Move script help strings to docstrings

Author: MatthewScholefield

precise/pocketsphinx/scripts/listen.py

@@ -12,6 +12,24 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+"""
+Run Pocketsphinx on microphone audio input
+
+:key_phrase str
+    Key phrase composed of words from dictionary
+
+:dict_file str
+    Filename of dictionary with word pronunciations
+
+:hmm_folder str
+    Folder containing hidden markov model
+
+:-th --threshold str 1e-90
+    Threshold for activations
+
+:-c --chunk-size int 2048
+    Samples between inferences
+"""
 from precise_runner import PreciseRunner
 from precise_runner.runner import ListenerEngine
 from prettyparse import Usage
@@ -23,24 +41,7 @@
 
 
 class PocketsphinxListenScript(BaseScript):
-    usage = Usage('''
-        Run Pocketsphinx on microphone audio input
-
-        :key_phrase str
-            Key phrase composed of words from dictionary
-
-        :dict_file str
-            Filename of dictionary with word pronunciations
-
-        :hmm_folder str
-            Folder containing hidden markov model
-
-        :-th --threshold str 1e-90
-            Threshold for activations
-
-        :-c --chunk-size int 2048
-            Samples between inferences
-    ''')
+    usage = Usage(__doc__)
 
     def run(self):
         def on_activation():

precise/pocketsphinx/scripts/test.py

@@ -12,6 +12,29 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+"""
+Test a dataset using Pocketsphinx
+
+:key_phrase str
+    Key phrase composed of words from dictionary
+
+:dict_file str
+    Filename of dictionary with word pronunciations
+
+:hmm_folder str
+    Folder containing hidden markov model
+
+:-th --threshold str 1e-90
+    Threshold for activations
+
+:-t --use-train
+    Evaluate training data instead of test data
+
+:-nf --no-filenames
+    Don't show the names of files that failed
+
+...
+"""
 import wave
 from prettyparse import Usage
 from subprocess import check_output, PIPE
@@ -23,29 +46,7 @@
 
 
 class PocketsphinxTestScript(BaseScript):
-    usage = Usage('''
-        Test a dataset using Pocketsphinx
-
-        :key_phrase str
-            Key phrase composed of words from dictionary
-
-        :dict_file str
-            Filename of dictionary with word pronunciations
-
-        :hmm_folder str
-            Folder containing hidden markov model
-
-        :-th --threshold str 1e-90
-            Threshold for activations
-
-        :-t --use-train
-            Evaluate training data instead of test data
-
-        :-nf --no-filenames
-            Don't show the names of files that failed
-
-        ...
-    ''') | TrainData.usage
+    usage = Usage(__doc__) | TrainData.usage
 
     def __init__(self, args):
         super().__init__(args)

precise/scripts/add_noise.py

@@ -12,6 +12,30 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+"""
+Create a duplicate dataset with added noise
+
+:folder str
+    Folder containing source dataset
+
+:-tg --tags-file str -
+    Tags file to optionally load from
+
+:noise_folder str
+    Folder with wav files containing noise to be added
+
+:output_folder str
+    Folder to write the duplicate generated dataset
+
+:-if --inflation-factor int 1
+    The number of noisy samples generated per single source sample
+
+:-nl --noise-ratio-low float 0.0
+    Minimum random ratio of noise to sample. 1.0 is all noise, no sample sound
+
+:-nh --noise-ratio-high float 0.4
+    Maximum random ratio of noise to sample. 1.0 is all noise, no sample sound
+"""
 from math import sqrt
 
 import numpy as np
@@ -68,30 +92,7 @@ def noised_audio(self, audio: np.ndarray, noise_ratio: float) -> np.ndarray:
 
 class AddNoiseScript(BaseScript):
     usage = Usage(
-        """
-            Create a duplicate dataset with added noise
-
-            :folder str
-                Folder containing source dataset
-
-            :-tg --tags-file str -
-                Tags file to optionally load from
-
-            :noise_folder str
-                Folder with wav files containing noise to be added
-
-            :output_folder str
-                Folder to write the duplicate generated dataset
-
-            :-if --inflation-factor int 1
-                The number of noisy samples generated per single source sample
-
-            :-nl --noise-ratio-low float 0.0
-                Minimum random ratio of noise to sample. 1.0 is all noise, no sample sound
-
-            :-nh --noise-ratio-high float 0.4
-                Maximum random ratio of noise to sample. 1.0 is all noise, no sample sound
-        """,
+        __doc__,
         tags_file=lambda args: abspath(args.tags_file) if args.tags_file else None,
         folder=lambda args: abspath(args.folder),
         output_folder=lambda args: abspath(args.output_folder)

precise/scripts/calc_threshold.py

@@ -12,6 +12,26 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License
+"""
+Update the threshold values of a model for a dataset.
+This makes the sensitivity more accurate and linear
+
+:model str
+    Either Keras (.net) or TensorFlow (.pb) model to adjust
+
+:input_file str
+    Input stats file that was outputted from precise-graph
+
+:-k --model-key str -
+    Custom model name to use from the stats.json
+
+:-s --smoothing float 1.2
+    Amount of extra smoothing to apply
+
+:-c --center float 0.2
+    Decoded threshold that is mapped to 0.5. Proportion of
+    false negatives at sensitivity=0.5
+"""
 from math import sqrt
 
 from os.path import basename, splitext
@@ -23,26 +43,7 @@
 
 
 class CalcThresholdScript(BaseScript):
-    usage = Usage('''
-        Update the threshold values of a model for a dataset.
-        This makes the sensitivity more accurate and linear
-
-        :model str
-            Either Keras (.net) or TensorFlow (.pb) model to adjust
-
-        :input_file str
-            Input stats file that was outputted from precise-graph
-
-        :-k --model-key str -
-            Custom model name to use from the stats.json
-
-        :-s --smoothing float 1.2
-            Amount of extra smoothing to apply
-
-        :-c --center float 0.2
-            Decoded threshold that is mapped to 0.5. Proportion of
-            false negatives at sensitivity=0.5
-    ''')
+    usage = Usage(__doc__)
 
     def __init__(self, args):
         super().__init__(args)

precise/scripts/collect.py

@@ -12,6 +12,18 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+"""
+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 1
+    Number of audio channels
+"""
 from select import select
 from sys import stdin
 from termios import tcsetattr, tcgetattr, TCSADRAIN
@@ -53,18 +65,7 @@ class CollectScript(BaseScript):
     RECORD_KEY = ' '
     EXIT_KEY_CODE = 27
 
-    usage = 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 1
-            Number of audio channels
-    ''')
+    usage = Usage(__doc__)
     usage.add_argument('file_label', nargs='?', help='File label (Ex. recording-##)')
 
     def __init__(self, args):

precise/scripts/convert.py

@@ -13,6 +13,15 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+"""
+Convert wake word model from Keras to TensorFlow
+
+:model str
+    Input Keras model (.net)
+
+:-o --out str {model}.pb
+    Custom output TensorFlow protobuf filename
+"""
 import os
 from os.path import split, isfile
 from prettyparse import Usage
@@ -22,15 +31,7 @@
 
 
 class ConvertScript(BaseScript):
-    usage = Usage('''
-        Convert wake word model from Keras to TensorFlow
-
-        :model str
-            Input Keras model (.net)
-
-        :-o --out str {model}.pb
-            Custom output TensorFlow protobuf filename
-    ''')
+    usage = Usage(__doc__)
 
     def run(self):
         args = self.args

precise/scripts/engine.py

@@ -12,6 +12,17 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
+"""
+stdin should be a stream of raw int16 audio, written in
+groups of CHUNK_SIZE samples. If no CHUNK_SIZE is given
+it will read until EOF. For every chunk, an inference
+will be given via stdout as a float string, one per line
+
+:model_name str
+    Keras or TensorFlow model to read from
+
+...
+"""
 import sys
 
 import os
@@ -27,17 +38,7 @@ def add_audio_pipe_to_parser(parser):
 
 
 class EngineScript(BaseScript):
-    usage = Usage('''
-        stdin should be a stream of raw int16 audio, written in
-        groups of CHUNK_SIZE samples. If no CHUNK_SIZE is given
-        it will read until EOF. For every chunk, an inference
-        will be given via stdout as a float string, one per line
-
-        :model_name str
-            Keras or TensorFlow model to read from
-
-        ...
-    ''')
+    usage = Usage(__doc__)
     usage.add_argument('-v', '--version', action='version', version=__version__)
     usage.add_argument('chunk_size', type=int, nargs='?', default=-1,
                        help='Number of bytes to read before making a prediction. '