Merge pull request #75 from finos/matthew/FixAudioSplitter

fix audio splitter

.travis.yml

@@ -8,6 +8,9 @@ python:
   - "3.7"
   - "3.7-dev"  # 3.7 development branch
 # command to install dependencies
+before_install:
+  - sudo apt-get update
+  - sudo apt-get install -y sox libsox-fmt-mp3
 install:
   - pip install -r requirements.txt
   - pip install numpy

asrtoolkit/convert_transcript.py

@@ -4,12 +4,24 @@
 """
 
 import argparse
+import sys
 import logging
 from asrtoolkit.file_utils.script_input_validation import assign_if_valid
 LOGGER = logging.getLogger(__name__)
 
 
-def convert():
+def check_input_file_validity(input_file):
+    if input_file is None:
+        LOGGER.error("Invalid input file {}".format(input_file))
+        sys.exit(1)
+
+
+def convert(input_file, output_file):
+    input_file = assign_if_valid(input_file)
+    input_file.write(output_file)
+
+
+def main():
     parser = argparse.ArgumentParser(
         description='convert between text file formats')
     parser.add_argument('input_file',
@@ -22,13 +34,10 @@ def convert():
                         help='output file')
     args = parser.parse_args()
 
-    input_file = assign_if_valid(args.input_file)
+    check_input_file_validity(args.input_file)
 
-    if input_file is not None:
-        input_file.write(args.output_file)
-    else:
-        LOGGER.error("Invalid input file {}".format(args.input_file))
+    convert(args.input_file, args.output_file)
 
 
 if __name__ == "__main__":
-    convert()
+    main()

asrtoolkit/file_utils/name_cleaners.py

@@ -5,6 +5,7 @@
 
 import os
 
+
 def get_extension(file_name):
     """
     Returns file extension
@@ -44,7 +45,8 @@ def replace_char(input_char, file_name):
                 "check to make sure your audio files and transcript files match"
             )
         return os.sep.join(
-            file_name.split(os.sep)[:-1] + [basename(file_name).replace(c, "_")])
+            file_name.split(os.sep)[:-1] +
+            [basename(file_name).replace(c, "_")])
 
     for c in chars_to_replace:
         file_name = replace_char(c, file_name)
@@ -63,5 +65,5 @@ def generate_segmented_file_name(target_dir, file_name, iseg):
     """
     Take a target location, a current location, and a segment number and generate a target filename
     """
-    return target_dir + os.sep + basename(strip_extension(file_name)) + \
-              "_seg_{:05d}.".format(iseg) + file_name.split(".")[-1]
+    return sanitize_hyphens(target_dir + os.sep + basename(strip_extension(file_name)) + \
+              "_seg_{:05d}.".format(iseg) + file_name.split(".")[-1])

asrtoolkit/split_audio_file.py

@@ -14,6 +14,29 @@
 LOGGER = logging.getLogger(__name__)
 
 
+def split_audio_file(source_audio_file, source_transcript, target_directory):
+    """
+    Execute the split logic
+    """
+    source_audio = audio_file(source_audio_file)
+    transcript = time_aligned_text(source_transcript)
+    source_audio.split(transcript, target_directory)
+
+
+def validate_transcript(transcript):
+    "exit if invalid transcript"
+    if not valid_input_file(transcript):
+        LOGGER.error("Invalid transcript file {}".format(transcript))
+        sys.exit(1)
+
+
+def validate_audio_file(source_audio_file):
+    if not valid_input_file(source_audio_file,
+                            ["mp3", "sph", "wav", "au", "raw"]):
+        LOGGER.error("Invalid audio file {}".format(source_audio_file))
+        sys.exit(1)
+
+
 def main():
     """
     Split audio file using transcript file
@@ -22,34 +45,22 @@ def main():
         description=
         "Split an audio file using valid segments from a transcript file. For this utility, transcript files must contain start/stop times."
     )
-    parser.add_argument('--target-dir',
-                        default='split',
+    parser.add_argument("--target-dir",
+                        default="split",
                         required=False,
                         help="Path to target directory")
-    parser.add_argument('audio_file',
-                        metavar='audio_file',
+    parser.add_argument("audio_file",
+                        metavar="audio_file",
                         type=str,
-                        help='input audio file')
-    parser.add_argument('transcript',
-                        metavar='transcript',
+                        help="input audio file")
+    parser.add_argument("transcript",
+                        metavar="transcript",
                         type=str,
-                        help='transcript')
+                        help="transcript")
 
     args = parser.parse_args()
 
-    if valid_input_file(args.transcript):
-        transcript = time_aligned_text(args.transcript)
-    else:
-        LOGGER.error("Invalid transcript file {}".format(args.transcript))
-        sys.exit(1)
-
-    if valid_input_file(args.audio_file, ["mp3", "sph", "wav", "au", "raw"]):
-        source_audio = audio_file(args.audio_file)
-    else:
-        LOGGER.error("Invalid audio file {}".format(args.audio_file))
-        sys.exit(1)
-
-    source_audio.split(transcript, args.target_dir)
+    split_audio_file(args.audio_file, args.transcript, args.target_dir)
 
 
 if __name__ == "__main__":

setup.py

@@ -12,7 +12,7 @@
 
 setup(
     name='asrtoolkit',
-    version='0.1.19',
+    version='0.1.20',
     description=
     'The GreenKey ASRToolkit provides tools for automatic speech recognition (ASR) file conversion and corpora organization.',
     long_description=long_description,
@@ -25,7 +25,7 @@
     keywords="asr speech recognition greenkey word error rate",
     entry_points={
         'console_scripts': [
-            'convert_transcript = asrtoolkit.convert_transcript:convert',
+            'convert_transcript = asrtoolkit.convert_transcript:main',
             'clean_formatting=asrtoolkit.clean_formatting:main',
             'prepare_audio_corpora=asrtoolkit.prepare_audio_corpora:main',
             'degrade_audio_file=asrtoolkit.degrade_audio_file:main',

tests/run_tests.sh

@@ -1,3 +1,3 @@
 #!/usr/bin/env sh
 
-python3 -m pytest -v --doctest-modules
+python3 -m pytest -vv --doctest-modules

tests/small-test-file.stm

@@ -0,0 +1,2 @@
+small-test-file 1 gk_speaker 0.765 2.881 <o,f0,male> one two three four five
+small-test-file 1 gk_speaker 5.008 6.966 <o,f0,male> six seven eight nine ten

tests/test_split_audio_file.py

@@ -0,0 +1,22 @@
+#!/usr/bin/env python
+"""
+Test audio file splitter
+"""
+import os
+from asrtoolkit.split_audio_file import split_audio_file
+
+
+def test_split_audio_file():
+    " test audio file splitter "
+    split_audio_file("tests/small-test-file.mp3", "tests/small-test-file.stm",
+                     "tests/split")
+    assert set(os.listdir("tests/split")) == {
+        'small_test_file_seg_00001.stm', 'small_test_file_seg_00000.mp3',
+        'small_test_file_seg_00001.mp3', 'small_test_file_seg_00000.stm'
+    }
+
+
+if __name__ == '__main__':
+    import sys
+    import pytest
+    pytest.main(sys.argv)