A utility for query-like predicate testing of Django model instances
Python
Latest commit 155f60c Jun 9, 2016 @lucaswiman lucaswiman committed on GitHub Merge pull request #35 from django-predicate/django-1.10
Django version updates
Permalink
Failed to load latest commit information.
predicate Remove pivot on version Jun 8, 2016
tests Drop django 1.7 support and add 1.10a1 to the build matrix Jun 8, 2016
.gitignore Update ignored coverage Dec 24, 2015
.travis.yml
AUTHORS Update authors to reflect recent pull requests Apr 9, 2016
CONTRIBUTING.rst
LICENSE initial commit Aug 15, 2012
MANIFEST.in
README.rst Fix examples so they are correct Feb 6, 2016
requirements-test.txt Drop django 1.7 support and add 1.10a1 to the build matrix Jun 8, 2016
runtests.py Add several extra kinds of models and associated tests. Dec 16, 2015
setup.py Drop python 2.6 support from metadata Apr 9, 2016
tox.ini Drop django 1.7 support and add 1.10a1 to the build matrix Jun 8, 2016

README.rst

django-predicate

django-predicate provides a Q like object to facilitate the question: "would this model instance be part of a query" but without running the query or even saving the object.

Quickstart

Install django-predicate:

pip install django-predicate

Then use the P object just as you would Q objects:

from predicate import P

p = P(some_field__startswith="hello", age__gt=20)

You can then call the eval method with a model instance to check whether it passes the conditions:

model_instance = MyModel(some_field="hello there", age=21)
other_model_instance = MyModel(some_field="hello there", age=10)
p.eval(model_instance)
>>> True
p.eval(other_model_instance)
>>> False

or you can use Python's in operator.

model_instance in p
>>> True

Even though a predicate is not a true container class - it can be used as (and was designed as being) a virtual "set" of objects that meets some condiiton.

Like Q objects, P objects can be &'ed and |'ed together to form more complex logic groupings.

In fact, P objects are actually a subclass of Q objects, so you can use them in queryset filter statements:

qs = MyModel.objects.filter(p)

P objects also support QuerySet-like filtering operations that can be applied to an arbitrary iterable: P.get(iterable), P.filter(iterable), and P.exclude(iterable):

model_instance = MyModel(some_field="hello there", age=21)
other_model_instance = MyModel(some_field="hello there", age=10)
p.filter([model_instance, other_model_instance]) == [model_instance]
>>> True
p.get([model_instance, other_model_instance]) == model_instance
>>> True
p.exclude([model_instance, other_model_instance]) == [other_model_instance]
>>> True

If you have a situation where you want to use querysets and predicates based on the same conditions, it is far better to start with the predicate. Because of the way querysets assume a SQL context, it is non-trivial to reverse engineer them back into a predicate. However as seen above, it is very straightforward to create a queryset based on a predicate.