`pfg gen seed beanie`

Capabilities

pfg gen seed beanie generates deterministic documents and inserts them into a MongoDB database using Beanie. It mirrors every planning option from the SQLModel variant but swaps in Beanie-specific behaviors such as cleanup mode and Mongo-safe URL allow-lists.

Typical use cases

Seed a local MongoDB/Atlas instance with fixture data before manual testing.
Populate QA clusters using deterministic seeds so smoke tests stay reproducible.
Run cleanup mode to insert documents for validation and then delete them automatically.

Inputs & outputs

Target: module path or --schema for JSON Schema ingestion.
Database: Mongo-style URI (for example mongodb://localhost:27017/app). Fixturegen refuses to connect unless the scheme matches --allow-url (defaults mongodb://, mongomock://).
Result: logs the number of documents inserted, whether cleanup took place, and whether the run was dry.

Flag reference

Shared planning options

--n/-n, --include, --exclude, --seed, --now, --preset, --profile, --freeze-seeds, --freeze-seeds-file, --link, --with-related, --max-depth, --on-cycle, --rng-mode, --respect-validators, --validator-max-retries, --schema — identical to the SQLModel command.

General generation flags (include/exclude, --seed, --preset, --profile, --freeze-seeds, --rng-mode, --link, --with-related, etc.) mirror pfg gen json. You can also override locales on demand via --locale or repeatable --locale-map pattern=locale so seeded documents match regional Faker data without editing config files.

Beanie-specific controls

--database/-d: Mongo connection string (can include credentials and query parameters).
--allow-url: repeatable whitelist of allowed URI prefixes. Extend this when targeting Atlas or SRV URIs.
--batch-size: number of documents per insertion chunk (default 50).
--cleanup/--keep: when enabled, wraps each insert in a matching delete so the database returns to its prior state after validation. Useful for doctor-style audits.
--dry-run: log generated payloads without talking to MongoDB.

Example workflows

Seed a local MongoDB instance

pfg gen seed beanie ./app/models.py \

  --database mongodb://localhost:27017/app \

  --n 250 --include app.models.Order \

  --seed 42 --preset realistic --with-related app.models.User

Inserts 250 `Order` documents (each with a related `User`) into the `app` database.

Sample output

[beanie_connect] url=mongodb://localhost:27017/app cleanup=False dry_run=False

Inserted documents: app.models.Order=250, related={'app.models.User':250}

Mongo preview (db.order.findOne({}, {_id:0}))

  "id": "be0773e6-ecbb-46e5-93d7-8db3cb8e9c8a",

  "total_cents": 4599,

  "user_id": "6ad0ab66-6c07-42c0-9e86-5b9292e70ac4",

  "status": "PENDING"

Cleanup mode for integration verification

pfg gen seed beanie ./app/models.py \

  --database mongodb://localhost:27017/app \

  --n 10 --cleanup --dry-run

Generates documents, logs them, and skips writes because of `--dry-run`. Drop `--dry-run` to perform insert/delete cycles automatically.

Sample output

[beanie_connect] url=mongodb://localhost:27017/app cleanup=True dry_run=True

Cleanup enabled; generated payloads discarded (0 writes executed)

In-memory `mongomock` quickstart

The repo ships runnable Beanie models at docs/examples/beanie_models.py. Copy them into your demo project and point fixturegen at mongomock://... to avoid running a real Mongo instance:

cp path/to/pydantic-fixturegen/docs/examples/beanie_models.py ./beanie_models.py

pfg gen seed beanie ./beanie_models.py \

  --database mongomock://localhost/pfg-demo \

  --include beanie_models.DemoPurchase \

  --with-related beanie_models.DemoCustomer \

  --n 10 --cleanup

`mongomock_motor` keeps everything in memory, so the command is safe in CI and during local smoke tests. Pair it with `--cleanup` when you want to validate generation without persisting documents between runs.

Operational notes

Fixturegen extracts the database name from the Mongo URI so Beanie can target the correct database even when SRV URIs are used.
Cleanup mode is invaluable when you want to observe generated payloads via MongoDB triggers without mutating long-lived data sets.
The CLI suppresses deprecated model_fields warnings emitted by Beanie with Pydantic v2 so logs stay clean.

Edit this page

pfg gen seed beanie