image-search-api / sdk /python /examples /plug_and_run_example.py
mbishu's picture
Initialize SDK and fix vector store issues
a475b1f
Raw History Blame Contribute Delete
8.75 kB
"""
Plug-and-Run Integration Example - Image Search SDK
This example demonstrates how customers can integrate image search
into their existing e-commerce systems with automatic embedding generation.
"""
from image_search_sdk import ImageSearchClient, Product, SearchOptions
def main():
print("πŸš€ Plug-and-Run Image Search Integration Example")
print("=" * 60)
# 1. Initialize client with different storage backends
print("\n1. Initializing Image Search Client...")
# Option A: Local file storage (simple, no database setup)
client_local = ImageSearchClient(
api_key="is_your_api_key_here",
base_url="http://localhost:8000",
embedding_storage="local",
storage_config={
'storage_path': './my_embeddings'
}
)
# Option B: PostgreSQL with pgvector (production-ready)
client_pgvector = ImageSearchClient(
api_key="is_your_api_key_here",
base_url="http://localhost:8000",
embedding_storage="pgvector",
storage_config={
'connection_string': 'postgresql://user:password@localhost:5432/my_ecommerce_db',
'table_name': 'product_embeddings'
}
)
# Use local storage for this example
client = client_local
# 2. Connect to existing product database
print("\n2. Connecting to existing product database...")
client.connect_database(
connection_string="sqlite:///my_products.db",
table_name="products",
id_column="product_id",
title_column="product_name",
image_url_column="image_url",
category_column="category",
brand_column="brand",
price_column="price",
rating_column="rating",
enable_auto_embedding=True # Enable automatic embedding generation
)
# 3. Add products with automatic embedding generation
print("\n3. Adding products with automatic embedding generation...")
# Create sample products
products = [
Product(
id="prod_001",
title="Wireless Bluetooth Headphones",
image_url="https://example.com/headphones.jpg",
category="Electronics",
brand="TechBrand",
price=99.99,
rating=4.5,
description="High-quality wireless headphones with noise cancellation"
),
Product(
id="prod_002",
title="Smart Fitness Watch",
image_url="https://example.com/watch.jpg",
category="Electronics",
brand="FitTech",
price=199.99,
rating=4.2,
description="Advanced fitness tracking smartwatch"
),
Product(
id="prod_003",
title="Coffee Maker Pro",
image_url="https://example.com/coffee.jpg",
category="Appliances",
brand="BrewMaster",
price=149.99,
rating=4.7,
description="Professional-grade coffee maker"
)
]
# Add products - embeddings are generated automatically in the background
for product in products:
success = client.add_product_with_auto_embedding(product)
print(f"Added product {product.id}: {'βœ…' if success else '❌'}")
# 4. Wait for embeddings to be generated
print("\n4. Waiting for automatic embedding generation...")
import time
time.sleep(5) # Give time for background processing
# Check embedding generation stats
stats = client.get_embedding_stats()
print(f"Embedding Stats: {stats}")
# 5. Search using local embeddings (fast, no API calls)
print("\n5. Searching using local embeddings...")
# Generate a query embedding (in real usage, this would come from an image or text)
# For demo purposes, we'll use a dummy embedding
import numpy as np
dummy_query_embedding = np.random.rand(768).tolist()
# Search with filters
options = SearchOptions(
top_k=5,
min_price=50.0,
max_price=200.0,
category="Electronics"
)
results = client.search_local(dummy_query_embedding, options)
print(f"Found {len(results)} results:")
for result in results:
print(f" - {result.product.title} (Score: {result.similarity_score:.3f})")
# 6. Update a product (embeddings regenerated automatically)
print("\n6. Updating product (automatic embedding regeneration)...")
updated_product = Product(
id="prod_001",
title="Wireless Bluetooth Headphones Pro", # Updated title
image_url="https://example.com/headphones_pro.jpg", # Updated image
category="Electronics",
brand="TechBrand",
price=129.99, # Updated price
rating=4.8, # Updated rating
description="Premium wireless headphones with advanced noise cancellation"
)
success = client.update_product_with_auto_embedding(updated_product)
print(f"Updated product: {'βœ…' if success else '❌'}")
# 7. Delete a product (embeddings cleaned up automatically)
print("\n7. Deleting product (automatic embedding cleanup)...")
success = client.delete_product_with_auto_embedding("prod_003")
print(f"Deleted product: {'βœ…' if success else '❌'}")
# 8. Integration with existing e-commerce workflow
print("\n8. Integration with existing e-commerce workflow...")
# Example: Customer uploads a product image
def handle_customer_image_upload(image_path: str):
"""Example function showing how to integrate with existing workflow."""
print(f"Customer uploaded image: {image_path}")
# Search for similar products
results = client.search_by_image(image_path, SearchOptions(top_k=3))
print("Similar products found:")
for result in results:
print(f" - {result.product.title} (Similarity: {result.similarity_score:.3f})")
return results
# Example: Customer searches by text
def handle_customer_text_search(query: str):
"""Example function for text-based search."""
print(f"Customer searched: '{query}'")
results = client.search_by_text(query, SearchOptions(top_k=5))
print("Search results:")
for result in results:
print(f" - {result.product.title} (Score: {result.similarity_score:.3f})")
return results
# Simulate customer interactions
print("\nSimulating customer interactions...")
handle_customer_text_search("wireless headphones")
# 9. System monitoring
print("\n9. System monitoring...")
status = client.get_status()
print(f"System Status: {status.is_healthy}")
print(f"Total Products: {status.total_products}")
print(f"Indexed Products: {status.indexed_products}")
print(f"Search Latency: {status.search_latency_ms}ms")
# 10. Cleanup
print("\n10. Cleaning up...")
client.close()
print("βœ… Integration complete!")
def demonstrate_different_storage_options():
"""Demonstrate different storage backend options."""
print("\n" + "="*60)
print("πŸ“Š Storage Backend Options")
print("="*60)
# Local file storage - good for development and small catalogs
print("\n1. Local File Storage:")
print(" - Simple setup, no database required")
print(" - Good for development and small catalogs (<10K products)")
print(" - Files stored in local directory")
# PostgreSQL with pgvector - production ready
print("\n2. PostgreSQL with pgvector:")
print(" - Production-ready, scalable")
print(" - Good for large catalogs (100K+ products)")
print(" - ACID compliance, backup support")
print(" - Advanced querying capabilities")
# Example configurations
print("\n3. Configuration Examples:")
print("\n Local Storage:")
print(" ```python")
print(" client = ImageSearchClient(")
print(" api_key='your_key',")
print(" embedding_storage='local',")
print(" storage_config={'storage_path': './embeddings'}")
print(" )")
print(" ```")
print("\n PostgreSQL Storage:")
print(" ```python")
print(" client = ImageSearchClient(")
print(" api_key='your_key',")
print(" embedding_storage='pgvector',")
print(" storage_config={")
print(" 'connection_string': 'postgresql://user:pass@localhost/db',")
print(" 'table_name': 'product_embeddings'")
print(" }")
print(" )")
print(" ```")
if __name__ == "__main__":
main()
demonstrate_different_storage_options()