H&M Fashion Search with LanceDB

A modern fashion search application built with FastAPI and LanceDB, featuring semantic search capabilities for H&M fashion items.

Features

• Semantic Search: Find fashion items using natural language queries
• Filtering: Filter by product groups and item types
• Real-time Results: Fast search results with pagination
• Modern UI: Clean, responsive interface with product details modal
• Vector Database: Powered by LanceDB for efficient similarity search

Technology Stack

• Backend: FastAPI (Python)
• Vector Database: LanceDB
• Embeddings: Sentence Transformers (CLIP ViT-B-32)
• Frontend: HTML, CSS, JavaScript
• AI Models: Hugging Face Transformers

Installation

1. Clone the repository:

git clone <repository-url>
cd lancedb-ecommerce

2. Install dependencies:

pip install -r requirements.txt

3. Create the data directory:

mkdir -p data

Usage

Starting the Application

1. Run the FastAPI server:

uvicorn app:app --reload --host 0.0.0.0 --port 8000

2. Open your browser and navigate to http://localhost:8000/static/index.html

Adding Data to LanceDB

The application will automatically create an empty table when it starts. To populate it with fashion data:

1. Prepare your data in a format compatible with the schema:

import pandas as pd
import lancedb
from sentence_transformers import SentenceTransformer

# Connect to the database
db = lancedb.connect("./data")
table = db.open_table("h&m-mini")

# Load your fashion data
data = pd.read_csv("your_fashion_data.csv")

# Generate embeddings for text fields (e.g., product descriptions)
encoder = SentenceTransformer("clip-ViT-B-32")
data['vector'] = data['detail_desc'].apply(lambda x: encoder.encode(x).tolist())

# Insert data into the table
table.add(data)

2. The expected schema includes:
   • image_url: URL to product image
   • prod_name: Product name
   • detail_desc: Product description
   • product_type_name: Type of product
   • index_group_name: Product category group
   • price: Product price
   • article_id: Unique article identifier
   • available: Availability status
   • color: Product color
   • size: Product size
   • vector: 512-dimensional embedding vector

API Endpoints

• GET /search: Search for fashion items with optional filters
• GET /groups: Get available product groups
• GET /static/index.html: Main application interface

Configuration

Edit the Config class in app.py to customize:

• Database path
• Table name
• Embedding model
• Product group ordering

Differences from Qdrant Version

This LanceDB version offers several advantages:

• Local Storage: No need for external database server
• Simplified Setup: Automatic table creation and management
• Better Performance: Optimized for local development and testing
• Easier Deployment: Single file database that can be easily shared

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Submit a pull request

License

This project is licensed under the MIT License.