Load YAML data fixtures for SQLAlchemy
This package allows you to define some data in YAML and load them into a DB. The yaml data should correspond to SQLAlchemy declarative mappers.
Example:
- User:
- __key__: joey
username: joey
email: joey@example.com
profile:
name: Jeffrey
- __key__: dee
username: deedee
email: deedee@example.com
- Profile:
- user: dee
name: Douglas
- Group:
- name: Ramones
members: [joey.profile, dee.profile]
- The root of YAML contains a sequence of
mapper names
e.g.- User
,- Profile
etc - The order of these names should follow relationship dependencies
- Every name should contain a sequence of instances
- Each instance is a mapping of attribute -> value
- the attributes are taken from the mapper
__init__()
(usually an attributes maps to a column) - The special field
__key__
can be used to identify this instnace in a relationship reference e.g. TheProfile.user
- Note that any
__key__
MUST be globally unique - In a to-one relationship the data can be directly nested in the parent data definition
- References can access attributes using a dot notation, e.g.
joey.profile
- to-many relationships can be added as a list of references
The mapper definition for this example is in the test file.
pip install sqla-yaml-fixtures
This module expose a single function
load(ModelBase, session, fixture_text, loader=None)
Where:
ModelBase
is SQLAlchemy declarative basesession
is SQLAlchemy sessionfixture_text
is a string containg the YAML fixtures
from sqlalchemy import create_engine
from sqlalchemy import Column, Integer, String
from sqlalchemy.ext.declarative import declarative_base
from sqlalchemy.orm import Session
import sqla_yaml_fixtures
BaseModel = declarative_base()
class User(BaseModel):
__tablename__ = 'user'
id = Column(Integer, primary_key=True)
username = Column(String(150), nullable=False, unique=True)
email = Column(String(254), unique=True)
def main():
engine = create_engine('sqlite://')
BaseModel.metadata.create_all(engine)
connection = engine.connect()
session = Session(bind=connection)
fixture = """
- User:
- username: deedee
email: deedee@example.com
- username: joey
email: joey@example.commit
"""
sqla_yaml_fixtures.load(BaseModel, session, fixture)
print('\n'.join(u.username for u in session.query(User).all()))
if __name__ == '__main__':
main()
Note: the load() function performs a session.commit().
load() returns an instance of Store. Using this object get() method you can passing a key as argument you get a reference to the object added into the database. This is useful to easily get attributes that are generated by the database.
store = sqla_yaml_fixtures.load(BaseModel, session, fixture)
my_obj = store.get('dee')
print('Created object id: {}'.format(my_obj.id))
Warning
By default YAML is loaded using yaml.FullLoader, this is insecure when loading unstrusted input. It is possible to overwrite the loaded by setting loader param in the load() function.
For basic usage there is also command line. Example:
$ python -m sqla_yaml_fixtures --db-url sqlite:///dev.db --db-base mypkg.models:Base --reset-db --alembic-stamp fixture.yaml
All available options:
$ python -m sqla_yaml_fixtures --help usage: sqla_yaml_fixtures [-h] --db-base DB_BASE --db-url DB_URL [--yes] [--reset-db] [--alembic-stamp] [--jinja2] FILE [FILE ...] load fixtures from yaml file into DB positional arguments: FILE YAML file with DB fixtures optional arguments: -h, --help show this help message and exit --db-base DB_BASE SQLAlchemy Base class with schema metadata in the format my_package.my_module:MyClass --db-url DB_URL Database URL in the format dialect+driver://username:password@host:port/database --yes Do NOT ask for confirmation before applying fixtures --reset-db Drop DB schema and data and re-create schema before loading fixtures --alembic-stamp Perform `alembic stamp head` --jinja2 load fixture files as jinja2 templates