app.py

import streamlit as st
from streamlit_mic_recorder import mic_recorder
from transformers import pipeline
import torch
from transformers import BertTokenizer, BertForSequenceClassification

def callback():
    if st.session_state.my_recorder_output:
        audio_bytes = st.session_state.my_recorder_output['bytes']
        st.audio(audio_bytes)

def transcribe(upload):
    pipe = pipeline("automatic-speech-recognition", model="openai/whisper-large")
    result = pipe(upload, generate_kwargs={'task': 'transcribe'})
    print(result['text'])
    return result['text']

def encode(docs, tokenizer):
    '''
    This function takes list of texts and returns input_ids and attention_mask of texts
    '''
    encoded_dict = tokenizer.batch_encode_plus(docs, add_special_tokens=True, max_length=128, padding='max_length',
                            return_attention_mask=True, truncation=True, return_tensors='pt')
    input_ids = encoded_dict['input_ids']
    attention_masks = encoded_dict['attention_mask']
    return input_ids, attention_masks


def load_model():
    CUSTOMMODEL_PATH = "./bert-itserviceclassification"
    PRETRAINED_LM = "bert-base-uncased"
    tokenizer = BertTokenizer.from_pretrained(PRETRAINED_LM, do_lower_case=True)
    model = BertForSequenceClassification.from_pretrained(PRETRAINED_LM,
                                                        num_labels=8,
                                                        output_attentions=False,
                                                        output_hidden_states=False)
    model.load_state_dict(torch.load(CUSTOMMODEL_PATH, map_location ='cpu'))
    return model, tokenizer


def predict(text, model, tokenizer):
    lookup_key ={0: 'Hardware',
    1: 'Access',
    2: 'Miscellaneous',
    3: 'HR Support',
    4: 'Purchase',
    5: 'Administrative rights',
    6: 'Storage',
    7: 'Internal Project'}
    with torch.no_grad():
        input_ids, att_mask = encode([text], tokenizer)
        logits = model(input_ids = input_ids, attention_mask=att_mask).logits
    predicted_class_id = logits.argmax().item()
    predicted_label = lookup_key.get(predicted_class_id)
    return predicted_label


def main():

    st.set_page_config(layout="wide", page_title="IT Service NLP Classification",)

    with st.sidebar:
        audio = mic_recorder(key='my_recorder', callback=callback)
        button = st.button('start classification')

    if button:
        st.write('Loading')
        text = transcribe(upload=audio["bytes"])
        st.write(f'Speech-to-text Result:')
        st.write(f'{text}')
        model, tokenizer = load_model()
        prediction = predict(text=text, model=model, tokenizer=tokenizer)
        st.write(f'Classifcation Result:')
        st.write(f'{prediction}')

if __name__ == '__main__':
    main()