Merge pull request 'voice recognition with whisper' (#40) from YasinOnm08/ai-virtual-assistant:main into main

Reviewed-on: https://interstellardevelopment.org/code/code/React-Group/ai-virtual-assistant/pulls/40

py/voice_recognition.py

@@ -1,31 +1,61 @@
-import speech_recognition as sr
+#pip install faster-whisper
+from api import API
+import os
+import wave
+from faster_whisper import WhisperModel
+import pyaudio
 
-class Voice: #create Class
-    @staticmethod
-    def listen():   #define function listen()
-        recognizer = sr.Recognizer()
-
-        try:
-            with sr.Microphone() as source:
-                print("Adjusting for ambient noise...")
-                recognizer.adjust_for_ambient_noise(source, duration=0.5) #listen to surrounding for .5sec to adjust backgroundnoise
-                print("Listening...")
-                audio_data = recognizer.listen(source) #listen to user until user stops speaking
-                print("Audio captured")
-                try:
-                    text = recognizer.recognize_sphinx(audio_data)  # Using Sphinx convert audio to text (also works offline)
-                #if any Exceptions or Errors eccur => return ERROR
-                except sr.UnknownValueError:
-                    text = "ERROR"
-                except sr.RequestError as e:
-                    text = "ERROR"
-                
-        except sr.RequestError as e:
-            text = "ERROR"
-        except sr.UnknownValueError:
-            text = "ERROR"
-        except Exception as e:
-            text = "ERROR"
-        return text
+def transcribe_chunk(model, file_path):
+    print("transcribing...")
+    segments, _ = model.transcribe(file_path)
+    transcription = ""
+    for segment in segments:
+        transcription += segment.text + " "
+    return transcription.strip()
 
+#Record voice
+#chunk_length = duration to record in seconds
+def record_chunk(p, stream, file_path, chunk_length=2):
+    frames=[]
+    for _ in range(0, int(16000/1024 * chunk_length)):
+        data = stream.read(1024)
+        frames.append(data)
 
+    wf = wave.open(file_path, 'wb')
+    wf.setnchannels(1)
+    wf.setsampwidth(p.get_sample_size(pyaudio.paInt16))
+    wf.setframerate(16000)
+    wf.writeframes(b''.join(frames))
+    wf.close()
+    
+def listen():
+    #model settings (tiny, base, small, medium, large)
+    model_size = "medium"
+    #what should it run on (cpu or cuda for gpu)
+    model = WhisperModel(model_size, device="cpu", compute_type="int8")
+    
+    p = pyaudio.PyAudio()
+    stream = p.open(format=pyaudio.paInt16, channels=1, rate = 16000, input = True,  frames_per_buffer=1024)
+    
+    try:
+        while True:
+            print("Recording...")
+            # CHANGE TEMP FILE PATH
+            chunk_file="temp_chunk.wav"
+            record_chunk(p, stream, chunk_file)
+            transcription = transcribe_chunk(model, chunk_file)
+            print(transcription)
+            try:
+                return transcription
+            except Exception as e:
+                return "ERROR"
+            finally:
+                if os.path.exists(chunk_file):
+                    os.remove(chunk_file)
+            break
+    except KeyboardInterrupt:
+        print("Stopping...")
+    finally:
+        stream.stop_stream()
+        stream.close()
+        p.terminate()