app.py

import os
import time
from typing import Any
import requests
import streamlit as st
from dotenv import find_dotenv, load_dotenv
from langchain.chains import LLMChain
from langchain.chat_models import ChatOpenAI
from langchain.prompts import PromptTemplate
from transformers import pipeline
from gtts import gTTS
import google.generativeai as genai

from custom import css_code

load_dotenv(find_dotenv())
HUGGINGFACE_API_TOKEN = os.getenv("HUGGINGFACE_API_TOKEN")
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")


def progress_bar(amount_of_time: int) -> Any:
    progress_text = "Please wait, Generative models hard at work"
    my_bar = st.progress(0, text=progress_text)

    for percent_complete in range(amount_of_time):
        time.sleep(0.04)
        my_bar.progress(percent_complete + 1, text=progress_text)
    time.sleep(1)
    my_bar.empty()


def generate_text_from_image(url: str) -> str:
    image_to_text: Any = pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
    generated_text: str = image_to_text(url)[0]["generated_text"]
    return generated_text


def generate_story_from_text(scenario: str) -> str:
    prompt_template: str = f"""
    You are a talented story teller who can create a story from a simple narrative.
    Create a story using the following scenario; the story should have a maximum of 1050 tokens;

    CONTEXT: {scenario}
    STORY:
    """

    prompt: PromptTemplate = PromptTemplate(template=prompt_template, input_variables=["scenario"])
    model = genai.GenerativeModel("gemini-1.5-flash")

    response = model.generate_content(scenario, generation_config=genai.types.GenerationConfig(max_output_tokens=1050))
    generated_story: str = response.text
    return generated_story


def generate_speech_from_text(message: str) -> str:
    # Generate audio file using gTTS
    tts = gTTS(text=message, lang='en')
    audio_file = "generated_audio.mp3"
    tts.save(audio_file)
    return audio_file


def main() -> None:
    st.set_page_config(page_title="IMAGE TO STORY CONVERTER", page_icon="🖼️")
    st.markdown(css_code, unsafe_allow_html=True)

    with st.sidebar:
        st.image("images.jpg")
        st.write("---")
        st.write("AI StoryTeller @ Anjali Dhoundiyal")

    st.header("Image-to-Story Converter")
    uploaded_file: Any = st.file_uploader("Please choose a file to upload", type="jpg")

    if uploaded_file is not None:
        bytes_data: Any = 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)
        progress_bar(100)
        scenario: str = generate_text_from_image(uploaded_file.name)
        story: str = generate_story_from_text(scenario)
        audio_file = generate_speech_from_text(story)

        with st.expander("Generated Image scenario"):
            st.write(scenario)
        with st.expander("Generated short story"):
            st.write(story)

        st.audio(audio_file)  # Play the generated audio file


if __name__ == "__main__":
    main()