Spaces:
Sleeping
Sleeping
| # 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. " \ | |
| f"The story should be about: {text}. " \ | |
| "Make it cheerful, imaginative, and suitable for young children. " \ | |
| "Avoid any scary or sad elements. The story must be exactly 95 words." | |
| # Generate the story | |
| story = text_generator(prompt, max_length=150, num_return_sequences=1)[0]["generated_text"] | |
| # Clean up the output to remove the prompt | |
| story = story.replace(prompt, "").strip() | |
| # Ensure the story is exactly 95 words | |
| words = story.split() | |
| if len(words) > 95: | |
| story = " ".join(words[:95]) | |
| elif len(words) < 95: | |
| # If the story is too short, pad it with a happy ending | |
| story = " ".join(words) + " They all laughed and played together, making it the best day ever!" | |
| words = story.split() | |
| if len(words) > 95: | |
| story = " ".join(words[:95]) | |
| return story | |
| # Function to convert text to audio using gTTS | |
| def text2audio(story_text): | |
| audio_file = "kids_playing_audio.wav" # Use the correct file name | |
| 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/wav") | |
| # 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!") |