[TFRecord] The easy way to verify your TFRecord file

There is a common situation when you build your TFRecord file ( your dataset ) and want to verify the correctness of the data in it. How to do it? I assume you don't have the problem to build your TFRecord file. So, the easy way to verify your TFRecord file is to use the API: tf.python_io.tf_record_iterator(). 

The following points need to check out carefully:
1. Use XXX.value without [] if the restored data is a list
2. For the image part, the data type should be consistent between writing and reading in TFRecord.

The example code:

#!/usr/bin/python
# -*- coding: utf-8 -*-
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt

def read_ftrecord(tffile):
  reader = tf.TFRecordReader()
  example = tf.train.Example()
  for r in tf.python_io.tf_record_iterator(tffile):
    example.ParseFromString(r)
    width = int(example.features.feature['image/width'].int64_list.value[0])
    height = int(example.features.feature['image/height'].int64_list.value[0])
    filename = example.features.feature['image/filename'].bytes_list.value[0]
    encoded = example.features.feature['image/encoded'].bytes_list.value[0]
    label = example.features.feature['image/class/label'].int64_list.value
    text = example.features.feature['image/class/text'].bytes_list.value[0]
  
    data = {
      "width": width,
      "height": height,
      "filename": filename,
      "encoded": encoded,
      "label": label,
      "text": text,
    }
    yield data

tfrecords_filename = 'my_dataset.tfrecords' 
gen_data = read_ftrecord(tfrecords_filename)

for i in range(5):
  my_data = gen_data.next()
  image_1d = np.fromstring(my_data["encoded"], dtype=np.int8)
  image = image_1d.reshape((my_data["width"], my_data["height"]))
  print(my_data["width"], 
      my_data["height"], 
      my_data["filename"], 
      image.shape, 
      my_data["label"], 
      my_data["text"])

  plt.title('image:' + str(i))
  plt.imshow(image, plt.cm.gray)
  plt.show()