RealTimeSTT 0.2.2__tar.gz → 0.2.4__tar.gz

{RealtimeSTT-0.2.2 → RealtimeSTT-0.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: RealtimeSTT
-Version: 0.2.2
+Version: 0.2.4
 Summary: A fast Voice Activity Detection and Transcription System
 Home-page: https://github.com/KoljaB/RealTimeSTT
 Author: Kolja Beigel
@@ -12,6 +12,16 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.6
 Description-Content-Type: text/markdown
 License-File: LICENSE
+Requires-Dist: PyAudio==0.2.14
+Requires-Dist: faster-whisper==1.0.3
+Requires-Dist: pvporcupine==1.9.5
+Requires-Dist: webrtcvad==2.0.10
+Requires-Dist: halo==0.0.31
+Requires-Dist: torch==2.3.1
+Requires-Dist: torchaudio==2.3.1
+Requires-Dist: scipy==1.12.0
+Requires-Dist: websockets==v12.0
+Requires-Dist: openwakeword==0.6.0
 
 
 # RealtimeSTT
@@ -331,6 +341,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **enable_realtime_transcription** (bool, default=False): Enables or disables real-time transcription of audio. When set to True, the audio will be transcribed continuously as it is being recorded.
 
+- **use_main_model_for_realtime** (bool, default=False): If set to True, the main transcription model will be used for both regular and real-time transcription. If False, a separate model specified by `realtime_model_type` will be used for real-time transcription. Using a single model can save memory and potentially improve performance, but may not be optimized for real-time processing. Using separate models allows for a smaller, faster model for real-time transcription while keeping a more accurate model for final transcription.
+
 - **realtime_model_type** (str, default="tiny"): Specifies the size or path of the machine learning model to be used for real-time transcription.
     - Valid options: 'tiny', 'tiny.en', 'base', 'base.en', 'small', 'small.en', 'medium', 'medium.en', 'large-v1', 'large-v2'.
 

{RealtimeSTT-0.2.2 → RealtimeSTT-0.2.4}/README.md

@@ -316,6 +316,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **enable_realtime_transcription** (bool, default=False): Enables or disables real-time transcription of audio. When set to True, the audio will be transcribed continuously as it is being recorded.
 
+- **use_main_model_for_realtime** (bool, default=False): If set to True, the main transcription model will be used for both regular and real-time transcription. If False, a separate model specified by `realtime_model_type` will be used for real-time transcription. Using a single model can save memory and potentially improve performance, but may not be optimized for real-time processing. Using separate models allows for a smaller, faster model for real-time transcription while keeping a more accurate model for final transcription.
+
 - **realtime_model_type** (str, default="tiny"): Specifies the size or path of the machine learning model to be used for real-time transcription.
     - Valid options: 'tiny', 'tiny.en', 'base', 'base.en', 'small', 'small.en', 'medium', 'medium.en', 'large-v1', 'large-v2'.
 

{RealtimeSTT-0.2.2 → RealtimeSTT-0.2.4}/RealtimeSTT/audio_recorder.py

@@ -107,6 +107,7 @@ class AudioToTextRecorder:
 
                  # Realtime transcription parameters
                  enable_realtime_transcription=False,
+                 use_main_model_for_realtime=False,
                  realtime_model_type=INIT_MODEL_TRANSCRIPTION_REALTIME,
                  realtime_processing_pause=INIT_REALTIME_PROCESSING_PAUSE,
                  on_realtime_transcription_update=None,
@@ -206,6 +207,15 @@ class AudioToTextRecorder:
         - enable_realtime_transcription (bool, default=False): Enables or
             disables real-time transcription of audio. When set to True, the
             audio will be transcribed continuously as it is being recorded.
+        - use_main_model_for_realtime (str, default=False):
+            If True, use the main transcription model for both regular and
+            real-time transcription. If False, use a separate model specified
+            by realtime_model_type for real-time transcription.
+            Using a single model can save memory and potentially improve
+            performance, but may not be optimized for real-time processing.
+            Using separate models allows for a smaller, faster model for
+            real-time transcription while keeping a more accurate model for
+            final transcription.
         - realtime_model_type (str, default="tiny"): Specifies the machine
             learning model to be used for real-time transcription. Valid
             options include 'tiny', 'tiny.en', 'base', 'base.en', 'small',
@@ -229,7 +239,7 @@ class AudioToTextRecorder:
             pre-trained model from Silero in the ONNX (Open Neural Network
             Exchange) format instead of the PyTorch format. This is
             recommended for faster performance.
-		- silero_deactivity_detection (bool, default=False): Enables the Silero
+        - silero_deactivity_detection (bool, default=False): Enables the Silero
             model for end-of-speech detection. More robust against background
             noise. Utilizes additional GPU resources but improves accuracy in
             noisy environments. When False, uses the default WebRTC VAD,
@@ -363,6 +373,7 @@ class AudioToTextRecorder:
         self.on_recorded_chunk = on_recorded_chunk
         self.on_transcription_start = on_transcription_start
         self.enable_realtime_transcription = enable_realtime_transcription
+        self.use_main_model_for_realtime = use_main_model_for_realtime
         self.realtime_model_type = realtime_model_type
         self.realtime_processing_pause = realtime_processing_pause
         self.on_realtime_transcription_update = (
@@ -410,6 +421,11 @@ class AudioToTextRecorder:
         self.initial_prompt = initial_prompt
         self.suppress_tokens = suppress_tokens
         self.use_wake_words = wake_words or wakeword_backend in {'oww', 'openwakeword', 'openwakewords'}
+        self.detected_language = None
+        self.detected_language_probability = 0
+        self.detected_realtime_language = None
+        self.detected_realtime_language_probability = 0
+        self.transcription_lock = threading.Lock()
 
         # Initialize the logging configuration with the specified level
         log_format = 'RealTimeSTT: %(name)s - %(levelname)s - %(message)s'
@@ -489,7 +505,7 @@ class AudioToTextRecorder:
             )
 
         # Initialize the realtime transcription model
-        if self.enable_realtime_transcription:
+        if self.enable_realtime_transcription and not self.use_main_model_for_realtime:
             try:
                 logging.info("Initializing faster_whisper realtime "
                              f"transcription model {self.realtime_model_type}"
@@ -758,17 +774,16 @@ class AudioToTextRecorder:
                 if conn.poll(0.5):
                     audio, language = conn.recv()
                     try:
-                        segments = model.transcribe(
+                        segments, info = model.transcribe(
                             audio,
                             language=language if language else None,
                             beam_size=beam_size,
                             initial_prompt=initial_prompt,
                             suppress_tokens=suppress_tokens
                         )
-                        segments = segments[0]
                         transcription = " ".join(seg.text for seg in segments)
                         transcription = transcription.strip()
-                        conn.send(('success', transcription))
+                        conn.send(('success', (transcription, info)))
                     except Exception as e:
                         logging.error(f"General transcription error: {e}")
                         conn.send(('error', str(e)))
@@ -964,16 +979,26 @@ class AudioToTextRecorder:
         """
         self._set_state("transcribing")
         audio_copy = copy.deepcopy(self.audio)
-        self.parent_transcription_pipe.send((self.audio, self.language))
-        status, result = self.parent_transcription_pipe.recv()
 
-        self._set_state("inactive")
-        if status == 'success':
-            self.last_transcription_bytes = audio_copy
-            return self._preprocess_output(result)
-        else:
-            logging.error(result)
-            raise Exception(result)
+
+        with self.transcription_lock:
+            try:
+                self.parent_transcription_pipe.send((self.audio, self.language))
+                status, result = self.parent_transcription_pipe.recv()
+
+                self._set_state("inactive")
+                if status == 'success':
+                    segments, info = result
+                    self.detected_language = info.language if info.language_probability > 0 else None
+                    self.detected_language_probability = info.language_probability
+                    self.last_transcription_bytes = audio_copy
+                    return self._preprocess_output(segments)
+                else:
+                    logging.error(f"Transcription error: {result}")
+                    raise Exception(result)
+            except Exception as e:
+                logging.error(f"Error during transcription: {str(e)}")
+                raise e
 
     def _process_wakeword(self, data):
         """
@@ -1446,14 +1471,41 @@ class AudioToTextRecorder:
                     audio_array = audio_array.astype(np.float32) / \
                         INT16_MAX_ABS_VALUE
 
-                    # Perform transcription and assemble the text
-                    segments = self.realtime_model_type.transcribe(
-                        audio_array,
-                        language=self.language if self.language else None,
-                        beam_size=self.beam_size_realtime,
-                        initial_prompt=self.initial_prompt,
-                        suppress_tokens=self.suppress_tokens,
-                    )
+                    if self.use_main_model_for_realtime#:
+                        with self.transcription_lock:
+                            try:
+                                self.parent_transcription_pipe.send((audio_array, self.language))
+                                if self.parent_transcription_pipe.poll(timeout=5):  # Wait for 5 seconds
+                                    status, result = self.parent_transcription_pipe.recv()
+                                    if status == 'success':
+                                        segments, info = result
+                                        self.detected_realtime_language = info.language if info.language_probability > 0 else None
+                                        self.detected_realtime_language_probability = info.language_probability
+                                        realtime_text = segments
+                                    else:
+                                        logging.error(f"Realtime transcription error: {result}")
+                                        continue
+                                else:
+                                    logging.warning("Realtime transcription timed out")
+                                    continue
+                            except Exception as e:
+                                logging.error(f"Error in realtime transcription: {str(e)}")
+                                continue
+                    else:
+                        # Perform transcription and assemble the text
+                        segments, info = self.realtime_model_type.transcribe(
+                            audio_array,
+                            language=self.language if self.language else None,
+                            beam_size=self.beam_size_realtime,
+                            initial_prompt=self.initial_prompt,
+                            suppress_tokens=self.suppress_tokens,
+                        )
+
+                        self.detected_realtime_language = info.language if info.language_probability > 0 else None
+                        self.detected_realtime_language_probability = info.language_probability
+                        realtime_text = " ".join(
+                            seg.text for seg in segments
+                        )
 
                     # double check recording state
                     # because it could have changed mid-transcription
@@ -1461,9 +1513,7 @@ class AudioToTextRecorder:
                             self.recording_start_time > 0.5:
 
                         logging.debug('Starting realtime transcription')
-                        self.realtime_transcription_text = " ".join(
-                            seg.text for seg in segments[0]
-                        )
+                        self.realtime_transcription_text = realtime_text
                         self.realtime_transcription_text = \
                             self.realtime_transcription_text.strip()
 

{RealtimeSTT-0.2.2 → RealtimeSTT-0.2.4}/RealtimeSTT.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: RealtimeSTT
-Version: 0.2.2
+Version: 0.2.4
 Summary: A fast Voice Activity Detection and Transcription System
 Home-page: https://github.com/KoljaB/RealTimeSTT
 Author: Kolja Beigel
@@ -12,6 +12,16 @@ Classifier: Operating System :: OS Independent
 Requires-Python: >=3.6
 Description-Content-Type: text/markdown
 License-File: LICENSE
+Requires-Dist: PyAudio==0.2.14
+Requires-Dist: faster-whisper==1.0.3
+Requires-Dist: pvporcupine==1.9.5
+Requires-Dist: webrtcvad==2.0.10
+Requires-Dist: halo==0.0.31
+Requires-Dist: torch==2.3.1
+Requires-Dist: torchaudio==2.3.1
+Requires-Dist: scipy==1.12.0
+Requires-Dist: websockets==v12.0
+Requires-Dist: openwakeword==0.6.0
 
 
 # RealtimeSTT
@@ -331,6 +341,8 @@ When you initialize the `AudioToTextRecorder` class, you have various options to
 
 - **enable_realtime_transcription** (bool, default=False): Enables or disables real-time transcription of audio. When set to True, the audio will be transcribed continuously as it is being recorded.
 
+- **use_main_model_for_realtime** (bool, default=False): If set to True, the main transcription model will be used for both regular and real-time transcription. If False, a separate model specified by `realtime_model_type` will be used for real-time transcription. Using a single model can save memory and potentially improve performance, but may not be optimized for real-time processing. Using separate models allows for a smaller, faster model for real-time transcription while keeping a more accurate model for final transcription.
+
 - **realtime_model_type** (str, default="tiny"): Specifies the size or path of the machine learning model to be used for real-time transcription.
     - Valid options: 'tiny', 'tiny.en', 'base', 'base.en', 'small', 'small.en', 'medium', 'medium.en', 'large-v1', 'large-v2'.
 

{RealtimeSTT-0.2.2 → RealtimeSTT-0.2.4}/setup.py

@@ -9,7 +9,7 @@ with open('requirements.txt') as f:
 
 setuptools.setup(
     name="RealtimeSTT",
-    version="0.2.2",
+    version="0.2.4",
     author="Kolja Beigel",
     author_email="kolja.beigel@web.de",
     description="A fast Voice Activity Detection and Transcription System",