django-konst
is a utility for Django that makes the definition, use and storage of both integer and string based constants easy and readable. It avoids passing of constants
into template contexts and makes evaluation concise and readable.
It also makes exposure of these constants via forms and DRF serializers simple.
Constants can be defined with friendly names, backed by either integers or text.
from konst import Constant, Constants
# states backed by integers
states = Constants(
Constant(pending=0),
Constant(active=1),
Constant(inactive=2)
)
# and another set of constants backed by strings
colours = Constants(
Constant(red="FF0000"),
Constant(green="00FF00"),
Constant(yellow="FFFF80"),
Constant(white="FFFFFF")
)
At times, it will be necessary to group constants and test membership within that group.
To achieve this, django-konst
provides a ConstantGroup
class.
from konst import Constant, ConstantGroup
# states backed by integers
states = Constants(
Constant(active=0),
Constant(cancelled_ontime=1),
Constant(cancelled_late=2),
ConstantGroup(
"cancelled",
("cancelled_ontime", "cancelled_late")
)
)
While not strictly necessary, it is advisable to effectively namespace your constants by defining them in the scope of a model definition. This means you have your constants wherever you have the model class, as well as any model instance.
from django.db import models
from django.utils.translation import ugettext_lazy as _
from konst import Constant, ConstantGroup, Constants
from konst.models.fields import (
ConstantChoiceCharField,
ConstantChoiceField
)
class Apple(models.Model):
purposes = Constants(
Constant(cooking=0, label=_("Cook me!")),
Constant(eating=1, label=_("Eat me!")),
Constant(juicing=2, label=_("Juice me!")),
Constant(ornamental=3, label=_("Just look how pretty I am!")),
ConstantGroup(
"culinary", ("cooking", "eating", "juicing")
)
)
colours = Constants(
Constant(red="FF0000", label=_("red")),
Constant(green="00FF00", label=_("green")),
Constant(yellow="FFFF80", label=_("yellow")),
Constant(white="FFFFFF", label=_("white")),
)
name = models.CharField(max_length=30)
purpose = ConstantChoiceField(constants=purposes)
colour = ConstantChoiceCharField(constants=colours, max_length=30)
The entire point of this library is to make the use of constants defined in this way easy and concise.
apple = Apple.objects.get(name='Granny Smith')
apple.purpose.cooking
True
apple.colour.red
False
apple.colour.green
True
# we don't care about the specific purpose, just whether it is as food
# or not, so use the ConstantGroup!
apple.purpose.culinary
True
{% if apple.purpose.eating %}
You should bite this {{ apple.name }}!
{% endif %}
red_apples = Apple.objects.filter(colour=Apple.colours.red)
culinary_apples = Apple.objects.filter(
purpose__in=Apple.purposes.culinary
)
from django import forms
from konst.forms.fields import ConstantChoiceField
class SomeForm(forms.Form):
purpose = ConstantChoiceField(constants=Apple.purposes)
Using the konst.extras.drf.fields.ConstantChoiceField
serializer field with the
Django Rest Framework it is possible to both output and receive constant values.
from konst.extras.drf.fields import ConstantChoiceField
from rest_framework import serializers
class AppleSerializer(serializers.ModelSerializer):
purpose = ConstantChoiceField(Apple.purposes)
colour = ConstantChoiceField(Apple.colours)
class Meta:
model = Apple
fields = (
"name", "purpose", "colour"
)
# let's see how it handles bad values
serializer = AppleSerializer(
data={
"name": "Fuji",
"colour": "blue",
"purpose": "dicing"
}
)
serializer.is_valid()
False
serializer.errors
{
'colour': [u'"blue" is not a valid choice.'],
'purpose': [u'"dicing" is not a valid choice.']
}
# and now how it handles some good values
serializer = AppleSerializer(
data={
"name": "Fuji",
"colour": "red",
"purpose": "eating"
}
)
serializer.is_valid()
True
# let's create a database entry!
instance = serializer.save()
# and now our instance can be interacted with neatly
instance.colour.red
True
# finally, let's see how this looks when rendering JSON
AppleSerializer(instance=instance).data
{
"name": "Fuji",
"colour": "red",
"purpose": "eating"
}
When the Django ORM instantiates a model instance from the database, it will ensure that any ConstantChoiceField
or ConstantChoiceCharField
fields have values set to Constant
instances.
If you create or modify a model instance using a raw, hard-coded constant value, you're likely to hit errors along the lines of "AttributeError: 'int' object has no attribute 'attribute-name'".
Take the example below where the caller sets the colour of a model instance to the underlying value for the Constant
:
apple = Apple.objects.get(name='Granny Smith')
apple.purpose = 1 # constant for `eating` .. I know, who on earth would eat a granny smith straight up?!
apple.save()
Django will happily persist the data correctly, but let's say you have a post_save
signal (or any other code, really) that does something further with the instance:
from django.db.models.signals import post_save
from django.dispatch import receiver
from .models import Apple
@receiver(post_save, sender=Apple)
def apple_updated_receiver(sender, instance, created, *args, **kwargs):
if not created and instance.purpose.eating:
print(f"Oh my, {instance.name} is now for eating!")
This code will sadly raise AttributeError: 'int' object has no attribute 'purpose'
, as instance.purpose
is just the integer we set it to before saving.
The good news is that you can avoid this by always setting django-konst
fields on instances to Constant
instances such that downstream code can happily handle the instance as if it'd come straight out of the database:
apple = Apple.objects.get(name='Granny Smith')
apple.purpose = Apple.purposes.eating
apple.save()
Regardless of whether you're using django-konst
for your constants, it's good practice to not hard-code constant values in order to avoid subtle mistakes and to ease changes in the future.
django-konst
supports a variety of Python and Django versions. It's best if you test each one of these before committing. Our Circle CI Integration will test these when you push but knowing before you commit prevents from having to do a lot of extra commits to get the build to pass.
In order to easily test on all these Pythons and run the exact same thing that CI will execute you'll want to setup pyenv and install the Python versions outlined in tox.ini.
If you are on Mac OS X, it's recommended you use brew. After installing brew
run:
brew install pyenv pyenv-virtualenv pyenv-virtualenvwrapper
Next, install the various python versions we want to test against and create a virtualenv specifically for django-konst
:
pyenv install 3.6.10
pyenv install 3.7.6
pyenv install 3.8.1
pyenv virtualenv 3.8.1 konst
pyenv activate konst
pip install detox
pyenv shell konst 3.6.10 3.7.6
Now ensure the konst
virtualenv is activated, make the other python versions also on our path, and run the tests!
pyenv shell konst 3.6.10 3.7.6
detox
This will execute the test environments in parallel as defined in the tox.ini
.