app.py

import gradio as gr
from speechbrain.inference.ASR import EncoderASR
from pydub import AudioSegment
import os
import numpy as np
from scipy.io.wavfile import write
import time


# Load the ASR model
asr_model = EncoderASR.from_hparams(
    source="YosefA/wave2vec2_amharic_stt", 
    savedir="pretrained_models/asr-wav2vec2-amharic"
)

# Directory to store converted audio files
os.makedirs("temp_audio", exist_ok=True)

def transcribe_audio(audio_file):
    """
    Converts the uploaded audio to .wav format, runs transcription, and returns the result.
    """
    # Save the uploaded audio to a temporary location
    temp_audio_path = "temp_audio/input_audio.wav"
    
    # Convert audio to .wav format
    sound = AudioSegment.from_file(audio_file)
    sound.export(temp_audio_path, format="wav")
    
    # Transcribe the audio
    transcription = asr_model.transcribe_file(temp_audio_path)
    
    # Clean up temporary files (optional)
    os.remove(temp_audio_path)
    
    return transcription

def process_audio(audio_data):
    """
    Processes recorded/uploaded audio, saves it, and sends it to the transcribe_audio function.
    """  
    transcription = transcribe_audio(audio_data)  
    return transcription


# Define the Gradio interface
with gr.Blocks() as app:
    gr.Markdown("### Amharic Speech-to-Text Transcription App")
    gr.Markdown("Upload or record an audio file in any format, and get its transcription.")
    
    with gr.Row():
        audio_input = gr.Audio(label="Upload or Record Audio", type="filepath") 
        transcription_output = gr.Textbox(label="Transcription")
    
    transcribe_button = gr.Button("Transcribe")
    transcribe_button.click(process_audio, inputs=audio_input, outputs=transcription_output)

# Launch the app
if __name__ == "__main__":
    app.launch()