dbseed

Auto-generate fake data for SQLAlchemy models — zero config

The Problem

Every project needs test data. You either write tedious factory code or install heavy libraries like Faker. For most cases, you just want 100 rows of realistic-looking data.

The Solution

pip install dbseed

from dbseed import seed

users = seed(User, SeedConfig(count=100))
session.add_all(users)
await session.commit()
# Done. 100 users with realistic names, emails, ages.

dbseed inspects your SQLAlchemy model and automatically generates the right kind of data for each column.

Smart Field Detection

Override Specific Fields

from dbseed import seed, SeedConfig

users = seed(User, SeedConfig(
    count=50,
    overrides={
        "role": "admin",                    # static value
        "age": lambda: random.randint(18, 65),  # dynamic callable
    },
))

Features

• Inspects SQLAlchemy model columns automatically
• Smart name-based field detection (email, name, url, etc.)
• Type-based fallback for unknown fields
• Field overrides (static values or callables)
• Skips primary keys and server-default columns
• No external dependencies beyond SQLAlchemy
• Works with SQLAlchemy 2.0 mapped columns

License

MIT