I'm working on an e-commerce framework for Django. The chief design goal is to provide the bare minimum functionality in terms of models and view, instead allowing the users of the library to extend or replace the components with their own.
The reasoning for this is that trying to develop a one-size-fits-all solution to e-commerce leads to overcomplicated code which is often far from optimal.
One approach to tackling this seems to be using inversion-of-control, either through Django's settings file or import hacks, but I've come up against a bit of a problem due to how Django registers its models.
The e-commerce framework provides a bunch of abstract models, as well as concrete versions in {app_label}/models.py. Views make use of Django's get_model(app_label,model) function to return the model class without having to hard-code the reference.
This approach has some problems:
Users have to mimic the structure of the framework's apps, ie the app_label and effectively replace our version of the app with their own
Because of the way the admin site works by looking for admin.py in each installed app, they have to mimic or explicitly import the framework's admin classes in order to use them. But by importing them, the register method gets called so they have to be unregistered if a user wants to customise them.
The user has to be extremely careful about how they import concrete models from the core framework. This is because Django's base model metaclass automatically registers a model with the app cache as soon as the class definition is read (ie upon __new__), and the first model registered with a specific label is the one you're stuck with. So you have to define all your override models BEFORE you import any of the core models. This means you end up with messy situations of having a bunch of imports at the bottom of your modules rather than the top.
My thinking is to go further down the inversion-of-control rabbit hole:
All references to core components (models, views, admin, etc) replaced with calls to an IoC container
For all the core (e-commerce framework) models, replace the use of Django's base model metaclass with one that doesn't automatically register the models, then have the container explicitly register them on startup.
My question:
Is there a better way to solve this problem? The goal is to make it easy to customise the framework and override functionality without having to learn lots of annoying tricks. The key seems to be with models and the admin site.
I appreciate that using an IoC container isn't a common pattern in the Django world, so I want to avoid it if possible, but it is seeming like the right solution.
Did you look at the code from other projects with a similar approach?
Not sure if this way covers your needs, but imo the code of django-shop is worth to look at.
This framework provides the basic logic, allowing you to provide custom logic where needed.
customize via models
eg see the productmodel.py
#==============================================================================
# Extensibility
#==============================================================================
PRODUCT_MODEL = getattr(settings, 'SHOP_PRODUCT_MODEL',
'shop.models.defaults.product.Product')
Product = load_class(PRODUCT_MODEL, 'SHOP_PRODUCT_MODEL')
customize via logic/urls
eg see the shop's simplevariation-plugin
It extends the cart-logic, so it hooks in via urlpattern:
(r'^shop/cart/', include(simplevariations_urls)),
(r'^shop/', include(shop_urls)),
and extends the views:
...
from shop.views.cart import CartDetails
class SimplevariationCartDetails(CartDetails):
"""Cart view that answers GET and POSTS request."""
...
The framework provides several points to hook-in, the simplevariation-plugin mentionned above additionally provides a cart-modifier:
SHOP_CART_MODIFIERS = [
...
'shop_simplevariations.cart_modifier.ProductOptionsModifier',
...
]
I worry that this explanation is not very understandable, it is difficult to briefly summarize this concept. But take a look at the django-shop project and some of its extensions: ecosystem
Related
I am starting a web application in pure Django. However, in the future, there might be a requirement for REST api. If it happens, the most obvious choice will be Django REST framework.
Both the "old-fashioned" and REST parts share the models, however, the views are slightly different (permissions definitions, for example) and forms are replaced with serializers. Doing it the most obvious way would mean to duplicate the application logic several times, and thus a failure to follow DRY principle, and so the code becomes unmaintainable.
I got an idea to write all the logic into models (since they are shared), but in such case, there will be no use of permission mixins, generic views and the code would not be among the nicest ones.
Now I ran out of ideas. What is the best practice here?
I'd try to keep things simple as you're not sure about the future requirements for the API, and guessing can introduce extra complexity that may not even be needed when requirements will be clear.
Both Django forms and Rest Framework serializers already offer you a declarative approach that abstracts away the boilerplate code needed for basic stuff, which normally accounts for most of your code anyway.
For example, one of your Django form could look like this:
class ArticleForm(ModelForm):
class Meta:
model = Article
fields = ['title', 'content']
And in the future the DRS serializer would be:
class ArticleSerializer(ModelSerializer):
class Meta:
model = Article
fields = ['title', 'content']
As you can see, if you try and stick to ModelForm and ModelSerializer, there won't be much duplication anyway. You can also simply store the fields list in a variable and just reuse that.
For more custom things, you can start by sharing logic into simple functions, for example:
def save_article_with_author(article_data, author_data):
# custom data manipulation before saving, consider that article_data will be a dictionary either if it comes from deserialized JSON (api) or POST data
# send email, whatever
This function can be shared between your form and serializer.
For everything related to data fetching, I'd try to use Model Managers as much as possible, defining custom querysets that can be resued e.g. for options by forms and serializers.
I tend to avoid writing any logic that doesn't directly read or write data into the model classes. I think that couples too much the business logic with the data layer. As an example, I never want to write any auth/permission checks into a save() method of a model, because that couples different layers too tightly.
As a rule of thumb, imagine this scenario: you add say permissions checks or the logic to send an email when a user is created overriding the save() method of your Article model.
Then, later on you're asked to write a simple manage command that batch-import users from a spreadsheet. At this point, what you did in your save() method really gets in the way, as you can freely access your data through your model without having to bother with permissions, emails and all of that.
Regarding the view layer and assuming you need to implement some shared auth/permission checks and you don't want to have separate views, you can use this approach:
https://www.django-rest-framework.org/topics/html-and-forms/
Blockquote
REST framework is suitable for returning both API style responses, and regular HTML pages. Additionally, serializers can be used as HTML forms and rendered in templates.
Here's some guidelines on how you could dynamically switch from HTML to JSON based to the request content type:
https://www.django-rest-framework.org/api-guide/renderers/#advanced-renderer-usage
This seems like a good option in your situation, I'd just write down a quick proof-of-concept before you go all in to see if you are not too limited for what you need to do.
I have a general algorithm design question. I am creating a Django app that will connect to an API, but I won't be storing these results (at least not at first). After I retrieve the data from the API I manipulate it accordingly and have already created a class with numerous methods to do this.
Should the programming logic for this be performed in the model or the view for the Django framework? Is one more sustainable than the other (e.g. in a few months I decide to store the information). Also, is it best to encapsulate my class in its own file, and them import into the model/view?
Thanks!
Rob
If you don't need to store data, don't use Djnago's built in models.
Write views and import your own modules/classes.
Bonus: If your views share a lot of logic (probably related to request/response handling), use class based views and write a common MyBaseView.
I'm learning Django and I've finished 2 tutorials - official and amazing tutorial called Tango With Django. Though, I got everything I need to work, I have one question:
In Tango with Django aren't used class-based views - only links to the official tutorial.
Why didn't they include this information?
When should we use class-based views and is it a good practice?
Class based views allow you to structure your views and reuse code by harnessing inheritance and mixins.
https://docs.djangoproject.com/en/1.8/topics/class-based-views/
For example you can inherit from "TemplateView" class which provides some features that you may need to use in your own view. For example you can reuse "get" method.
# some_app/views.py
from django.views.generic import TemplateView
class AboutView(TemplateView):
template_name = "about.html"
We use class based views (CBV's) to reduce the amount of code required to do repetitive tasks such as rendering a form or a template, listing the items from a queryset etc.
Using CBV's drastically reduces the amount of code required and should be used where it can be.
It is good to be used in a CRM system. You have a a list view, item view, delete view etc, such as an blog. CBV could help you write less code.
But also because it has do too much for you. Sometimes it will be a little bit troublesome to make some customization or add some extra logic. In this situation, it is more suitable for those really experienced and familiar with CBV so that they could change it easily.
While working on my Django-based projects I'm always trying to follow Django's approach to reusable apps - I'm trying to decouple my applications from each other and especially trying to avoid cross references but sometimes it does not seem to be possible.
Let's consider a simple example with 2 applications: articles and users. Articles application defines article model, articles list view and single article view, users application defines user model and user profile view. Article is referencing user from the author field, so articles application is obviously dependent on users application which is fine.
But when it comes to user profile, I want to display latest articles authored by the user (and may be latest articles viewed by the user) on that page but that makes users application aware of articles application which is what I'm trying to avoid.
I can obviously try to push all such references to the template level but it still does not solve the issue completely and at the same time may be very inefficient in terms of database queries sometimes.
What do you guys do in such cases?
If you are really set on not having any conversation between the 'user' app and the 'article' app, then you need a third app to act as interface. That would know about users and articles and define all the coupling between them. Your article view would be in there, since it has to get the user data, and your user profile view would be in there, because it needs to say "Fred wrote 5 articles".
Whether this level of decoupling is worth it, I don't know. Sometimes programming for re-usability gets in the way of making the thing usable in the first place.
The standard (or preferred) way of keeping coupled apps decoupled is to add a conditional coupling - like in some apps that try to import django-notification and only if they find it, they report events to it.
Still, if you have two apps that talks to each other by design, then I don't see any point in decoupling them - there are plenty of examples in Django world of apps that just require other apps. Note that I'm talking here about writing real-world software, not about some academic delibrations :-)
It seems that in this case, the dependency of users on articles is in a method, not a field. (Whether it's a model method, a model class method, or a manager method is immaterial). If that's so, you can do a lazy import of articles inside the method. By the time this import is performed, users.models will be fully loaded, so even if this is a circular import, it will not cause problems. The "import users" statement in articles will not have to re-load users and will have the full users namespace available.
I've begun diving into Django again and I'm having trouble finding the parallel to some common concepts from my life in C#. While using .NET MVC I very often find myself creating a base controller which will provide a base action implementation to take care of the type of stuff I want to do on every request, like retrieving user information, getting localization values.
Where I'm finding myself confused is how to do this in Django. I am getting more familiar with the MVT concept but I can't seem to find how to solve this scenario. I've looked at class based views and the generic views yet they didn't seem to work how I expected. What am I missing? How can i create default logic that each view will be instructed to run but not have to write it in each view method?
If it is truly common for your whole site you use middleware. If it is only common for some views, the way to go in my opinion is to create decorators for those views. I never use class-based views because I tend to keep views simple and put more logic into models, so I have no need for classes there.