backend/check_db_clustering.py

import os
from supabase import create_client, Client
from dotenv import load_dotenv

load_dotenv()

url = os.environ.get("SUPABASE_URL")
key = os.environ.get("SUPABASE_SERVICE_ROLE_KEY") or os.environ.get("SUPABASE_KEY")
client: Client = create_client(url, key)

def check_clustering_status():
    print("Checking profiles table for cluster labels...")
    resp = client.table("profiles").select("id, cluster_label").limit(20).execute()
    data = resp.data
    
    if not data:
        print("No profiles found.")
        return

    # Count how many have labels
    labeled = [d for d in data if d.get("cluster_label")]
    print(f"Sample size: {len(data)}")
    print(f"Profiles with cluster_label: {len(labeled)}")
    
    if labeled:
        print("Sample labels:")
        for d in labeled[:5]:
            print(f" - {d['id']}: {d['cluster_label']}")
    else:
        print("No profiles have cluster labels in this sample.")

if __name__ == "__main__":
    check_clustering_status()