import streamlit as st
from transformers import pipeline

# function part
# img2text
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"]
    return text

# text2story
def text2story(text):
    text_to_story_model = pipeline("text-generation", model="pranavpsv/genre-story-generator-v2")
    story_text = text_to_story_model(text)[0]["generated_text"]
    return story_text

# text2audio
def text2audio(story_text):
    text_to_audio_model = pipeline("text-to-speech", model="facebook/mms-tts-eng")
    audio_data = text_to_audio_model(story_text)
    return audio_data


# Main Part
st.set_page_config(page_title="Your Image to Audio Story",
                   page_icon="🦜")
st.header("Turn Your Image to Audio Story")
uploaded_file = st.file_uploader("Select an Image...")

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_container_width=True)

    # Stage 1: Image to Text
    if "scenario" not in st.session_state:
        st.text('Processing img2text...')
        st.session_state.scenario = img2text(uploaded_file.name)
    st.subheader("Caption:")
    st.write(st.session_state.scenario)

    # Stage 2: Text to Story
    if "story" not in st.session_state:
        st.text('Generating a story...')
        st.session_state.story = text2story(st.session_state.scenario)
    st.subheader("Story:")
    st.write(st.session_state.story)

    # Stage 3: Story to Audio
    if "audio_data" not in st.session_state:
        st.text('Generating audio...')
        st.session_state.audio_data = text2audio(st.session_state.story)

    # Play Button
    if st.button("Play Audio"):
        st.audio(
            st.session_state.audio_data['audio'],
            format="audio/wav",
            start_time=0,
            sample_rate=st.session_state.audio_data['sampling_rate']
        )