lancedb/ecommerce

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