Make torchaudio work on Python 2. (#209)

- Apply __future__ imports uniformly (future division is the biggy, but absolute
  imports mattered too)
- Hotfix use of tempfile.TemporaryDirectory using a BC library (DO NOT
  add this library as a dependency to torchaudio; it's only for testing)
- Replace math.gcd with fractions.gcd
- Fix a weird pytest collection bug involving parametrized tests
- Turn on Python 2 and Python 3.7 in Travis.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>

torchaudio/datasets/yesno.py

-from __future__ import print_function
+from __future__ import absolute_import, division, print_function, unicode_literals
 import torch.utils.data as data
 import os
 import os.path

torchaudio/functional.py

-from __future__ import division
+from __future__ import absolute_import, division, print_function, unicode_literals
 import math
 import torch
 

torchaudio/kaldi_io.py

+from __future__ import absolute_import, division, print_function, unicode_literals
 # To use this file, the dependency (https://github.com/vesis84/kaldi-io-for-python)
 # needs to be installed. This is a light wrapper around kaldi_io that returns
 # torch.Tensors.

torchaudio/sox_effects.py

-from __future__ import division, print_function
+from __future__ import absolute_import, division, print_function, unicode_literals
 import torch
 import _torch_sox
 

torchaudio/transforms.py

-from __future__ import division, print_function
+from __future__ import absolute_import, division, print_function, unicode_literals
 from warnings import warn
 import math
 import torch