app.py

import warnings
warnings.filterwarnings("ignore")

import librosa    # Library for loading and processing audio files.
import numpy as np      # Library for numerical computations, used for signal processing.
import gradio as gr     # Library for creating a web-based user interface for inference.
from transformers import pipeline     # Import pipeline for automatic speech recognition (ASR).
from scipy.signal import butter, lfilter, wiener  # Signal processing functions for noise reduction.

# Importing custom utility functions for text processing.
from text2int import text_to_int   # Converts text numbers (e.g., "one") into integers (e.g., 1).
from isNumber import is_number   # Checks if a string is a number.
from Text2List import text_to_list   # Converts a text string into a list of words.
from convert2list import convert_to_list   # Converts processed text into a structured list.
from processDoubles import process_doubles   # Handles repeated words or numbers in speech recognition output.
from replaceWords import replace_words   # Replaces specific words in the recognized text with alternatives.
from highPassFiler import high_pass_filter
form WaveletDenoise import wavelet_denoise
fomr applyWienerFilter import apply_wiener_filter

# Initialize ASR model pipeline
asr_model = pipeline("automatic-speech-recognition", model="cdactvm/w2v-bert-punjabi")


# Function to handle speech recognition
def recognize_speech(audio_file):
    audio, sr = librosa.load(audio_file, sr=16000)
    audio = high_pass_filter(audio, sr)
    audio = apply_wiener_filter(audio)
    denoised_audio = wavelet_denoise(audio)
    result = asr_model(denoised_audio)
    text_value = result['text']
    cleaned_text = text_value.replace("[PAD]", "")
    converted_to_list = convert_to_list(cleaned_text, text_to_list())
    processed_doubles = process_doubles(converted_to_list)
    replaced_words = replace_words(processed_doubles)
    converted_text = text_to_int(replaced_words)
    return converted_text


def sel_lng(lng, mic=None, file=None):
    if mic is not None:
        audio = mic
    elif file is not None:
        audio = file
    else:
        return "You must either provide a mic recording or a file"
    
    if lng == "model_1":
        return recognize_speech(audio)
        

# Create a Gradio interface
demo = gr.Interface(
    fn=sel_lng, 
    inputs=[
        gr.Dropdown(["model_1"], label="Select Model"),
        gr.Audio(sources=["microphone", "upload"], type="filepath"),
    ],
    outputs=["textbox"],
    title="Automatic Speech Recognition",
    description="Demo for Automatic Speech Recognition. Use microphone to record speech. Please press Record button. Initially, it will take some time to load the model. The recognized text will appear in the output textbox"
)

demo.launch()