support for snowboy; fixes alexa-pi#53

.travis.yml

@@ -18,11 +18,15 @@ sudo: required
 # Some discussion: https://github.com/koalaman/shellcheck/issues/785
 before_install:
   - sudo apt-get update
-  - sudo apt-get install -y wget git swig portaudio19-dev libpulse-dev vlc-nox shellcheck
+  - sudo apt-get install -y wget git portaudio19-dev libpulse-dev vlc-nox shellcheck sox libatlas-base-dev
+  - sudo wget http://mirrors.kernel.org/ubuntu/pool/universe/s/swig/swig3.0_3.0.10-1.1_amd64.deb
+  - sudo dpkg -i swig3.0_3.0.10-1.1_amd64.deb
+  - sudo ln -s /usr/bin/swig3.0 /usr/local/bin/swig
 
 install:
   - pip install -r src/requirements.txt
   - pip install -r src/dev-requirements.txt
+  - pip install git+https://github.com/Kitt-AI/snowboy.git || true
   - yes | pip install git+https://github.com/duxingkei33/orangepi_PC_gpio_pyH3.git -v
 
 script:

CHANGELOG.md

@@ -5,6 +5,9 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/).
 
 ## [Unreleased]
 
+### Added
+- Support for snowboy - the awesome trigger word detector
+
 ## [1.6] - 2017-10-08
 
 ### Added

pylintrc

@@ -158,7 +158,7 @@ notes=XXX
 [SIMILARITIES]
 
 # Minimum lines number of a similarity.
-min-similarity-lines=5
+min-similarity-lines=6
 
 # Ignore comments when computing similarities.
 ignore-comments=yes

src/alexapi/triggers/snowboytrigger.py

@@ -0,0 +1,154 @@
+import os
+import logging
+import time
+import collections
+import site
+
+import pyaudio
+from snowboy import snowboydetect  # pylint: disable=import-error
+
+from .voicetrigger import VoiceTrigger
+from ..capture import DeviceInfo
+from ..exceptions import ConfigurationException
+
+SNOWBOY_FOLDER = ''
+try:
+	for packages_path in site.getsitepackages():  # pylint: disable=no-member
+		path_candidate = os.path.join(packages_path, 'snowboy')
+		if os.path.exists(path_candidate):
+			SNOWBOY_FOLDER = path_candidate
+except AttributeError:
+	SNOWBOY_FOLDER = os.path.dirname(snowboydetect.__file__)
+
+logger = logging.getLogger(__name__)
+
+
+# Copied from snowboy
+class RingBuffer(object):
+	"""Ring buffer to hold audio from PortAudio"""
+	def __init__(self, size=4096):
+		self._buf = collections.deque(maxlen=size)
+
+	def extend(self, data):
+		"""Adds data to the end of buffer"""
+		self._buf.extend(data)
+
+	def get(self):
+		"""Retrieves data from the beginning of buffer and clears it"""
+		tmp = bytes(bytearray(self._buf))
+		self._buf.clear()
+		return tmp
+
+
+class SnowboyTrigger(VoiceTrigger):
+
+	name = 'snowboy'
+
+	_sleep_time = 0.03
+	_ring_buffer = None
+
+	_pa = None
+	_device_info = None
+
+	def __init__(self, config, trigger_callback, capture):  # pylint: disable=unused-argument
+		super(SnowboyTrigger, self).__init__(config, trigger_callback)
+
+		self._model = self._tconfig['model'].replace('{distribution}', os.path.join(SNOWBOY_FOLDER, 'resources'))
+		self._sensitivity = self._tconfig['sensitivity']
+
+		self._device_info = DeviceInfo()
+
+	def validate_config(self):
+
+		model_path = self._tconfig['model'].replace('{distribution}', os.path.join(SNOWBOY_FOLDER, 'resources'))
+		if not os.path.isfile(model_path):
+			raise ConfigurationException("Invalid snowboy model path: '" + model_path + "'")
+
+	def setup(self):
+		# """
+		# :param decoder_model: decoder model file path, a string or a list of strings
+		# :param resource: resource file path.
+		# :param sensitivity: decoder sensitivity, a float of a list of floats.
+		# 						  The bigger the value, the more senstive the
+		# 						  decoder. If an empty list is provided, then the
+		# 						  default sensitivity in the model will be used.
+		# :param audio_gain: multiply input volume by this factor.
+		# """
+
+		audio_gain = 1
+
+		tm = type(self._model)
+		ts = type(self._sensitivity)
+		if tm is not list:
+			self._model = [self._model]
+		if ts is not list:
+			self._sensitivity = [self._sensitivity]
+		model_str = ",".join(self._model)
+
+		resource_filename = os.path.join(SNOWBOY_FOLDER, "resources/common.res")
+		self._detector = snowboydetect.SnowboyDetect(resource_filename.encode(), model_str.encode())
+		self._detector.SetAudioGain(audio_gain)
+		num_hotwords = self._detector.NumHotwords()
+
+		if len(self._model) > 1 and len(self._sensitivity) == 1:
+			self._sensitivity = self._sensitivity * num_hotwords
+
+		if self._sensitivity:
+			assert num_hotwords == len(self._sensitivity), \
+				"number of hotwords in self._model (%d) and sensitivity " \
+				"(%d) does not match" % (num_hotwords, len(self._sensitivity))
+
+			sensitivity_str = ",".join([str(t) for t in self._sensitivity])
+			self._detector.SetSensitivity(sensitivity_str.encode())
+
+		self._ring_buffer = RingBuffer(self._detector.NumChannels() * self._detector.SampleRate() * 5)
+
+		self._pa = pyaudio.PyAudio()
+
+	def _audio_callback(self, in_data, frame_count, time_info, status):  # pylint: disable=unused-argument
+		self._ring_buffer.extend(in_data)
+		play_data = chr(0) * len(in_data)
+
+		return play_data, pyaudio.paContinue
+
+	def thread(self):
+
+		while True:
+			self._enabled_lock.wait()
+
+			stream_in = self._pa.open(
+				input=True,
+				input_device_index=self._device_info.get_device_index(self._config['sound']['input_device']),
+				format=self._pa.get_format_from_width(self._detector.BitsPerSample() / 8),
+				channels=self._detector.NumChannels(),
+				rate=self._detector.SampleRate(),
+				frames_per_buffer=2048,
+				stream_callback=self._audio_callback)
+
+			triggered = False
+			while not triggered:
+				if not self._enabled_lock.isSet():
+					break
+
+				data = self._ring_buffer.get()
+				if not data:
+					time.sleep(self._sleep_time)
+					continue
+
+				ans = self._detector.RunDetection(data)
+				if ans == -1:
+					logger.warning("Error initializing streams or reading audio data")
+				elif ans > 0:
+					triggered = True
+
+			stream_in.stop_stream()
+			stream_in.close()
+
+			self._disabled_sync_lock.set()
+
+			if triggered:
+				self._trigger_callback(self)
+
+	def cleanup(self):
+		self.disable()
+		self._pa.terminate()

src/config.template.yaml

@@ -76,6 +76,21 @@ triggers:
     phrase: "alexa"
     threshold: 1e-10
 
+  snowboy:
+    enabled: false
+    voice_confirm: true
+
+    # Use your own model or a file from the default snowboy distribution:
+    #
+    # {distribution}/alexa.umdl
+    # {distribution}/snowboy.umdl
+    # {distribution}/alexa/alexa_02092017.umdl
+    # {distribution}/alexa/alexa-avs-sample-app/alexa.umdl
+
+    model: "{distribution}/alexa/alexa_02092017.umdl"
+    sensitivity: 0.5
+
+
 # Commands to run before and after an interaction begins. Leave empty to disable.
 event_commands: