Missing 2 required positional arguments.

[Lothedr]

Short description
Calling tfds.builder('mnist') causes exception:

TypeError: __init__() missing 2 required positional arguments: 'op' and 'message'

Environment information

• Operating System: Windows 10 Pro
• Python version: 3.8.3 (default, Jul 2 2020, 17:30:36) [MSC v.1916 64 bit (AMD64)]
• tensorflow-datasets version: 3.2.0
• tensorflow version: 2.2.0

Reproduction instructions

import tensorflow as tf
import tensorflow_datasets as tfds
builder = tfds.builder('mnist')

Link to logs
Traceback

Expected behavior
I expected that tfds will not throw an exception.

Additional context
There seems to be an implementation error here. In my case, exc_type is UnimplmentedError and its constructor needs more arguments according to the implementation.

[vijayphoenix]

@Conchylicultor
Minimal code to reproduce the issue is

import tensorflow as tf
tf.io.gfile.exists("gs://tfds-data/dataset_info/mnist/3.0.1")

If I remember correctly, TensorFlow has does not support for Google Cloud Storage (GCS) on Windows operating systems.

However, I already have commented on the existing TF issue tensorflow/tensorflow#38477 (comment)

[Conchylicultor]

Sorry for the breakage, it was introduced by using tf.io.gfile instead of request when we released 3.2.0. In the meantime you can downgrade to 3.1.0.

Marking this as duplicate of #2190