Merge pull request #1 from s0d3s/main

AudioRecorder.py

@@ -1,10 +1,14 @@
 import custom_speech_recognition as sr
 import os
 from datetime import datetime
+
 try:
     import pyaudiowpatch as pyaudio
 except ImportError:
-    import pyaudio
+    if os.name != "nt":
+        import pyaudio
+    else:
+        raise
 
 RECORD_TIMEOUT = 3
 ENERGY_THRESHOLD = 1000
@@ -41,45 +45,44 @@ def __init__(self):
         self.adjust_for_noise("Default Mic", "Please make some noise from the Default Mic...")
 
 class DefaultSpeakerRecorder(BaseRecorder):
-    def __init__(self):
-        #with pyaudio.PyAudio() as p:
-        p = pyaudio.PyAudio() 
-        try:
-            if os.name == "nt":
-                wasapi_info = p.get_host_api_info_by_type(pyaudio.paWASAPI)
-                default_speakers = p.get_device_info_by_index(wasapi_info["defaultOutputDevice"])
-            else:
-                codeaudio_info = p.get_host_api_info_by_type(pyaudio.paCoreAudio)
-                default_speakers = p.get_device_info_by_index(codeaudio_info["defaultOutputDevice"])
+
+    # Different implementations of obtaining the info dict of a default speaker, for different platforms
+    if os.name == "nt":
+        def _get_default_speaker(self):
+            # Requires PyAudioWPatch >= 0.2.12.6
+            with pyaudio.PyAudio() as p:
+                try:
+                    # Get loopback of default WASAPI speaker
+                    return p.get_default_wasapi_loopback()
+
+                except OSError:                
+                    print("[ERROR] Looks like WASAPI is not available on the system.")
+
+                except LookupError:
+                    print("[ERROR] No loopback device found.")
+
+    else:
+        def _get_default_speaker(self):
+            # At the moment, recording from speakers is only available under Windows
+            # raise NotImplementedError("Recording from speakers is only available under Windows")
+
+            # As far as I understand, now the code style does not provide
+            # for error handling - only print them.
+            print("[ERROR] Recording from speakers is only available under Windows.")
 
-            isLoopbackDevice = False
-            if os.name == "nt":
-                isLoopbackDevice = default_speakers["isLoopbackDevice"]
-            else:
-                isLoopbackDevice = default_speakers.get("isLoopbackDevice", False)
-            if not isLoopbackDevice:
-                if os.name != "nt":
-                    for i in range(p.get_device_count()):
-                        device_info = p.get_device_info_by_index(i)
-                        if device_info['maxInputChannels'] > 0 and device_info['hostApi'] == p.get_default_host_api_info()['index']:
-                            default_speakers = loopback = device_info
-                            break
-                        else:
-                            print("[ERROR] No loopback device found.")
-
-                else:
-                    for loopback in p.get_loopback_device_info_generator():
-                        if default_speakers["name"] in loopback["name"]:
-                            default_speakers = loopback
-                            break
-                    else:
-                        print("[ERROR] No loopback device found.")
-        finally:
-            p.terminate()
+
+    def __init__(self):
+        default_speaker = self._get_default_speaker()
+
+        if not default_speaker:
+            print("[ERROR] Something went wrong while trying to get the default speakers.")
+            super().__init__(source=sr.Microphone(sample_rate=16000), source_name="Speaker")
+            return
+
         source = sr.Microphone(speaker=True,
-                               device_index= default_speakers["index"],
-                               sample_rate=int(default_speakers["defaultSampleRate"]),
+                               device_index=default_speaker["index"],
+                               sample_rate=int(default_speaker["defaultSampleRate"]),
                                chunk_size=pyaudio.get_sample_size(pyaudio.paInt16),
-                               channels=default_speakers["maxInputChannels"])
+                               channels=default_speaker["maxInputChannels"])
         super().__init__(source=source, source_name="Speaker")
         self.adjust_for_noise("Default Speaker", "Please make or play some noise from the Default Speaker...")

AudioTranscriber.py

@@ -8,10 +8,14 @@
 import io
 from datetime import timedelta
 from heapq import merge
+
 try:
-    import pyaudiowpatch as pyaudio 
+    import pyaudiowpatch as pyaudio
 except ImportError:
-    import pyaudio
+    if os.name != "nt":
+        import pyaudio
+    else:
+        raise
 
 
 PHRASE_TIMEOUT = 3.05

custom_speech_recognition/__init__.py

@@ -107,9 +107,12 @@ def get_pyaudio():
         Imports the pyaudio module and checks its version. Throws exceptions if pyaudio can't be found or a wrong version is installed
         """
         try:
-            import pyaudiowpatch as pyaudio 
+            import pyaudiowpatch as pyaudio
         except ImportError:
-            import pyaudio
+            if os.name != "nt":
+                import pyaudio
+            else:
+                raise
         from distutils.version import LooseVersion
         if LooseVersion(pyaudio.__version__) < LooseVersion("0.2.11"):
             raise AttributeError("PyAudio 0.2.11 or later is required (found version {})".format(pyaudio.__version__))

requirements.txt

@@ -3,7 +3,7 @@ openai-whisper==20230314
 Wave==0.0.2
 openai==0.27.6
 customtkinter==5.1.3
-PyAudioWPatch==0.2.12.5
+PyAudioWPatch==0.2.12.6; sys_platform == 'win32'
+pyaudio==0.2.13; sys_platform != 'win32'
 --extra-index-url https://download.pytorch.org/whl/cu117
 torch
-pyaudio==0.2.13