Move from tf flags to absl flags, since they seem to be gone in tf 1.15

PiperOrigin-RevId: 281869018

setup.py

@@ -33,6 +33,7 @@
         'tensor2tensor/bin/t2t-translate-all',
     ],
     install_requires=[
+        'absl-py',
         'bz2file',
         'dopamine-rl',
         'flask',
@@ -64,7 +65,6 @@
         'tensorflow': ['tensorflow>=1.15.0'],
         'tensorflow-hub': ['tensorflow-hub>=0.1.1'],
         'tests': [
-            'absl-py',
             # Needed to fix a Travis pytest error.
             # https://github.com/Julian/jsonschema/issues/449#issuecomment-411406525
             'attrs>=17.4.0',

tensor2tensor/data_generators/audio.py

@@ -19,14 +19,12 @@
 from __future__ import print_function
 
 import os
-from subprocess import call
+import subprocess
 import tarfile
 import wave
-# from tensor2tensor.data_generators import generator_utils
-
+from absl import flags
 import tensorflow as tf
 
-flags = tf.flags
 FLAGS = flags.FLAGS
 
 flags.DEFINE_string("timit_paths", "",
@@ -78,7 +76,7 @@ def _get_audio_data(filepath):
   # Construct a true .wav file.
   out_filepath = filepath.strip(".WAV") + ".wav"
   # Assumes sox is installed on system. Sox converts from NIST SPHERE to WAV.
-  call(["sox", filepath, out_filepath])
+  subprocess.call(["sox", filepath, out_filepath])
   wav_file = wave.open(open(out_filepath))
   frame_count = wav_file.getnframes()
   byte_array = wav_file.readframes(frame_count)