import streamlit as st
import torch
from transformers import pipeline
from moviepy.editor import *
from faster_whisper import WhisperModel
import os
import random
import requests

# Set page configuration
st.set_page_config(page_title="AI Text-to-Video Generator", layout="wide")

# Streamlit UI
st.title("🚀 AI Text-to-Video Generator")
st.markdown("Generate videos from text using AI-powered tools. 🎥🎙️")

# Text-to-Video Model
text_to_video = pipeline("text-to-video", model="damo-vilab/modelscope-text-to-video-synthesis")

# Text-to-Speech Model
tts_pipeline = pipeline("text-to-speech", model="coqui-ai/TTS")

# Whisper Speech-to-Text Model
whisper_model = WhisperModel("base", device="cuda" if torch.cuda.is_available() else "cpu")

# Function to Generate Video from Text
def generate_video(text_script, use_ai_material=True):
    output_video_path = f"output_{random.randint(1000,9999)}.mp4"
    
    if use_ai_material:
        video = text_to_video(text_script)
        video_path = video["video"]
    else:
        video_path = "local_video.mp4"  # Placeholder for user-uploaded video

    return video_path

# Function to Generate Voiceover
def generate_voice(text_script, voice_choice="default"):
    audio_path = f"voice_{random.randint(1000,9999)}.wav"
    tts_audio = tts_pipeline(text_script)
    
    with open(audio_path, "wb") as f:
        f.write(tts_audio["audio"])
    
    return audio_path

# Function to Auto-Generate Captions
def generate_captions(audio_path):
    segments, _ = whisper_model.transcribe(audio_path)
    captions = "\n".join([segment.text for segment in segments])
    
    return captions

# User Inputs
text_script = st.text_area("✍️ Enter your script:", "Type your video script here...")
use_ai_material = st.checkbox("Use AI-Generated Material", value=True)
voice_choice = st.selectbox("🎙️ Choose Voice Type:", ["default", "robotic", "natural"])

# Generate Button
if st.button("🎬 Generate Video"):
    if text_script:
        with st.spinner("Generating Video... 🚀"):
            video_path = generate_video(text_script, use_ai_material)
            audio_path = generate_voice(text_script, voice_choice)
            
            # Combine video with AI-generated voice
            video_clip = VideoFileClip(video_path)
            audio_clip = AudioFileClip(audio_path)
            final_clip = video_clip.set_audio(audio_clip)
            
            final_video_path = f"final_{random.randint(1000,9999)}.mp4"
            final_clip.write_videofile(final_video_path, codec="libx264")

            # Generate Auto-Captions
            captions = generate_captions(audio_path)
            
            # Display Results
            st.video(final_video_path)
            st.markdown("### 📝 Auto-Generated Captions:")
            st.write(captions)

    else:
        st.warning("Please enter a script to generate a video.")

# Video Search System
st.markdown("---")
st.markdown("## 🔍 Search AI-Generated Videos")

search_query = st.text_input("Search for a video by script content:")
if st.button("🔍 Search"):
    # Simulated Search Logic (Replace with actual database search)
    dummy_videos = {
        "AI Video 1": "final_1234.mp4",
        "AI Video 2": "final_5678.mp4"
    }
    
    found_videos = {title: path for title, path in dummy_videos.items() if search_query.lower() in title.lower()}
    
    if found_videos:
        for title, path in found_videos.items():
            st.markdown(f"### 🎥 {title}")
            st.video(path)
    else:
        st.write("❌ No videos found matching your search.")