a-ledger

Embedded Python accounting SDK for a-options and a-trader. It owns balanced journal facts, correction history, lot computation, reservations and batch reads. Applications own business data and compose it through explicit ledger identifiers.

Use

from sqlalchemy import create_engine, event
from a_ledger import Ledger
from a_ledger.schema import install

engine = create_engine("sqlite:///accounts.db")

@event.listens_for(engine, "connect")
def configure_sqlite(dbapi_connection, _):
    dbapi_connection.isolation_level = None
    dbapi_connection.execute("PRAGMA foreign_keys=ON")

@event.listens_for(engine, "begin")
def begin_sqlite(connection):
    connection.exec_driver_sql("BEGIN")

with engine.begin() as connection:
    install(connection)  # Explicit installation on an empty/current store.
    ledger = Ledger(connection)
    ledger.accounts.create_portfolio("main", code="main", currency="CNY")
    ledger.accounts.create_account(
        "cash", portfolio_id="main", code="cash", category="ASSET"
    )

The caller owns the connection and outer transaction. Use the same connection for application writes and SDK calls; the SDK never commits that transaction. All posting, replacement and reversal commands use ledger.journal.

• Public contracts
• Data model and business references
• Ledger page queries
• Whole-source option booking
• Domain terminology

Verification

Run .venv/bin/python -m pytest. Source and test modules are limited to 500 physical lines. Tests use synthetic data and public accounting interfaces. Normal startup and restore support the current schema; historical audit evidence and command identities remain immutable.