app.py

# Import necessary libraries
import streamlit as st
from transformers import pipeline
from gtts import gTTS
import os

# Function to convert image to text
def img2text(url):
    image_to_text_model = pipeline("image-to-text", model="Salesforce/blip-image-captioning-base")
    text = image_to_text_model(url)[0]["generated_text"]
    
    # Remove unwanted words like "illustration"
    unwanted_words = ["illustration", "painting", "drawing", "artwork"]
    for word in unwanted_words:
        text = text.replace(word, "")
    return text.strip()

# Function to generate a story from text using GPT-Neo
def text2story(text):
    # Use the GPT-Neo model for storytelling
    text_generator = pipeline("text-generation", model="EleutherAI/gpt-neo-125M")
    
    # Add a strong prompt to guide the model
    prompt = f"Write a short, happy, and fun story for kids aged 3-10 based on the following description: {text}. " \
             "The story should be cheerful, imaginative, and suitable for young children. " \
             "Avoid any scary or sad elements. Keep the story under 95 words and make sure it has a clear beginning, middle, and end."
    
    # Generate the story
    story = text_generator(prompt, max_length=95, num_return_sequences=1)[0]["generated_text"]
    
    # Clean up the output to remove the prompt (if necessary)
    story = story.replace(prompt, "").strip()
    return story

# Function to convert text to audio using gTTS
def text2audio(story_text):
    audio_file = "story_audio.mp3"
    tts = gTTS(story_text, lang="en")
    tts.save(audio_file)
    return audio_file

# Main application
st.set_page_config(page_title="Image to Story", page_icon="📖")
st.header("📖 Image to Story")
st.markdown("### Turn your image into a fun story!")

# Initialize session state
if "story" not in st.session_state:
    st.session_state.story = None
if "audio_file" not in st.session_state:
    st.session_state.audio_file = None

uploaded_file = st.file_uploader("Select an Image...", type=["jpg", "jpeg", "png"])

if uploaded_file is not None:
    # Save the uploaded file
    bytes_data = uploaded_file.getvalue()
    with open(uploaded_file.name, "wb") as file:
        file.write(bytes_data)

    st.image(uploaded_file, caption="Your Uploaded Image", use_container_width=True)

    # Stage 1: Image to Text
    st.text('🖼️ Processing image...')
    scenario = img2text(uploaded_file.name)
    st.write("**What I see:**", scenario)

    # Stage 2: Text to Story
    if st.session_state.story is None or st.button("🔄 Generate New Story"):
        st.text('📝 Creating a story...')
        st.session_state.story = text2story(scenario)
        st.write("**Your Story:**", st.session_state.story)

        # Stage 3: Story to Audio
        st.text('🎙️ Turning your story into audio...')
        st.session_state.audio_file = text2audio(st.session_state.story)

    # Play button for audio
    if st.session_state.audio_file and st.button("🎧 Listen to the Story"):
        st.audio(st.session_state.audio_file, format="audio/mp3")

    # Clean up the generated audio file
    if st.session_state.audio_file and os.path.exists(st.session_state.audio_file):
        os.remove(st.session_state.audio_file)

# Add some fun prompts for kids
st.markdown("### 🎨 Tips for a Great Story!")
st.write("1. Upload a picture of your favorite animal, place, or toy!")
st.write("2. Imagine what's happening in the picture and let the story begin!")
st.write("3. Listen to your story and share it with your friends!")