Skip to main content

Provides tools to auto generate test data.

Project description

This app aims to provide a simple way of loading masses of randomly generated test data into your development database. You can use a management command to load test data through command line.

It is named autofixture because of the similarity of how I mainly used django’s fixtures. Usually you add test data through the admin to see how your site looks with non static pages. You export data by using dumpdata to send it to your colleagues or to preserve it before you make a manage.py reset app and so on. Your site gets more and more complex and adding test data gets more and more annoying.

This is the usecase where autofixtures should help you to save time that can actually be spent on hacking.

Installation

You must make the autofixture package available on your python path. Either drop it into your project directory or install it from the python package index with pip install django-autofixture. You can also use easy_install django-autofixture if you don’t have pip available.

To use the management command you must add 'autofixture' to the INSTALLED_APPS setting in your django settings file. You don’t need to do this if you want to use the autofixture package only as library.

Management command

The loadtestdata accepts the following syntax:

django-admin.py loadtestdata [options] app.Model:# [app.Model:# ...]

Its nearly self explanatory. Supply names of models, prefixed with its app name. After that, place a colon and tell the command how many objects you want to create. Here is an example how to create three categories and twenty entries for you blogging app:

django-admin.py loadtestdata blog.Category:3 blog.Entry:20

Voila! You have ready to use testing data populated to your database. The model fields are filled with data by producing randomly generated values depending on the type of the field. E.g. text fields are filled with lorem ipsum dummies, date fields are populated with random dates from the last years etc.

There are a few command line options available. Mainly to control the behavior of related fields. If foreingkey or many to many fields should be populated with existing data or if the related models are also generated on the fly. Please have a look at the help page of the command for more information:

django-admin.py help loadtestdata

Using autofixtures as tool for unittests

It has proofed that autofixtures have a great use for unittests. It has always bugged me that creating complex models for testing their behaviour was complicated. Sometimes models have strict restrictions or many related objects which they depend on. One solution would be to use traditional fixtures dumped from your production database. But while in development when database schemes are changing frequently, its hard to maintain all fixtures and to know exactly which objects are contained in the dumps etc…

Autofixtures to the rescue! It lets you automatically generate models and all of their dependecies on the fly. Have a look at the following examples.

Lets start with the very basics. We create an AutoFixture instance for the Entry model and tell it to create ten model instances:

from autofixture import AutoFixture
fixture = AutoFixture(Entry)
entries = fixture.create(10)

Now you can play around and test your blog entries. By default dependecies of foreignkeys and many to many relations are solved by randomly selecting an already existing object of the related model. What if you don’t have one yet? You can provide the generate_fk attribute which allows the autofixture instance to follow foreignkeys by generating new related models:

fixture = AutoFixture(Entry, generate_fk=True)

This generates new instance for all foreignkey fields of Entry. Its possible to limit this behaviour to single fields:

fixture = AutoFixture(Entry, generate_fk=['author'])

This will only create new authors automatically and doesn’t touch other tables. The same is possible with many to many fields. But you need additionally specify how many objects should be created for the m2m relation:

fixture = AutoFixture(Entry, generate_m2m={'categories': (1,3)})

All created entry models get one to three new categories assigned.

Setting custom values for fields

However its often necessary to be sure that a specific field must have a specific value. This is easily achieved with the field_values attribute of AutoFixture:

fixture = AutoFixture(Entry,
    field_values={'pub_date': datetime(2010, 2, 1)})

Custom autofixtures

To have custom autofixtures for your model, you can easily subclass AutoFixture somewhere (e.g. in myapp/autofixtures.py)

from models import MyModel
from autofixture import generators, register, AutoFixture

class MyModelAutoFixture(AutoFixture):
    field_values = {
        'name': generators.StaticGenerator('this_is_my_static_name'),
    }

register(MyModel, MyModelAutoFixture)

Then, loadtestdata will automatically use your custom fixtures.

django-admin.py loadtestdata app.MyModel:10

You can load all autofixtures.py files of your installed apps automatically like you can do with the admin autodiscover. Do so by running autofixture.autodiscover() somewhere in the code, preferably in the urls.py.

More

There is so much more to explore which might be useful for you and your projects:

  • There are ways to register custom AutoFixture subclasses with models that are automatically used when calling loadtestdata on the model.

  • More control for related models, even with relations of related models… (e.g. by using generate_fk=['author', 'author__user'])

  • Custom constraints that are used to ensure that created the models are valid (e.g. unique and unique_together constraints which are already handled by default)

I hope to explain this in the future with more details in a documentation. It will be written but is not finished yet. I wanted to get this project out to support you in development. But since its only python code you can easily study the source on your own and see in which ways it can be used. There are already some parts documented with doc strings which might also be helpful for you.

Future development

The autofixture app is nearly feature complete from the point I wanted to have while starting development. But there is still much room for improvements. One feature you can expect in the future is for example support for multiple databases which was introduced by django 1.2. If you have any ideas or interests to contribute: Feel free to contact me or just start hacking.

Email me (gregor@muellegger.de), contact me on twitter (@gregmuellegger) or fork the git repository on github (git clone git://github.com/gregmuellegger/django-autofixture.git).

Happy autofixturing!

Changelog

0.3.0

  • Adding better support for subclassing AutoFixture through merging of nested Values classes.

  • Renamed attribute and argument none_chance to better matching name empty_p for generators and none_p for AutoFixture.

  • Fixed some issues with management command options. Thanks Mikko Hellsing for his hard work.

  • Fixed issues in unregister(). Thanks Mikko Hellsing for the report.

  • Adding support for FloatField. Thanks to Jyr Gaxiola for the report.

0.2.5

  • Fixing issue with --generate-fk option in management command. Thanks Mikko Hellsing for the report and fix.

0.2.4

  • Using Autofixture.Values for defining initial values in Autofixture subclasses.

  • Making autodiscover more robust. Don’t break if some module can’t be imported or throws any other exception.

0.2.3

  • Fixing bug when a CharField with max_length smaller than 15 is used.

  • AutoFixture.field_values accepts callables as values.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

django-autofixture-0.3.0.tar.gz (20.1 kB view details)

Uploaded Source

File details

Details for the file django-autofixture-0.3.0.tar.gz.

File metadata

File hashes

Hashes for django-autofixture-0.3.0.tar.gz
Algorithm Hash digest
SHA256 7b766182e0b265130380a2da890a4d93c141cfb764d1a604be467bf3681e426d
MD5 b095a51ae56eb4be41993ace00f1b4d0
BLAKE2b-256 15fad69799eeee0a200c08eb3efcfdc85ddd7057ecd1c48d35b4f811c5131da3

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page