tensorflow tf.python_io模块源码阅读

1.序言
该模块是tensorflow用来处理tfrecords文件的接口，定义在tensorflow/python/lib/io/python_io.py，主要包含了四个部分：
class TFRecordCompressionType：记录的压缩类型。
class TFRecordOptions：用于操作TFRecord文件的选项。
class TFRecordWriter：将记录写入TFRecords文件的类。
tf_record_iterator(…)：从TFRecords文件中读取记录的迭代器

2.源码解析

from __future__ import absolute_import
from __future__ import division
from __future__ import print_function

from tensorflow.python import pywrap_tensorflow
from tensorflow.python.framework import errors
from tensorflow.python.util import compat

#该类定义了tfrecords文件压缩类型：无，ZLIB，GZIP三种
class TFRecordCompressionType(object):
  """The type of compression for the record."""
  NONE = 0
  ZLIB = 1
  GZIP = 2


# 这个类会转换为proto格式，以便与C++接口对接
class TFRecordOptions(object):
  """Options used for manipulating TFRecord files."""
  compression_type_map = {
      TFRecordCompressionType.ZLIB: "ZLIB",
      TFRecordCompressionType.GZIP: "GZIP",
      TFRecordCompressionType.NONE: ""
  }

  def __init__(self, compression_type):
    self.compression_type = compression_type

  @classmethod
  def get_compression_type_string(cls, options):
    if not options:
      return ""
    return cls.compression_type_map[options.compression_type]


def tf_record_iterator(path, options=None):
  """从tfrecords文件读取数据的迭代器.
  参数:
    path: TFRecords文件路径.
    options: 读取选项，主要是压缩类型，TFRecordOptions对象.
  yields:
    Strings.
  异常:
    IOError: 路径不正确是引发.
  """
  compression_type = TFRecordOptions.get_compression_type_string(options)
  with errors.raise_exception_on_not_ok_status() as status:
    reader = pywrap_tensorflow.PyRecordReader_New(
        compat.as_bytes(path), 0, compat.as_bytes(compression_type), status)#读取器，pywarp_tensorflow包装所以的符号，这里定义了一个文件读取器对象

  if reader is None:
    raise IOError("Could not open %s." % path)
  while True:
    try:
      with errors.raise_exception_on_not_ok_status() as status:
        reader.GetNext(status)
    except errors.OutOfRangeError:
      break
    yield reader.record() #逐步读取文件
  reader.Close()


class TFRecordWriter(object):
  """tfrecords文件写操作类，由于实施了__enter__和__exit__接口，根据Python的上下文管理机制，可以用with语句
  """

  # TODO(josh11b): Support appending?
  def __init__(self, path, options=None):
    """打开文件，并初始化写对象
    参数:
      path: 文件路径
      options: 选项，TFRecordOptions对象
    Raises:
      IOError: If `path` cannot be opened for writing.
    """
    compression_type = TFRecordOptions.get_compression_type_string(options)#获取压缩类型

    with errors.raise_exception_on_not_ok_status() as status:
      self._writer = pywrap_tensorflow.PyRecordWriter_New(
          compat.as_bytes(path), compat.as_bytes(compression_type), status)#定义writer

  def __enter__(self):
    """进入with语句块"""
    return self

  def __exit__(self, unused_type, unused_value, unused_traceback):
    """退出with语句块，并关闭文件"""
    self.close()

  def write(self, record):
    """想文件中写入一条记录.
    Args:
      record: str
    """
    self._writer.WriteRecord(record)#实际是由writer实现

  def flush(self):
    """刷新缓冲区内容到磁盘文件"""
    with errors.raise_exception_on_not_ok_status() as status:
      self._writer.Flush(status)

  def close(self):
    """关闭文件"""
    with errors.raise_exception_on_not_ok_status() as status:
      self._writer.Close(status)