_size) const; enum {MAX_SEARCH_TOKENS = 20}; enum {MAX_RESULTS = 100}; };