app.py

import streamlit as st
import firebase_admin
from firebase_admin import credentials
from firebase_admin import firestore
import datetime
from transformers import pipeline
import gradio as gr

@st.experimental_singleton
def get_db_firestore():
    cred = credentials.Certificate('test.json')
    firebase_admin.initialize_app(cred, {'projectId': u'clinical-nlp-b9117',})
    db = firestore.client()
    return db


db = get_db_firestore()
asr = pipeline("automatic-speech-recognition", "facebook/wav2vec2-base-960h")

def transcribe(audio):
    text = asr(audio)["text"]
    return text

classifier = pipeline("text-classification")

def speech_to_text(speech):
    text = asr(speech)["text"]
    return text

def text_to_sentiment(text):
    sentiment = classifier(text)[0]["label"]
    return sentiment 

def upsert(text):
    date_time =str(datetime.datetime.today())
    doc_ref = db.collection('Text2SpeechSentimentSave').document('Text2SpeechSentimentSave')
    doc_ref.set({u'firefield': date_time, u'first': 'Text2SpeechSentimentSave', u'last': text, u'born': date_time,})
    saved = select('Text2SpeechSentimentSave','Text2SpeechSentimentSave')
    return saved
      
def select(collection, document):
    doc_ref = db.collection(collection).document(document)
    doc = doc_ref.get()
    docid = ("The id is: ", doc.id)
    contents = ("The contents are: ", doc.to_dict())
    return contents
    
demo = gr.Blocks()

with demo:
    #audio_file = gr.Audio(type="filepath")
    audio_file = gr.inputs.Audio(source="microphone", type="filepath")
    text = gr.Textbox()
    label = gr.Label()
    saved = gr.Textbox()

    b1 = gr.Button("Recognize Speech")
    b2 = gr.Button("Classify Sentiment")
    b3 = gr.Button("Save Speech to Text")

    b1.click(speech_to_text, inputs=audio_file, outputs=text)
    b2.click(text_to_sentiment, inputs=text, outputs=label)
    b3.click(upsert, inputs=text, outputs=saved)

demo.launch(share=True)