app.py

from dotenv import find_dotenv, load_dotenv
from transformers import pipeline
import streamlit as st
import os


# load env variables from .env file
load_dotenv(find_dotenv())

# img to text
def img_to_text(url):
    image_to_text = pipeline("image-to-text", model="Salesforce/blip-image-captioning-large")

    text = image_to_text(url)[0]["generated_text"]
    return text


# llm
def generate_story(text):
    generator = pipeline("text-generation", model="distilgpt2")

    result = generator(text, max_length=20, num_return_sequences=1)
    return result[0]['generated_text']


#
# text-to-speech
def text_to_speech(text):
    import requests

    API_URL = "https://api-inference.huggingface.co/models/espnet/kan-bayashi_ljspeech_vits"
    headers = {"Authorization": f"Bearer {os.environ.get('HUGGINGFACE_API_TOKEN')}"}
    payload = {
        "inputs": text
    }

    response = requests.post(API_URL, headers=headers, json=payload)
    response.raise_for_status()
    with open('audio.flac', 'wb') as file:
        file.write(response.content)



def main():
    st.set_page_config(page_title="img to audio story")
    st.header("turn image to audio story")
    uploaded_file = st.file_uploader("Choose an image ... ", type="jpg")

    if uploaded_file is not None:
        print(uploaded_file)
        bytes_data = uploaded_file.getvalue()
        with open(uploaded_file.name, "wb") as file:
            file.write(bytes_data)
        st.image(uploaded_file, caption="Uploaded image", use_column_width=True)
        text = img_to_text(uploaded_file.name)
        story = generate_story(text)
        text_to_speech(story)

        with st.expander("text"):
            st.write(text)
        with st.expander("story"):
            st.write(story)
        st.audio("audio.flac")

main()