postseed (Python)

Realistic Postgres seed data. In seconds. Postseed generates seed data that respects your foreign keys, enums and uniqueness, never leaves your machine, and is deterministic: the same schema, seed and version give byte-identical output.

This package bundles the Postseed engine and runs it with Node.js 20 or newer (it must be on your PATH, or set POSTSEED_NODE). Nothing else to install.

pip install postseed
postseed            # opens Studio, the local web app, on 127.0.0.1

Command line

postseed generate --schema schema.sql --rows 50 --seed 42 --out seed.sql
postseed insert --url "$DATABASE_URL" --dry-run     # validate against your database, then roll back
postseed insert --url "$DATABASE_URL" --yes         # really insert, in one transaction

See the documentation for every option.

From Python

import postseed

sql = postseed.generate(
    """
    CREATE TABLE users (id serial PRIMARY KEY, email text UNIQUE NOT NULL);
    CREATE TABLE posts (id serial PRIMARY KEY, user_id int NOT NULL REFERENCES users(id), title text);
    """,
    rows=25,
    seed=42,
)
# `sql` is a transactional script: every post points at a real user, every email is unique.

generate(schema_sql, *, rows=50, seed=42, format="sql", tables=None, table_rows=None, null_ratio=None, now=None) returns the text (format is "sql" or "json") and raises postseed.PostseedError with the engine's message if it cannot. The same arguments always return the same text.

postseed.run(args, input=None) runs any CLI command and returns a subprocess.CompletedProcess.

Notes

• Python 3.8 or newer. The package is pure Python plus a bundled JavaScript engine, so there is one wheel for every platform.
• It is the same engine as the postseed npm package and the postseed gem, so output is identical across all three.
• MIT licensed.