Skip to content

Supabase Migration Summary

Overview

The application has been updated to use Supabase's REST API instead of direct PostgreSQL connections. This provides better reliability and easier access to your database.

What Changed

1. Dependencies Added

  • supabase - Official Python client for Supabase

2. Database Module Updated (trade_calc/database.py)

All database functions now use the Supabase client instead of psycopg2:

  • get_supabase_client() - New function to create Supabase client with automatic authentication
  • get_db_connection() - Now returns Supabase client (backwards compatible)
  • fetch_price_history() - Uses Supabase REST API
  • get_current_price_from_db() - Uses Supabase REST API
  • get_available_tickers() - Uses Supabase REST API

Important: The client automatically signs in with your username/password if provided in .env, which is required for tables with Row Level Security (RLS) enabled.

3. Scripts Updated

  • scripts/next_friday_calc.py - Updated to use Supabase client

4. Environment Variables

Required credentials in .env: 

SUPABASE_URL=http://192.168.1.147:8000
SUPABASE_ANON_KEY=your-anon-key
SUPABASE_USERNAME=your-email@example.com
SUPABASE_PASSWORD=your-password

Benefits

  1. More Reliable - No more "server closed connection" errors
  2. RESTful API - Uses HTTP/REST instead of PostgreSQL wire protocol
  3. Better Error Handling - Clearer error messages
  4. Automatic Connection Pooling - Supabase handles connection management

Available Tables

Currently available in your Supabase instance: - ✅ earnings_calendar - Earnings report dates for symbols

Not yet available (may need to be created): - ❌ tcs_ohlc_daily - Historical OHLC price data

Testing

The script was successfully tested: 

uv run scripts/next_friday_calc.py

Result: ✅ Connected successfully to Supabase and queried earnings_calendar

Note: Initial connection returned 0 results because Row Level Security (RLS) was enabled on the table. After adding authentication with username/password, the script successfully found 1,442 symbols with earnings in the specified date range.

Backwards Compatibility

The old get_db_connection() function still exists but now returns a Supabase client instead of a psycopg2 connection. All code has been updated to use the new client API.

Next Steps

If you need historical price data (OHLC), you'll need to: 1. Create the tcs_ohlc_daily table in Supabase 2. Populate it with historical data 3. Ensure the table is exposed through Supabase's REST API

For now, the application will fall back to using yfinance for fetching current prices and options data.