Source code for

# Copyright 2019 The Texar Authors. All Rights Reserved.
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# See the License for the specific language governing permissions and
# limitations under the License.
Base text data class that is inherited by all text data classes.
import io
import locale
from abc import ABC
from typing import IO, Iterator, List, Optional, TypeVar

import torch
from import DatasetBase, DataSource
from texar.torch.utils.types import MaybeList

__all__ = [

RawExample = TypeVar('RawExample')
Example = TypeVar('Example')

[docs]class TextLineDataSource(DataSource[List[str]]): r"""Data source for reading from (multiple) text files. Each line is tokenized and yielded as an example. This data source does not support indexing. Args: file_paths (str or list[str]): Paths to the text files. compression_type (str, optional): The compression type for the text files, ``"gzip"`` and ``"zlib"`` are supported. Default is `None`, in which case files are treated as plain text files. encoding (str, optional): Encoding for the files. By default uses the default locale of the system (usually UTF-8). delimiter (str, optional): Delimiter for tokenization purposes. This is used in combination with ``max_length``. If `None`, text is split on any blank character. max_length (int, optional): Maximum length for data examples. Length is measured as the number of tokens in a line after being tokenized using the provided ``delimiter``. Lines with more than ``max_length`` tokens will be dropped. """ def __init__(self, file_paths: MaybeList[str], compression_type: Optional[str] = None, encoding: Optional[str] = None, delimiter: Optional[str] = None, max_length: Optional[int] = None): if compression_type is not None: compression_type = compression_type.lower() if compression_type not in ['gzip', 'zlib']: raise ValueError( f"Unsupported compression type: {compression_type}") if isinstance(file_paths, str): file_paths = [file_paths] self._compression_type = compression_type self._encoding = encoding or locale.getpreferredencoding() self._file_paths = file_paths self._max_length = max_length self._delimiter = delimiter class _ZlibWrapper(io.BufferedReader): def __init__(self, raw: IO[bytes]): super().__init__(raw) # type: ignore import zlib self.file = raw self.zlib = zlib.decompressobj() self.buffer = b'' @property def closed(self) -> bool: return self.file.closed def readable(self) -> bool: return True def close(self) -> None: self.file.close() def read1(self, n: int = -1) -> bytes: # Our implementation does not really satisfy the definition for # `read1`, but whatever, it seems to work. if n == -1: raw = b = self.buffer self.buffer = b'' if raw: cur = self.zlib.decompress(raw) if len(cur) > 0: if len(b) > 0: b = b + cur else: b = cur else: if len(self.buffer) > 0: b = self.buffer[:n] self.buffer = self.buffer[n:] return b while True: raw = if not raw: return b'' b = self.zlib.decompress(raw) if len(b) > 0: break if len(b) > n: self.buffer = b[n:] return b[:n] return b def read(self, n: Optional[int] = -1) -> bytes: if n is None or n < 0: n = -1 return self.read1(n) def __getattr__(self, item): return getattr(self.file, item) def _open_file(self, path: str) -> IO[str]: if self._compression_type == 'zlib': f: IO[str] = io.TextIOWrapper( self._ZlibWrapper(open(path, 'rb')), encoding=self._encoding) elif self._compression_type == 'gzip': import gzip f =, 'rt', encoding=self._encoding) else: f = open(path, 'r', encoding=self._encoding) return f def __iter__(self) -> Iterator[List[str]]: for path in self._file_paths: with self._open_file(path) as f: for line in f: tokens = line.split(self._delimiter) if (self._max_length is not None and len(tokens) > self._max_length): continue yield tokens
class TextDataBase(DatasetBase[RawExample, Example], ABC): r"""Base class inherited by all text data classes. """ def __init__(self, source: DataSource[RawExample], hparams, device: Optional[torch.device] = None): super().__init__(source, hparams, device=device) @staticmethod def default_hparams(): r"""Returns a dictionary of default hyperparameters. See the specific subclasses for the details. """ hparams = DatasetBase.default_hparams() hparams.update({ "bucket_boundaries": [], "bucket_batch_sizes": None, "bucket_length_fn": None}) return hparams