#!/usr/bin/env python3
"""
Image-to-video test script for Veo 3
This script demonstrates image-to-video generation using the MCP Veo 3 server.
It creates a sample image first, then generates a video from it.
"""
import asyncio
import os
import sys
from pathlib import Path
# Add parent directory to path to import the MCP server
sys.path.insert(0, str(Path(__file__).parent.parent))
from mcp_veo3 import Veo3Client, VideoGenerationConfig
try:
from PIL import Image, ImageDraw, ImageFont
except ImportError:
Image = ImageDraw = ImageFont = None
def create_sample_image():
"""Create a sample image for testing image-to-video"""
if not Image:
print("❌ PIL (Pillow) not installed. Install with: pip install Pillow")
return None
# Create a simple landscape image
width, height = 1280, 720 # 16:9 aspect ratio
image = Image.new('RGB', (width, height), color='skyblue')
draw = ImageDraw.Draw(image)
# Draw a simple landscape
# Ground
draw.rectangle([0, height//2, width, height], fill='green')
# Mountains
mountain_points = [
(0, height//2),
(width//4, height//4),
(width//2, height//3),
(3*width//4, height//5),
(width, height//2)
]
draw.polygon(mountain_points, fill='gray')
# Sun
sun_center = (width - 100, 100)
sun_radius = 50
draw.ellipse([
sun_center[0] - sun_radius,
sun_center[1] - sun_radius,
sun_center[0] + sun_radius,
sun_center[1] + sun_radius
], fill='yellow')
# Save the image
image_path = Path(__file__).parent / "sample_landscape.jpg"
image.save(image_path, "JPEG", quality=95)
print(f"📸 Created sample image: {image_path}")
return str(image_path)
async def test_image_to_video():
"""Test image-to-video generation"""
print("🖼️ Testing Image-to-Video Generation")
print("=" * 50)
# Check API key
api_key = os.getenv("GEMINI_API_KEY")
if not api_key:
print("❌ Error: GEMINI_API_KEY environment variable not set")
print("Please set your Gemini API key:")
print("export GEMINI_API_KEY='your_api_key_here'")
return
# Create sample image
image_path = create_sample_image()
if not image_path:
return
try:
# Initialize client
print("🔧 Initializing Veo 3 client...")
client = Veo3Client(api_key)
# Configure generation
config = VideoGenerationConfig(
model="veo-3.0-fast-generate-preview",
aspect_ratio="16:9",
output_dir="test_videos"
)
# Motion prompt for the landscape
prompt = "The sun slowly moves across the sky while clouds drift peacefully overhead, creating moving shadows on the mountains"
print(f"🖼️ Source image: {image_path}")
print(f"📝 Motion prompt: {prompt}")
print(f"🎬 Model: {config.model}")
print("\n🚀 Starting image-to-video generation...")
print("⏱️ This may take 1-6 minutes...")
# Generate video from image
result = await client.generate_video(prompt, config, image_path=image_path)
# Display results
if result["success"]:
print("\n✅ Image-to-video generation completed!")
print(f"📁 Video file: {result['video_path']}")
print(f"🖼️ Source image: {image_path}")
print(f"⏱️ Generation time: {result['generation_time']:.1f} seconds")
print(f"📏 File size: {result['file_size'] / 1024 / 1024:.1f} MB")
else:
print(f"\n❌ Image-to-video generation failed: {result['error']}")
except Exception as e:
print(f"❌ Test failed with error: {str(e)}")
async def test_with_existing_image():
"""Test with an existing image file"""
print("\n📁 Testing with User-Provided Image")
print("=" * 50)
# Look for common image files in the current directory
current_dir = Path(__file__).parent
image_extensions = ['.jpg', '.jpeg', '.png', '.webp', '.bmp']
existing_images = []
for ext in image_extensions:
existing_images.extend(current_dir.glob(f"*{ext}"))
existing_images.extend(current_dir.glob(f"*{ext.upper()}"))
if not existing_images:
print("📂 No existing image files found in examples directory")
print("💡 To test with your own image:")
print(" 1. Copy an image file to the examples directory")
print(" 2. Run this script again")
return
# Use the first found image
image_path = str(existing_images[0])
api_key = os.getenv("GEMINI_API_KEY")
if not api_key:
print("❌ Error: GEMINI_API_KEY environment variable not set")
return
try:
client = Veo3Client(api_key)
config = VideoGenerationConfig(
model="veo-3.0-fast-generate-preview",
aspect_ratio="16:9",
output_dir="test_videos"
)
# Generic motion prompt
prompt = "Gentle camera movement revealing more of the scene, with natural lighting changes and subtle motion"
print(f"🖼️ Using existing image: {image_path}")
print(f"📝 Motion prompt: {prompt}")
print("\n🚀 Starting generation...")
result = await client.generate_video(prompt, config, image_path=image_path)
if result["success"]:
print("\n✅ Generation with existing image completed!")
print(f"📁 Video file: {result['video_path']}")
else:
print(f"\n❌ Generation failed: {result['error']}")
except Exception as e:
print(f"❌ Test failed with error: {str(e)}")
async def main():
"""Run all image-to-video tests"""
print("🎯 Image-to-Video Test Suite")
print("=" * 50)
await test_image_to_video()
await test_with_existing_image()
print("\n🎉 Image-to-video test suite completed!")
print("\n💡 Tips:")
print("- Image-to-video works best with clear, well-lit images")
print("- Describe the motion you want to see in your prompt")
print("- Higher resolution source images generally produce better results")
if __name__ == "__main__":
asyncio.run(main())
