app.py

# import part
import streamlit as st
from transformers import pipeline
import os

# function part
# img2text
def img2text(url):
    try:
        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"
        text = text.replace("illustration", "").strip()
        # Make the caption more fun and happy
        fun_caption = f"🌟 Wow! This picture shows {text.lower()}. Let’s turn it into a fun story! 🌟"
        return fun_caption
    except Exception as e:
        st.error(f"Oops! Something went wrong while looking at your picture. Please try again! 😊")
        return None

# text2story
def text2story(text):
    try:
        # Use a better model for text generation (e.g., GPT-Neo)
        story_generator = pipeline("text-generation", model="EleutherAI/gpt-neo-125M")
        # Add a fun and happy prompt to guide the story generation
        prompt = f"One sunny day, {text}. "
        story = story_generator(prompt, max_length=100, num_return_sequences=1)[0]["generated_text"]
        # Remove any unwanted text (e.g., usernames, special characters)
        story = " ".join([word for word in story.split() if not word.startswith("@") and not word.startswith("http")])
        # Make the story more fun by adding a happy ending
        happy_story = story + " And everyone had a big smile on their faces at the end of the day! 😄🌈"
        return happy_story
    except Exception as e:
        st.error(f"Oops! Something went wrong while creating your story. Please try again! 😊")
        return None

# text2audio
def text2audio(story_text):
    try:
        # Use a reliable TTS model (e.g., ESPnet's VITS model)
        tts_pipeline = pipeline("text-to-speech", model="espnet/kan-bayashi_ljspeech_vits")
        audio_output = tts_pipeline(story_text)
        audio_file = "story_audio.wav"
        # Save the audio file
        with open(audio_file, "wb") as f:
            f.write(audio_output["audio"])
        return audio_file
    except Exception as e:
        st.error(f"Oops! Something went wrong while turning your story into audio. Please try again! 😊")
        return None

# main part
st.set_page_config(page_title="Story Maker", page_icon="😊")
st.header("😄 Story Maker: Turn Your Picture into a Happy Story! 😄")
uploaded_file = st.file_uploader("📷 Choose a picture to create a fun story...", type=["jpg", "jpeg", "png"])

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

    st.image(uploaded_file, caption="Your fun picture!", use_container_width=True)

    # Stage 1: Image to Text
    st.write("🌟 Let’s see what’s in your picture... 🌟")
    scenario = img2text(uploaded_file.name)
    if scenario:
        st.write(scenario)

        # Stage 2: Text to Story
        st.write("📖 Turning your picture into a fun story... 📖")
        story = text2story(scenario)
        if story:
            st.write("📖 Here’s your fun story:")
            st.write(story)

            # Stage 3: Story to Audio data
            st.write("🎤 Getting ready to tell your story... 🎤")
            audio_file = text2audio(story)
            if audio_file:
                # Play button
                if st.button("🎧 Listen to Your Story!"):
                    st.audio(audio_file, format="audio/wav")
                # Clean up the audio file after playing
                os.remove(audio_file)