Aitool

# -*- coding: utf-8 -*-
"""AI Video Generator - Complete Text to Video + Audio System"""

# Install dependencies
print("📦 Installing dependencies...")
!pip install diffusers accelerate torch transformers coqui-tts opencv-python pillow ffmpeg-python xformers -q
!apt update && apt install -y ffmpeg > /dev/null 2>&1

import torch
import numpy as np
from diffusers import StableVideoDiffusionPipeline
from coqui_tts import TTS
import subprocess
import os
from IPython.display import display, HTML
import base64
import cv2
import ffmpeg
from pathlib import Path

class AIVideoGenerator:
    def __init__(self):
        self.video_pipe = None
        self.tts_model = None
        
    def load_models(self):
        """Load AI models"""
        print("🔄 Loading models...")
        
        # Load video generation model
        self.video_pipe = StableVideoDiffusionPipeline.from_pretrained(
            "stabilityai/stable-video-diffusion-img2vid-xt",
            torch_dtype=torch.float16,
            variant="fp16"
        )
        self.video_pipe.enable_model_cpu_offload()
        
        # Load TTS model
        self.tts_model = TTS(model_name="tts_models/en/ljspeech/tacotron2-DDC", progress_bar=False)
        print("✅ Models loaded successfully!")
    
    def create_video_from_frames(self, frames, output_path, fps=8):
        """Convert frames to video using OpenCV"""
        if not frames:
            return None
        
        height, width = frames[0].shape[:2]
        fourcc = cv2.VideoWriter_fourcc(*'mp4v')
        out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))
        
        for frame in frames:
            # Convert RGB to BGR for OpenCV
            frame_bgr = cv2.cvtColor(frame, cv2.COLOR_RGB2BGR)
            out.write(frame_bgr)
        out.release()
        return output_path
    
    def generate_video(self, prompt, duration=4, seed=42):
        """Generate video from text prompt"""
        print(f"🎬 Generating video: {prompt}")
        
        if self.video_pipe is None:
            self.load_models()
        
        generator = torch.manual_seed(seed)
        frames = self.video_pipe(
            prompt,
            num_frames=duration * 8,
            generator=generator,
            decode_chunk_size=4
        ).frames[0]
        
        video_path = "generated_video.mp4"
        self.create_video_from_frames(frames, video_path)
        return video_path
    
    def generate_audio(self, text, output_path="generated_audio.wav"):
        """Generate audio from text using TTS"""
        print(f"🔊 Generating audio: {text}")
        
        if self.tts_model is None:
            self.load_models()
        
        try:
            self.tts_model.tts_to_file(text=text, file_path=output_path)
            return output_path
        except Exception as e:
            print(f"❌ TTS failed: {e}")
            return None
    
    def merge_audio_video(self, video_path, audio_path, output_path="final_video.mp4"):
        """Merge audio and video using FFmpeg"""
        print("🔄 Merging audio and video...")
        
        try:
            input_video = ffmpeg.input(video_path)
            input_audio = ffmpeg.input(audio_path)
            
            ffmpeg.output(
                input_video, 
                input_audio, 
                output_path,
                vcodec='libx264',
                acodec='aac',
                strict='experimental'
            ).overwrite_output().run(quiet=True)
            
            return output_path
        except Exception as e:
            print(f"❌ Merge failed: {e}")
            return video_path
    
    def display_video(self, video_path):
        """Display video in Colab"""
        try:
            with open(video_path, 'rb') as f:
                video_bytes = f.read()
            
            video_b64 = base64.b64encode(video_bytes).decode()
            html = f'''
            <video width="640" height="480" controls autoplay>
                <source src="data:video/mp4;base64,{video_b64}" type="video/mp4">
            </video>
            '''
            display(HTML(html))
        except Exception as e:
            print(f"❌ Display failed: {e}")

# MAIN EXECUTION
def main():
    generator = AIVideoGenerator()
    
    # Configuration
    prompt = "A beautiful sunset over mountains, cinematic style, 4K, high quality"
    audio_text = "This is a beautiful sunset scene generated by artificial intelligence."
    duration = 4
    seed = 42
    
    print("🚀 Starting AI Video Generation Pipeline...")
    
    try:
        # Generate video
        video_path = generator.generate_video(prompt, duration, seed)
        
        # Generate audio
        audio_path = generator.generate_audio(audio_text)
        
        # Merge if audio generated successfully
        if audio_path and os.path.exists(audio_path):
            final_path = generator.merge_audio_video(video_path, audio_path)
        else:
            final_path = video_path
        
        # Display result
        print(f"✅ Final video saved: {final_path}")
        generator.display_video(final_path)
        
        # Provide download link
        if os.path.exists(final_path):
            file_size = os.path.getsize(final_path) / (1024 * 1024)
            print(f"📁 File size: {file_size:.2f} MB")
            
    except Exception as e:
        print(f"❌ Generation failed: {e}")
    
    finally:
        # Cleanup
        print("🧹 Cleaning up temporary files...")
        for path in ["generated_video.mp4", "generated_audio.wav", "final_video.mp4"]:
            if os.path.exists(path):
                os.remove(path)

# Run the complete pipeline
if __name__ == "__main__":
    main()
    fastapi==0.104.1
uvicorn==0.24.0
diffusers==0.24.0
accelerate==0.24.1
torch==2.1.0
transformers==4.35.0
coqui-tts==0.11.0
ffmpeg-python==0.2.0
python-multipart==0.0.6
aiofiles==23.2.1
pillow==10.0.1
numpy==1.24.3
opencv-python==4.8.1.78
python-dotenv==1.0.0