| import os |
| import json |
| import vosk |
|
|
| class SpeechRecognitionService: |
| def __init__(self, model_path=None): |
| if model_path is None: |
| base_dir = os.path.dirname(os.path.dirname(os.path.abspath(__file__))) |
| model_path = os.path.join(base_dir, "models", "vosk-model-small-en-us-0.15") |
| |
| if not os.path.exists(model_path): |
| raise FileNotFoundError(f"Vosk model not found at {model_path}") |
|
|
| print(f"Loading Vosk model from {model_path}...") |
| self.model = vosk.Model(model_path) |
|
|
| def transcribe(self, audio_data, samplerate): |
| recognizer = vosk.KaldiRecognizer(self.model, samplerate) |
| if recognizer.AcceptWaveform(audio_data): |
| result = json.loads(recognizer.Result()) |
| return result.get("text", "") |
| else: |
| partial = json.loads(recognizer.PartialResult()) |
| return partial.get("partial", "") |
|
|
| def transcribe_full(self, audio_data, samplerate): |
| recognizer = vosk.KaldiRecognizer(self.model, samplerate) |
| recognizer.AcceptWaveform(audio_data) |
| result = json.loads(recognizer.FinalResult()) |
| return result.get("text", "") |
|
|
