mypy config update with type fixes (#36)

Author: will-rice

.circleci/config.yml

@@ -32,7 +32,7 @@ jobs:
                   black spokestack tests
             - run: |
                   source ~/venv/bin/activate
-                  mypy spokestack tests
+                  mypy spokestack
             - run: |
                   source ~/venv/bin/activate
                   python setup.py build_ext -i

.pre-commit-config.yaml

@@ -23,6 +23,7 @@ repos:
       rev: v0.790
       hooks:
           - id: mypy
+            exclude: ^tests/
     - repo: https://github.com/asottile/blacken-docs
       rev: v1.8.0
       hooks:

mypy.ini

@@ -0,0 +1,33 @@
+[mypy]
+disallow_subclassing_any = True
+check_untyped_defs = True
+disallow_untyped_defs = True
+disallow_incomplete_defs = True
+no_implicit_optional = True
+warn_unused_ignores = True
+warn_no_return = True
+warn_unreachable = True
+[mypy-numpy.*]
+ignore_missing_imports = True
+[mypy-pynvml.*]
+ignore_missing_imports = True
+[mypy-tensorflow.*]
+ignore_missing_imports = True
+[mypy-google.*]
+ignore_missing_imports = True
+[mypy-websocket.*]
+ignore_missing_imports = True
+[mypy-pyaudio.*]
+ignore_missing_imports = True
+[mypy-tflite_runtime.*]
+ignore_missing_imports = True
+[mypy-tokenizers.*]
+ignore_missing_imports = True
+[mypy-spokestack.extensions.*]
+ignore_missing_imports = True
+[mypy-streamp3.*]
+ignore_missing_imports = True
+[mypy-pytest.*]
+ignore_missing_imports = True
+[mypy-Cython.*]
+ignore_missing_imports = True

setup.py

@@ -2,7 +2,7 @@
 
 import numpy as np
 import setuptools
-from Cython.Build import cythonize  # type: ignore
+from Cython.Build import cythonize
 
 SOURCES = [
     os.path.join("spokestack/extensions/webrtc", source)
@@ -61,7 +61,7 @@
 
 setuptools.setup(
     name="spokestack",
-    version="0.0.9",
+    version="0.0.10",
     author="Spokestack",
     author_email="support@spokestack.io",
     description="Spokestack Library for Python",
@@ -75,14 +75,16 @@
         "Operating System :: OS Independent",
     ],
     python_requires=">=3.8",
+    setup_requires=["setuptools", "cython", "numpy"],
     install_requires=[
+        "numpy",
         "pyaudio",
         "webrtcvad",
-        "numpy",
         "websocket_client",
         "tokenizers",
         "requests",
         "streamp3",
+        "cython",
     ],
     ext_modules=cythonize(EXTENSIONS),
     include_dirs=[np.get_include()],

spokestack/activation_timeout.py

@@ -1,6 +1,10 @@
 """
 This module manages the timeout for speech pipeline activation.
 """
+from typing import Any, Union
+
+import numpy as np
+
 from spokestack.context import SpeechContext
 
 
@@ -14,15 +18,21 @@ class ActivationTimeout:
     """
 
     def __init__(
-        self, frame_width=20, min_active=500, max_active=5000, **kwargs
+        self,
+        frame_width: int = 20,
+        min_active: int = 500,
+        max_active: int = 5000,
+        **kwargs: Any
     ) -> None:
 
         self._min_active = min_active / frame_width
         self._max_active = max_active / frame_width
         self._is_speech = False
         self._active_length = 0
 
-    def __call__(self, context: SpeechContext, frame=None) -> None:
+    def __call__(
+        self, context: SpeechContext, frame: Union[np.ndarray, None] = None
+    ) -> None:
         """Main entry point that manages timeout
 
         Args:

spokestack/agc/webrtc.py

@@ -1,10 +1,12 @@
 """
 This module contains the class for webrtc's automatic gain control
 """
-import numpy as np  # type: ignore
+from typing import Any
+
+import numpy as np
 
 from spokestack.context import SpeechContext
-from spokestack.extensions.webrtc.agc import WebRtcAgc  # type: ignore
+from spokestack.extensions.webrtc.agc import WebRtcAgc
 
 
 class AutomaticGainControl:
@@ -20,12 +22,12 @@ class AutomaticGainControl:
 
     def __init__(
         self,
-        sample_rate=16000,
-        frame_width=20,
-        target_level_dbfs=3,
-        compression_gain_db=15,
-        limit_enable=True,
-        **kwargs
+        sample_rate: int = 16000,
+        frame_width: int = 20,
+        target_level_dbfs: int = 3,
+        compression_gain_db: int = 15,
+        limit_enable: bool = True,
+        **kwargs: Any
     ) -> None:
         # validate sample rate
         self._sample_rate = sample_rate

spokestack/asr/google/speech_recognizer.py

@@ -4,15 +4,14 @@
 import logging
 from queue import Queue
 from threading import Thread
-from typing import Any, Union
+from typing import Any, Generator, Union
 
-import numpy as np  # type: ignore
-from google.cloud import speech  # type: ignore
-from google.oauth2 import service_account  # type: ignore
+import numpy as np
+from google.cloud import speech
+from google.oauth2 import service_account
 
 from spokestack.context import SpeechContext
 
-
 _LOG = logging.getLogger(__name__)
 
 
@@ -37,7 +36,7 @@ def __init__(
         language: str,
         credentials: Union[None, str, dict] = None,
         sample_rate: int = 16000,
-        **kwargs,
+        **kwargs: Any,
     ) -> None:
         if credentials:
             if isinstance(credentials, str):
@@ -83,14 +82,14 @@ def __call__(self, context: SpeechContext, frame: np.ndarray) -> None:
         if context.is_active:
             self._send(frame)
 
-    def _begin(self, context) -> None:
+    def _begin(self, context: SpeechContext) -> None:
         self._thread = Thread(
             target=self._receive,
             args=(context,),
         )
         self._thread.start()
 
-    def _receive(self, context):
+    def _receive(self, context: SpeechContext) -> None:
         for response in self._client.streaming_recognize(self._config, self._drain()):
             for result in response.results[:1]:
                 for alternative in result.alternatives[:1]:
@@ -107,7 +106,7 @@ def _receive(self, context):
                         context.event("timeout")
                         _LOG.debug("timeout event")
 
-    def _drain(self):
+    def _drain(self) -> Generator:
         while data := self._queue.get():
             yield data
 
@@ -116,7 +115,7 @@ def _commit(self) -> None:
         self._thread.join()
         self._thread = None
 
-    def _send(self, frame) -> None:
+    def _send(self, frame: np.ndarray) -> None:
         self._queue.put(speech.StreamingRecognizeRequest(audio_content=frame.tobytes()))
 
     def reset(self) -> None:

spokestack/asr/keyword/tflite.py

@@ -3,9 +3,9 @@
 from an audio stream.
 """
 import os
-from typing import List
+from typing import Any, List
 
-import numpy as np  # type: ignore
+import numpy as np
 
 from spokestack.context import SpeechContext
 from spokestack.models.tensorflow import TFLiteModel
@@ -34,7 +34,7 @@ def __init__(
         fft_hop_length: int = 10,
         model_dir: str = "",
         posterior_threshold: float = 0.5,
-        **kwargs
+        **kwargs: Any
     ) -> None:
 
         self.classes = classes
@@ -96,7 +96,7 @@ def __init__(
         self._prev_sample: float = 0.0
         self._is_active = False
 
-    def __call__(self, context: SpeechContext, frame) -> None:
+    def __call__(self, context: SpeechContext, frame: np.ndarray) -> None:
 
         self._sample(context, frame)
 
@@ -105,7 +105,7 @@ def __call__(self, context: SpeechContext, frame) -> None:
 
         self._is_active = context.is_active
 
-    def _sample(self, context: SpeechContext, frame) -> None:
+    def _sample(self, context: SpeechContext, frame: np.ndarray) -> None:
         # convert the PCM-16 audio to float32 in (-1.0, 1.0)
         frame = frame.astype(np.float32) / (2 ** 15 - 1)
         frame = np.clip(frame, -1.0, 1.0)
@@ -138,7 +138,7 @@ def _analyze(self, context: SpeechContext) -> None:
         # compute mel spectrogram
         self._filter(context, frame)
 
-    def _filter(self, context: SpeechContext, frame) -> None:
+    def _filter(self, context: SpeechContext, frame: np.ndarray) -> None:
         # add the batch dimension and compute the mel spectrogram with filter model
         frame = np.expand_dims(frame, 0)
         frame = self.filter_model(frame)[0]

spokestack/asr/spokestack/cloud_client.py

@@ -8,8 +8,8 @@
 import json
 from typing import Any, Dict, List, Union
 
-import numpy as np  # type: ignore
-from websocket import WebSocket  # type: ignore
+import numpy as np
+from websocket import WebSocket
 
 
 class CloudClient:
@@ -137,7 +137,7 @@ def disconnect(self) -> None:
             self._socket.close()
             self._socket = None
 
-    def send(self, frame: np.ndarray):
+    def send(self, frame: np.ndarray) -> None:
         """sends a single frame of audio
 
         Args:
@@ -149,14 +149,14 @@ def send(self, frame: np.ndarray):
         else:
             raise ConnectionError("Not Connected")
 
-    def end(self):
+    def end(self) -> None:
         """ sends empty string in binary to indicate last frame """
         if self._socket:
             self._socket.send_binary(b"")
         else:
             raise ConnectionError("Not Connected")
 
-    def receive(self):
+    def receive(self) -> None:
         """ receives the api response """
         if self._socket:
             timeout = self._socket.timeout
@@ -191,7 +191,7 @@ def idle_count(self) -> int:
         return self._idle_count
 
     @idle_count.setter
-    def idle_count(self, value: int):
+    def idle_count(self, value: int) -> None:
         """ sets the idle counter"""
         self._idle_count = value
 

spokestack/asr/spokestack/speech_recognizer.py

@@ -3,13 +3,13 @@
 the speech pipeline
 """
 import logging
+from typing import Any
 
-import numpy as np  # type: ignore
+import numpy as np
 
 from spokestack.asr.spokestack.cloud_client import CloudClient
 from spokestack.context import SpeechContext
 
-
 _LOG = logging.getLogger(__name__)
 
 
@@ -33,7 +33,7 @@ def __init__(
         sample_rate: int = 16000,
         frame_width: int = 20,
         idle_timeout: int = 5000,
-        **kwargs,
+        **kwargs: Any,
     ) -> None:
 
         self._client: CloudClient = CloudClient(
@@ -78,10 +78,10 @@ def _begin(self) -> None:
         self._is_active = True
         self._client.idle_count = 0
 
-    def _send(self, frame) -> None:
+    def _send(self, frame: np.ndarray) -> None:
         self._client.send(frame)
 
-    def _receive(self, context):
+    def _receive(self, context: SpeechContext) -> None:
         self._client.receive()
         hypotheses = self._client.response.get("hypotheses")
         if hypotheses: