treebank.hpp

/*
 * CoNLL-U - https://github.com/tugrulgungor/CoNLL-U
 *
 * Copyright (c) 2026. All rights reserved.
 * Tuğrul Güngör - https://conll-u.tugrulgungor.me
 *
 * Distributed under the MIT License.
 * https://opensource.org/license/mit/
 */

 
#ifndef TG_CONLLU_TREEBANK_HPP
#define TG_CONLLU_TREEBANK_HPP 1
 
#include "conllu/error.hpp"
#include "conllu/internal/string_hash.hpp"
#include "conllu/sentence.hpp"
 
#include <deque>
#include <expected>
#include <ranges>
#include <string>
#include <unordered_map>
#include <unordered_set>
#include <vector>
 
namespace tg::conllu {
    class Treebank final {
        std::deque<std::string> columns_{};
        std::vector<std::string_view> column_views_{};
        std::unordered_map<std::string_view, std::size_t, internal::StringHash, std::equal_to<>> column_idx_{};
 
        std::vector<std::unique_ptr<Sentence>> sentences_{};
        std::unordered_map<std::string_view, std::size_t, internal::StringHash, std::equal_to<>> sentence_idx_{};
 
        friend class Parser;
        friend class Sentence;
        friend class Token;
        friend class Word;
        friend class Writer;
 
    public:
        constexpr Treebank() noexcept = default;
        constexpr Treebank(const Treebank&) noexcept = delete;
        constexpr Treebank(Treebank&& other) noexcept = delete;
        [[nodiscard]] constexpr const std::vector<std::string_view>& columns() const noexcept;
        [[nodiscard]] constexpr std::size_t num_columns() const noexcept;
        [[nodiscard]] constexpr std::optional<std::size_t> get_column_pos(std::string_view name) const noexcept;
        [[nodiscard]] constexpr std::optional<std::string_view> get_column_name(std::size_t position) const noexcept;
        constexpr std::expected<std::size_t, std::error_code> add_column(std::string_view name) noexcept;
        [[nodiscard]] constexpr auto sentences() const noexcept;
        [[nodiscard]] constexpr std::size_t num_sentences() const noexcept;
        [[nodiscard]] constexpr std::expected<Sentence*, std::error_code> add_sentence(
        std::string_view sent_id, std::string_view text = std::string_view()) noexcept;
        [[nodiscard]] constexpr std::optional<Sentence*> get_sentence_by_id(std::string_view sent_id) const noexcept;
        constexpr std::error_code remove_sentence(std::string_view sent_id) noexcept;
        constexpr std::error_code remove_sentence(const Sentence* sentence) noexcept;
        [[nodiscard]] constexpr bool empty() const noexcept;
        constexpr void clear() noexcept;
        [[nodiscard]] constexpr std::error_code validate() const noexcept;
        constexpr Treebank& operator=(Treebank&& other) noexcept = delete;
        constexpr Treebank& operator=(const Treebank&) noexcept = delete;
        constexpr ~Treebank() noexcept = default;
    };
} // namespace tg::conllu
 
#include "conllu/impl/sentence.ipp"
#include "conllu/impl/token.ipp"
#include "conllu/impl/treebank.ipp"
#include "conllu/impl/word.ipp"
 
#endif // TG_CONLLU_TREEBANK_HPP