itinerary

A tool to apply pure SQL migrations to PostgreSQL databases at runtime.

Itinerary uses PostgreSQL’s advisory locks feature to prevent multiple instances from running migrations simultaneously. As such, it’s suitable for use in horizontally scalable applications.

Installation

pip install itinerary

Usage

Put raw SQL files in your migration directory, e.g.:

migrations/0001_create_table.sql
migrations/0002_add_column.sql
migrations/0003_create_index.sql

Then run the auto_migrate function before your application starts.

Flask example

import itinerary
from flask import Flask
app = Flask(__name__)

if __name__ == '__main__':
    conn = psycopg2.connect(host='localhost')
    itinerary.auto_migrate(conn)
    app.run()

uWSGI example

import itinerary
from uwsgidecorators import postfork

@postfork
def migrate_db():
    conn = psycopg2.connect(host='localhost')
    itinerary.auto_migrate(conn)

Configuration

itinerary.auto_migrate(connection, path='migrations',
                       version_table='_version', lock_id=0)

• path – Directory to scan for migration files.
• version_table – Table to use to keep track of applied migrations. If it doesn’t exist, itinerary will create it.
• lock_id – The key to use when acquiring and releasing a PostgreSQL advisory lock.