src/transcription/file_transcriber.py

import numpy as np
from faster_whisper import WhisperModel
import librosa
import logging

logger = logging.getLogger(__name__)

class FileTranscriber:
    def __init__(self, model_size="base", device="cpu", compute_type="int8"):
        self.model_size = model_size
        self.device = device
        self.compute_type = compute_type
        self.model = None

    def load_model(self):
        if self.model is None:
            logger.info(f"Loading Whisper model: {self.model_size}")
            try:
                self.model = WhisperModel(self.model_size, device=self.device, compute_type=self.compute_type)
                logger.info("Model loaded successfully")
            except Exception as e:
                logger.error(f"Failed to load model: {e}")
                raise

    def transcribe_file(self, file_path):
        self.load_model()
        try:
            # Load audio
            audio_data, sample_rate = librosa.load(file_path, sr=16000)
            audio_data = audio_data.astype(np.float32)

            # Transcribe
            segments, _ = self.model.transcribe(audio_data, beam_size=5, vad_filter=True)

            transcription = ""
            for segment in segments:
                text = segment.text.strip()
                if text:
                    transcription += text + " "

            return transcription.strip()

        except Exception as e:
            logger.error(f"Error transcribing file {file_path}: {e}")
            return f"Error: {str(e)}"