New precise-collect script and minor fixes

- Fix precise-convert script
 - Fix pyinstaller spec
 - Change keyword folder to wake-word. This is more consistent with other labels

.gitignore

@@ -1,23 +1,24 @@
-dist/
-build/
-.cache/
-.idea/
+/dist/
+/build/
+/.cache/
+/.idea/
 __pycache__/
-*.egg-info/
-*.pb
-*.params
-*.net
-*.pbtxt
-*.txt
-other/
-.venv/
-stats.json
+/*.egg-info/
+/*.pb
+/*.params
+/*.net
+/*.pbtxt
+/*.txt
+/*.wav
+/other/
+/.venv/
+/stats.json
 
 !requirements.txt
 
 # jenkins branch
-wake-words.txt
-all/
-mycroft-precise/
-models/
+/wake-words.txt
+/all/
+/mycroft-precise/
+/models/
 

precise.stream.spec

@@ -2,7 +2,7 @@
 block_cipher = None
 
 
-a = Analysis(['precise/stream.py'],
+a = Analysis(['precise/scripts/stream.py'],
              pathex=['.'],
              binaries=[],
              datas=[],

precise/model.py

@@ -17,7 +17,7 @@ def load_precise_model(model_name: str) -> Any:
     return load_keras().models.load_model(model_name)
 
 
-def create_model(model_name: str, skip_acc: bool = False) -> Any:
+def create_model(model_name: str, skip_acc=False, extra_metrics=False) -> Any:
     """
     Load or create a precise model
 
@@ -42,6 +42,6 @@ def create_model(model_name: str, skip_acc: bool = False) -> Any:
         model.add(Dense(1, activation='sigmoid'))
 
     load_keras()
-    metrics = ['accuracy', false_pos, false_neg]
+    metrics = ['accuracy'] + extra_metrics * [false_pos, false_neg]
     model.compile('rmsprop', weighted_log_loss, metrics=(not skip_acc) * metrics)
     return model

precise/scripts/collect.py

@@ -0,0 +1,146 @@
+#!/usr/bin/env python3
+# Copyright (c) 2017 Mycroft AI Inc.
+import tty
+import wave
+from os.path import isfile
+from select import select
+from sys import stdin
+from termios import tcsetattr, tcgetattr, TCSADRAIN
+from prettyparse import create_parser
+
+import pyaudio
+
+usage = '''
+    Record audio samples for use with precise
+    
+    :-w --width int 2
+        Sample width of audio
+    
+    :-r --rate int 16000
+        Sample rate of audio
+    
+    :-c --channels int 2
+        Number of audio channels
+'''
+
+
+def key_pressed():
+    return select([stdin], [], [], 0) == ([stdin], [], [])
+
+
+def termios_wrapper(main):
+    global orig_settings
+    orig_settings = tcgetattr(stdin)
+    try:
+        hide_input()
+        main()
+    finally:
+        tcsetattr(stdin, TCSADRAIN, orig_settings)
+
+
+def show_input():
+    tcsetattr(stdin, TCSADRAIN, orig_settings)
+
+
+def hide_input():
+    tty.setcbreak(stdin.fileno())
+
+
+orig_settings = None
+
+RECORD_KEY = ' '
+EXIT_KEY_CODE = 27
+
+
+def record_until(p, should_return, args):
+    chunk_size = 1024
+    stream = p.open(format=p.get_format_from_width(args.width), channels=args.channels, rate=args.rate,
+                    input=True, frames_per_buffer=chunk_size)
+
+    frames = []
+    while not should_return():
+        frames.append(stream.read(chunk_size))
+
+    stream.stop_stream()
+    stream.close()
+
+    return b''.join(frames)
+
+
+def save_audio(name, data, args):
+    wf = wave.open(name, 'wb')
+    wf.setnchannels(args.channels)
+    wf.setsampwidth(args.width)
+    wf.setframerate(args.rate)
+    wf.writeframes(data)
+    wf.close()
+
+
+def next_name(name):
+    name += '.wav'
+    pos, num_digits = None, None
+    try:
+        pos = name.index('#')
+        num_digits = name.count('#')
+    except ValueError:
+        print("Name must contain at least one # to indicate where to put the number.")
+        raise
+
+    def get_name(i):
+        nonlocal name, pos
+        return name[:pos] + str(i).zfill(num_digits) + name[pos + num_digits:]
+
+    i = 0
+    while True:
+        if not isfile(get_name(i)):
+            break
+        i += 1
+
+    return get_name(i)
+
+
+def wait_to_continue():
+    while True:
+        c = stdin.read(1)
+        if c == RECORD_KEY:
+            return True
+        elif ord(c) == EXIT_KEY_CODE:
+            return False
+
+
+def record_until_key(p, args):
+    def should_return():
+        return key_pressed() and stdin.read(1) == RECORD_KEY
+
+    return record_until(p, should_return, args)
+
+
+def _main():
+    args = create_parser(usage).parse_args()
+    show_input()
+    audio_name = input("Audio name (Ex. recording-##): ")
+    hide_input()
+
+    p = pyaudio.PyAudio()
+
+    while True:
+        print('Press space to record (esc to exit)...')
+
+        if not wait_to_continue():
+            break
+
+        print('Recording...')
+        d = record_until_key(p, args)
+        name = next_name(audio_name)
+        save_audio(name, d, args)
+        print('Saved as ' + name)
+
+    p.terminate()
+
+
+def main():
+    termios_wrapper(_main)
+
+
+if __name__ == '__main__':
+    main()

precise/scripts/convert.py

@@ -8,7 +8,7 @@ from shutil import copyfile
 from prettyparse import create_parser
 
 usage = '''
-    Convert keyword model from Keras to TensorFlow
+    Convert wake word model from Keras to TensorFlow
     
     :model str
         Input Keras model (.net)
@@ -65,8 +65,12 @@ def convert(model_path: str, out_file: str):
     del sess
 
 
-if __name__ == '__main__':
+def main():
     args = create_parser(usage).parse_args()
 
     model_name = args.model.replace('.net', '')
     convert(args.model, args.out.format(model=model_name))
+
+
+if __name__ == '__main__':
+    main()

precise/scripts/train.py

@@ -25,6 +25,9 @@ usage = '''
     :-mm --metric-monitor str loss
         Metric used to determine when to save
     
+    :-em --extra-metrics
+        Add extra metrics during training
+    
     ...
 '''
 
@@ -50,7 +53,7 @@ def main():
         print('Not enough data to train')
         exit(1)
 
-    model = create_model(args.model, args.no_validation)
+    model = create_model(args.model, args.no_validation, args.extra_metrics)
     model.summary()
 
     from keras.callbacks import ModelCheckpoint

precise/scripts/train_incremental.py

@@ -40,6 +40,9 @@ usage = '''
     :-mm --metric-monitor str loss
         Metric used to determine when to save
     
+    :-em --extra-metrics
+        Add extra metrics during training
+    
     :-nv --no-validation
         Disable accuracy and validation calculation
         to improve speed during training
@@ -85,7 +88,7 @@ class IncrementalTrainer:
         self.db_data = data.load(True, not args.no_validation)
 
         if not isfile(args.model):
-            create_model(args.model, args.no_validation).save(args.model)
+            create_model(args.model, args.no_validation, args.extra_metrics).save(args.model)
         self.listener = Listener(args.model, args.chunk_size, runner_cls=KerasRunner)
 
     def retrain(self):
@@ -118,7 +121,7 @@ class IncrementalTrainer:
             if conf > 0.5:
                 samples_since_train += 1
                 name = splitext(basename(fn))[0] + '-' + str(i) + '.wav'
-                name = join(self.args.data_dir, 'test' if save_test else '', 'not-keyword',
+                name = join(self.args.data_dir, 'test' if save_test else '', 'not-wake-word',
                             'generated', name)
                 save_audio(name, audio_buffer)
                 print()
@@ -152,8 +155,8 @@ def main():
     args = TrainData.parse_args(create_parser(usage))
 
     for i in (
-            join(args.data_dir, 'not-keyword', 'generated'),
-            join(args.data_dir, 'test', 'not-keyword', 'generated')
+            join(args.data_dir, 'not-wake-word', 'generated'),
+            join(args.data_dir, 'test', 'not-wake-word', 'generated')
     ):
         makedirs(i, exist_ok=True)
 

precise/train_data.py

@@ -23,14 +23,14 @@ class TrainData:
         """
         Load a set of data from a structured folder in the following format:
         {prefix}/
-            keyword/
+            wake-word/
                 *.wav
-            not-keyword/
+            not-wake-word/
                 *.wav
             test/
-                keyword/
+                wake-word/
                     *.wav
-                not-keyword/
+                not-wake-word/
                     *.wav
         """
         return cls(find_wavs(folder), find_wavs(join(folder, 'test')))
@@ -99,7 +99,7 @@ class TrainData:
         return self.__load(self.__load_files, train, test)
 
     def load_inhibit(self, train=True, test=True) -> tuple:
-        """Generate data with inhibitory inputs created from keyword samples"""
+        """Generate data with inhibitory inputs created from wake word samples"""
 
         def loader(kws: list, nkws: list):
             from precise.params import pr
@@ -162,10 +162,10 @@ class TrainData:
             inputs.extend(load_vector(f, vectorizer) for f in filenames)
             outputs.extend(np.array([output]) for _ in filenames)
 
-        print('Loading keyword...')
+        print('Loading wake-word...')
         add(kw_files, 1.0)
 
-        print('Loading not-keyword...')
+        print('Loading not-wake-word...')
         add(nkw_files, 0.0)
 
         from precise.params import pr

precise/util.py

@@ -52,5 +52,5 @@ def glob_all(folder: str, filt: str) -> List[str]:
 
 
 def find_wavs(folder: str) -> Tuple[List[str], List[str]]:
-    """Finds keyword and not-keyword wavs in folder"""
-    return glob_all(folder + '/keyword', '*.wav'), glob_all(folder + '/not-keyword', '*.wav')
+    """Finds wake-word and not-wake-word wavs in folder"""
+    return glob_all(folder + '/wake-word', '*.wav'), glob_all(folder + '/not-wake-word', '*.wav')

precise/vectorization.py

@@ -40,7 +40,7 @@ def vectorize(audio: np.ndarray) -> np.ndarray:
 def vectorize_inhibit(audio: np.ndarray) -> np.ndarray:
     """
     Returns an array of inputs generated from the
-    keyword audio that shouldn't cause an activation
+    wake word audio that shouldn't cause an activation
     """
 
     def samp(x):

setup.py

@@ -1,15 +1,21 @@
 #!/usr/bin/env python3
 
-from setuptools import setup, find_packages
+from setuptools import setup
 
 from precise import __version__
 
 setup(
     name='mycroft-precise',
     version=__version__,
-    packages=find_packages(),
+    packages=[
+        'precise',
+        'precise.scripts',
+        'precise.pocketsphinx',
+        'precise.pocketsphinx.scripts'
+    ],
     entry_points={
         'console_scripts': [
+            'precise-collect=precise.scripts.collect:main',
             'precise-convert=precise.scripts.convert:main',
             'precise-eval=precise.scripts.eval:main',
             'precise-record=precise.scripts.record:main',