Golden Rule

Data goes in the database. No random CSVs/JSON outside data/ directory.

JSON/CSV = raw, unclean, not for production
PostgreSQL = production data storage

DBManager Interface

from db_manager import DBManager, default_db

# Use default connection
df = default_db.fetch_df("SELECT * FROM unified_games WHERE sport = 'nba'")

# Or create custom connection
db = DBManager(connection_string="postgresql://user:pass@host:5432/db")

Key Tables

Table	Purpose
`unified_games`	All games across sports
`game_odds`	Kalshi/sportsbook odds

df = default_db.fetch_df("""
    SELECT * FROM unified_games
    WHERE sport = :sport AND game_date >= :start_date
""", {"sport": "nba", "start_date": "2024-01-01"})

default_db.execute("""
    UPDATE placed_bets SET status = 'won'
    WHERE bet_id = :bet_id
""", {"bet_id": "123"})

df.to_sql("unified_games", default_db.get_engine(),
          if_exists="append", index=False)

default_db.execute("""
    INSERT INTO game_odds (game_id, source, yes_price)
    VALUES (:game_id, :source, :yes_price)
    ON CONFLICT (game_id, source)
    DO UPDATE SET yes_price = EXCLUDED.yes_price
""", params)

from data_validation import validate_nba_data

report = validate_nba_data()
if not report.is_valid:
    raise ValueError(f"Validation failed: {report.errors}")

Database Operations | Skills Pool

Database Operations

Database Operations

Golden Rule

DBManager Interface

Key Tables

Common Operations

Fetch as DataFrame

Execute Query

Insert DataFrame

Upsert Pattern

Connection Defaults

Data Validation

Files to Reference

Postgres Patterns

Postgres Patterns

Database Migrations

Postgres Patterns

Postgres Patterns

Jpa Patterns