tfrecords 0.0.8

tfrecords

simplify and transplant the tfrecord dataset

update information

    2022-10-17:  Add shared memory read mode with more accelerated Reading.
    2022-02-01:  simplify and transplant the tfrecord dataset

record read and write demo , with_share_memory flags will Accelerated Reading

record 读写demo , 建议开启 参数with_share_memory , 将会大大加速文件读取

import tfrecords
options = tfrecords.TFRecordOptions(compression_type=tfrecords.TFRecordCompressionType.NONE)

def test_write(filename, N=3, context='aaa'):
    with tfrecords.TFRecordWriter(filename, options=options) as file_writer:
        for _ in range(N):
            # x, y = np.random.random(), np.random.random()
            file_writer.write(context + '____' + str(_))

def test_record_iterator(example_paths):
    print('test_record_iterator')
    for example_path in example_paths:
        iterator = tfrecords.tf_record_iterator(example_path, options=options,with_share_memory=False)
        num = 0
        for iter in iterator:
            num += 1
            print(iter)

def test_random_reader(example_paths):
    print('test_random_reader')
    for example_path in example_paths:
        file_reader = tfrecords.tf_record_random_reader(example_path, options=options,with_share_memory=False)
        last_pos = 0
        while True:
            try:
                x,pos = file_reader.read(last_pos)
                print(x)
                last_pos = pos
            except:
                break

def test_random_reader2(example_paths):
    print('test_random_reader2')
    for example_path in example_paths:
        file_reader = tfrecords.tf_record_random_reader(example_path, options=options,with_share_memory=False)
        skip_bytes = 0
        offset_list = file_reader.read_offsets(skip_bytes)
        for offset,length in offset_list:
            x, _ = file_reader.read(offset)
            print(x)

test_write('d:/example.tfrecords0',3,'file0')
test_write('d:/example.tfrecords1',10,'file1')
test_write('d:/example.tfrecords2',12,'file2')
example_paths = tfrecords.glob('d:/example.tfrecords*')


test_record_iterator(example_paths)
print()
test_random_reader2(example_paths)
print()
test_random_reader2(example_paths)
print()