Port of DDF for algorithmically creating an initial dataset
Project description
At a glance
To run the command: python manage.py populate_db
*You have to run the command on a completely empty database with migrations applied, else it will crash.*
Description
Applys a simple, guided DFS technique to traverse a schema, and create a concrete / compact database. The meat of the population engine is a modified version of django dynamic fixture by Paulo Cheque.
The populate_db management command iterates through all models recognized by the ORM, attempting to save each instance to the database. If the instance has a non-empty chain of dependencies, the code fills out the dependencies before saving the instance. Our code differs from ddf in that it stores all previously created objects. When dependencies are filled out, previously created objects are used to fill the chain unless otherwise specified.
The following three lists of classes are used to alter the behavior of the command. They should be set in your app’s settings.py. Note that all classes must be given in their fully qualified class name (with all modules and submodules).
IGNORE
The IGNORE list specifies classes that will not be used in the top level constructor function. Most of the classes currently in the IGNORE list are ones which require special fixture data to be saved. The app works okay without them or auto-generates these, usually.
ALLOW_MULTIPLE
The allow multiple specifies classes that will be created as though they are not already in the list of created instances. I.e. a profile may be dependent on a user, but it may be nonsensical for all profiles to be associated to a single user
currently anything that is on the receiving end of a one_to_one will need to be in ALLOW_MULTIPLE for proper behavior.
REQUIREMENTS
Sometimes a web app may require a subgraph with specific properties. The requirements list allows you to specify a subgraph in a json-like format. Requirements are stored as a list of (possibly) nested dictionaries, keyed by the field name (edge of the graph) that connects nodes with specific properties.
A useful example imported from our configuration:
REQUIREMENTS = [
# Examples of chained dependencies. This foo.bar.Baz selects a 'foo.Quux' with 'daedalus'
# in the 'foo.Quux' field named 'name'
# through its field with name 'qux'
{'foo.bar.Baz': {'qux': {'foo.Quux': {'name': 'daedalus'}}}},
{'foo.bar.Baz': {'qux': {'foo.Quux': {'name': 'tiresias'}}}},
{'corge.Waldo': {'qux': {'foo.Quux': {'name': 'daedalus'}}, 'is_prophet': False}},
{'corge.Waldo': {'qux': {'foo.Quux': {'name': 'tiresias'}}, 'is_prophet': True}},
# Objects with specific properties may be declared without necessarily chaining
{'django.contrib.auth.User': {'email': EMAIL_1}},
{'django.contrib.auth.User': {'email': EMAIL_2}}
]
