vocabulary.py

# coding: utf-8
"""
Vocabulary module
"""
import sys
from collections import Counter
from pathlib import Path
from types import SimpleNamespace
from typing import Dict, List, Tuple

import numpy as np

from joeynmt.datasets import BaseDataset
from joeynmt.helpers import flatten, read_list_from_file, write_list_to_file
from joeynmt.helpers_for_ddp import get_logger

logger = get_logger(__name__)


class Vocabulary:
    """Vocabulary represents mapping between tokens and indices."""

    def __init__(self, tokens: List[str], cfg: SimpleNamespace) -> None:
        """
        Create vocabulary from list of tokens.
        Special tokens are added if not already in list.

        :param tokens: list of tokens
        :param cfg: special symbols defined in config
        """
        # warning: stoi grows with unknown tokens, don't use for saving or size

        # special symbols
        self.specials = [cfg.unk_token, cfg.pad_token, cfg.bos_token, cfg.eos_token]
        self.lang_tags = cfg.lang_tags
        if cfg.sep_token:
            self.specials.append(cfg.sep_token)

        # don't allow to access _stoi and _itos outside of this class
        self._stoi: Dict[str, int] = {}  # string to index
        self._itos: List[str] = []  # index to string

        # construct
        self.add_tokens(tokens=self.specials + self.lang_tags + tokens)
        assert len(self._stoi) == len(self._itos)

        # assign after stoi is built
        self.pad_index = cfg.pad_id
        self.bos_index = cfg.bos_id
        self.eos_index = cfg.eos_id
        self.unk_index = cfg.unk_id
        self.sep_index = cfg.sep_id if cfg.sep_token else None
        assert self.pad_index == self.lookup(cfg.pad_token)
        assert self.bos_index == self.lookup(cfg.bos_token)
        assert self.eos_index == self.lookup(cfg.eos_token)
        assert self.unk_index == self.lookup(cfg.unk_token)
        assert self._itos[cfg.unk_id] == cfg.unk_token

        if cfg.sep_token:
            assert self.sep_index == self.lookup(cfg.sep_token)

    def add_tokens(self, tokens: List[str]) -> None:
        """
        Add list of tokens to vocabulary

        :param tokens: list of tokens to add to the vocabulary
        """
        for t in tokens:
            new_index = len(self._itos)
            # add to vocab if not already there
            if t not in self._itos:
                self._itos.append(t)
                self._stoi[t] = new_index

    def to_file(self, file: Path) -> None:
        """
        Save the vocabulary to a file, by writing token with index i in line i.

        :param file: path to file where the vocabulary is written
        """
        write_list_to_file(file, self._itos)

    def is_unk(self, token: str) -> bool:
        """
        Check whether a token is covered by the vocabulary

        :param token:
        :return: True if covered, False otherwise
        """
        return self.lookup(token) == self.unk_index

    def lookup(self, token: str) -> int:
        """
        look up the encoding dictionary. (needed for multiprocessing)

        :param token: surface str
        :return: token id
        """
        return self._stoi.get(token, self.unk_index)

    def __len__(self) -> int:
        return len(self._itos)

    def __eq__(self, other) -> bool:
        if isinstance(other, Vocabulary):
            return self._itos == other._itos
        return False

    def _array_to_sentence(
        self,
        array: np.ndarray,
        cut_at_eos: bool = True,
        skip_pad: bool = True
    ) -> List[str]:
        """
        Converts an array of IDs to a sentence, optionally cutting the result off at the
        end-of-sequence token.

        :param array: 1D array containing indices
        :param cut_at_eos: cut the decoded sentences at the first <eos>
        :param skip_pad: skip generated <pad> tokens
        :return: list of strings (tokens)
        """
        sentence = []
        for i in array:
            if skip_pad and i == self.pad_index:
                continue

            s = self._itos[i]  # decode back to token surface
            sentence.append(s)

            # break at the position AFTER eos
            if cut_at_eos and i == self.eos_index:
                break
        return sentence

    def arrays_to_sentences(
        self,
        arrays: np.ndarray,
        cut_at_eos: bool = True,
        skip_pad: bool = True
    ) -> List[List[str]]:
        """
        Convert multiple arrays containing sequences of token IDs to their sentences,
        optionally cutting them off at the end-of-sequence token.

        :param arrays: 2D array containing indices
        :param cut_at_eos: cut the decoded sentences at the first <eos>
        :param skip_pad: skip generated <pad> tokens
        :return: list of list of strings (tokens)
        """
        return [
            self._array_to_sentence(array, cut_at_eos, skip_pad) for array in arrays
        ]

    def sentences_to_ids(
        self,
        sentences: List[List[str]],
        bos: bool = True,
        eos: bool = True,
    ) -> Tuple[List[List[int]], List[int], List[int]]:
        """
        Encode sentences to indices and pad sequences to the maximum length of the
        sentences given

        :param sentences: list of tokenized sentences
        :param bos: whether to add <bos>
        :param eos: whether to add <eos>
        :return:
            - padded ids
            - original lengths before padding
            - prompt_mask
        """
        max_len = max([len(sent) for sent in sentences])
        if bos:
            max_len += 1
        if eos:
            max_len += 1
        padded, lengths, prompt_mask = [], [], []
        for sent in sentences:
            encoded = [self.lookup(s) for s in sent]
            if bos:
                encoded = [self.bos_index] + encoded
            if eos:
                encoded = encoded + [self.eos_index]
            offset = max(0, max_len - len(encoded))
            padded.append(encoded + [self.pad_index] * offset)
            lengths.append(len(encoded))

            try:
                sep_pos = encoded.index(self.sep_index) + 1
                prompt_mask.append([1] * sep_pos + [0] * (max_len - sep_pos))
            except ValueError as e:  # pylint: disable=unused-variable # noqa: F841
                # if SEP_ID not found, fill with zeros
                prompt_mask.append([0] * max_len)
        return padded, lengths, prompt_mask

    def log_vocab(self, k: int) -> str:
        """first k vocab entities"""
        return " ".join(f"({i}) {t}" for i, t in enumerate(self._itos[:k]))

    def __repr__(self) -> str:
        return (
            f"{self.__class__.__name__}(len={self.__len__()}, "
            f"specials={self.specials}, lang_tags={self.lang_tags})"
        )


def sort_and_cut(counter: Counter,
                 max_size: int = sys.maxsize,
                 min_freq: int = -1) -> List[str]:
    """
    Cut counter to most frequent, sorted numerically and alphabetically
    :param counter: flattened token list in Counter object
    :param max_size: maximum size of vocabulary
    :param min_freq: minimum frequency for an item to be included
    :return: list of valid tokens
    """
    # filter counter by min frequency
    if min_freq > -1:
        counter = Counter({t: c for t, c in counter.items() if c >= min_freq})

    # sort by frequency, then alphabetically
    tokens_and_frequencies = sorted(counter.items(), key=lambda tup: tup[0])
    tokens_and_frequencies.sort(key=lambda tup: tup[1], reverse=True)

    # cut off
    vocab_tokens = [i[0] for i in tokens_and_frequencies[:max_size]]
    assert len(vocab_tokens) <= max_size, (len(vocab_tokens), max_size)
    return vocab_tokens


def _build_vocab(
    cfg: Dict,
    special_symbols: SimpleNamespace,
    dataset: BaseDataset = None
) -> Vocabulary:
    """
    Builds vocabulary either from file or sentences.

    :param cfg: data cfg
    :param special_symbols: special symbols
    :param dataset: dataset object which contains preprocessed sentences
    :return: Vocabulary created from either `tokens` or `vocab_file`
    """
    vocab_file = cfg.get("voc_file", None)
    min_freq = cfg.get("voc_min_freq", 1)  # min freq for an item to be included
    max_size = int(cfg.get("voc_limit", sys.maxsize))  # max size of vocabulary
    assert max_size > 0

    if vocab_file is not None:
        # load it from file (not to apply `sort_and_cut()`)
        unique_tokens = read_list_from_file(Path(vocab_file))
    elif dataset is not None:
        # tokenize sentences (no subsampling)
        sents = dataset.get_list(lang=cfg["lang"], tokenized=True, subsampled=False)

        # newly create unique token list (language-wise, no joint-vocab)
        counter = Counter(flatten(sents))
        unique_tokens = sort_and_cut(counter, max_size, min_freq)
    else:
        raise ValueError("Please provide a vocab file path or dataset.")

    vocab = Vocabulary(unique_tokens, special_symbols)
    assert len(vocab) <= max_size + len(vocab.specials + vocab.lang_tags), \
        (len(vocab), max_size)

    # check for all special symbols except for UNK token whether they are not OOVs
    for s in vocab.specials[1:] + vocab.lang_tags:
        assert not vocab.is_unk(s)

    return vocab


def build_vocab(cfg: Dict,
                dataset: BaseDataset = None,
                model_dir: Path = None) -> Tuple[Vocabulary, Vocabulary]:
    # use the vocab file saved in model_dir
    if model_dir is not None and cfg["src"].get("voc_file", None) is None:
        assert (model_dir / "src_vocab.txt").is_file()
        cfg["src"]["voc_file"] = (model_dir / "src_vocab.txt").as_posix()
    if model_dir is not None and cfg["trg"].get("voc_file", None) is None:
        assert (model_dir / "trg_vocab.txt").is_file()
        cfg["trg"]["voc_file"] = (model_dir / "trg_vocab.txt").as_posix()

    src_vocab = _build_vocab(cfg["src"], cfg["special_symbols"], dataset)
    trg_vocab = _build_vocab(cfg["trg"], cfg["special_symbols"], dataset)

    assert src_vocab.pad_index == trg_vocab.pad_index
    assert src_vocab.bos_index == trg_vocab.bos_index
    assert src_vocab.eos_index == trg_vocab.eos_index
    assert src_vocab.sep_index == trg_vocab.sep_index
    return src_vocab, trg_vocab