How can I monitor which routes and urls commonly requested? - django

We have an app made from reactjs and django and it is hosted on heroku. We use axios in reactjs to make and get and post request to the django side, now are there any method in heroku that we could monitor these url requests, e.g. which url path is most commonly requested and how much are they costing at the end of the month.

If Heroku doesn't offer anything out-of-the-box, you could utilize Google Analytics to fire off an event using the library react-ga. You could put this in the finally() portion of your fetch call after the then() and catch() functions, this way you're able to view the number of requests attempted regardless of whether or not it was successful.

Related

Is separating django frontend and backend with API viable?

I'm used to Django and already developed website with a whole Django or with Django API + React.
For a side project, I'm thinking about creating a Django project with 2 apps:
Django API 2) Django Front.
I absolutely want to decouple front/back to be able to reuse the API in the future, and I like the view system from Django.
Is my idea making sense?
Edit 1:
To clarify.
App 1: Django API serving JSON.
App 2: Django App using API calls in the controllers to generate the views.
Edit 2:
I did a proof of concept and it works like a charm.
The only drawback is that I have to use Cookies to store the JWT token and maintain the session state, which is not convenient
it is possible, but completely wrong with idea.
How it possible. Wrong version:
Try to remember, how we can create integrate test for our view. We Should create client and send request to Django to your view-url with args kwargs querystring e.t.c.
In response you have already answer. Try to imagine: part with client - is your Django front, requested part - your backend.
The same works not only in test, you can create request to ask something on the completely other server.
Redis/MemCashed e.t.c. (pattern sender/receiver) Wrong version:
The Front Django speaks with Backend through Third part application. It can be faster.
"Pythonic" version. Right in your case:
You can create Backend Django part like a library with views as interfaces.
Frontend Django part is completely standalone, it import an use interfaces from library, which is your "BackEnd Module".
If you want - you can use "BackEnd Module" everywhere as import, or you can start it as server and ask info per requests.
Completely independent in this case means, you can install "BackEnd Module" without "FrontEnd Module". "FrontEnd Module" can be installed standalone too and it should take from settings, which interfaces should be used as data-source.
I hope, i am right understand your question.
You could definitely separate front and back, remember, django just creates endpoints and consumes them with its own views, you can even use both.
Here is the documentation for django views: https://docs.djangoproject.com/en/4.0/#the-view-layer
You can use a librarie like React as frontend and connect to your api(django app) and use both.

Is there a way in django to update the same page with the response without totally rendering it?

Is there a way in django to update the same page with the response without totally rendering it. I am trying to create a code editor to test. But when I am returning the results my contents are removed. I understand it is because I am rendering the page . I need the contents to retain. How can I do it using redirect? I am including the render statement I used and a screenshot of how it looks here:
Steps:
Handle post request
Program execution code
Save the result in a variable called "message". Then I used
return render(request, 'editor.html', {'message': message})
I want to redirect the message to the same page without rendering a new page.
[Before submission][1]
[After submission][2]
[1]: https://i.stack.imgur.com/BxoLU.png
[2]: https://i.stack.imgur.com/uiEOU.png
Any help will be appreciated. Thank you.
it is possible. using ajax in front-end and render or render_to_string in back-end(Django). using ajax you're able to call a url (with/without data you want to send), then in the views.py you can do anything needed and return render(request,'template.html, context). then in ajax you have success: function (res) { $('#id_of_element').append(res)}. this function in ajax will append the recived response which is a rendered HTML template to the target element.
For that, you have to switch to a different web software paradigm called "single page application", which implies that both, backend and frontend, are functional software components on their own, instead of having a "dumb" HTML frontend that only displays what the backend renders.
In a regular web application, the front end is served from a backend with all the information that is going to display. In a Single Page Application, the front end is served by a server independent of the backend server, and the frontend and backend interact through an API served by the backend.
With this architecture, the frontend component is responsible for requesting and providing data from and to the backend, as well as for displaying the data and getting user's interaction, and the mean for interchanging data with the backend is called an ajax, that is an asynchronous request.
The only language accepted by web browsers is javascript, but there are many frameworks and second level languages that can render a javascript application, like React, Angular, Vue, and many others.

Populate web page using API result directly from API server

Firstly sorry if this question has been asked before but I'm a novice so even if it has I'm unaware of the language I'd even use to try and seek it out.
I'm beginning to learn about REST API's and it got me thinking. Is it possible to load the JSON response directly from the API server into the user's browser and bypass your own server?
Imagine you have say a Django app running on a server that accesses email messages from Outlook.com using the graph API.
I assume an ordinary flow would go something like:
User request->your server->graph api-> your server-> user browser.
It seems like a waste for it to hit your server that second time before it goes on to be presented to the user's browser.
Is there a way the Django app can render a template and effectively tell the browser "expect some data from X source, and place it in y location in this template"?
You could do that with javascript. You'd have to include either a script tag in your template, or create and include some static javascript files with the code.
I'd recommend learning and using the jQuery javascript library, as it makes what you're talking about much easier to implement. Research ajax requests, those are what you'll need to make requests directly to another server, bypassing your own

Communication between Django and React

I'm trying to setup a project using Django for backend and React for frontend. The project has several screens, a lot of information in DB and images generated by the backend, and will include some authentication and user permissions for different screens.
According to what I found - the best way to do it is having Django render an html file:
def index(request):
return render(request, 'frontend/index.html')
which references a .js file:
<script src="{% static "frontend/main.js" %}"></script>
Which is created using Webpack.
This main.js retrieves the data it needs from Django using a REST api:
fetch("...some Django endpoint..").then(response => ... this.setState(...retrieved data...))
Unlike when just using Django for backend + Django templates for frontend where the backend can just send the context directly to the template:
def index(request):
context = {'information': .... retrieve info from DB}
return HttpResponse(loader.get_template('bla/index.html').render(context, request))
and the template can use this info directly, without referencing the backend again:
{% for bla in information %}
I'm wondering if it is a reasonable setup?
It seems excessive to have the frontend use REST for retrieving each piece of information it needs and the backend exposing another REST api for each part of data it needs to supply (Instead of just pushing all of the information to a single dict and sending it over along with the template),
Also, it requires at least 2 RTTs to render the full page (which I guess usually is okay)
According to what I found - the best way to do it is having Django render an html file:
I disagree with this line. I would say it would be best to keep the react app and Django app totally separate.
I believe, the Django application should solely provide APIs and adminsite(maybe, depending on your needs).
And the frontend should be a standalone app which can be served through NGINX/ExpressJs/Apache etc.
There are several advantages of this setup.
From Django application's perspective, the advantages are:
Django will not be burdened to serve the Frontend. Use gunicorn or uwsgi to serve the Django APIs.
As Django will provide data through API only, it will provide clarity on how the frontend application will communicate with the backend. I know that you can send data using context when Django serves the react app, but this might cause confusion because of API and context's co-existence.
You can use Token based authentication, JWT etc instead of Django's own session based authentication, which have a lot of other advantages.
Freeing your frontend application from backend is the best thing can happen for the frontend. Like for example:
if you had Django to serve the frontend, you were almost forced to use session based auth(its not like you can't use other auths, but whats the point of having multiple auth systems)
You couldn't have used server side rendering with Django rendering the frontend.
Lets say, you are have no idea about how Django works, but you will be forced to setup a Django application in your local machine, because it serves the frontend.
You couldn't have used ExpressJs to serve the frontend, or use the advantages of using NGINX to serve those contents.
Deployment would be complicated if you have docker setup. In this case, you would have had to use one Docker Container to serve everything, else you could have used multiple docker containers to serve backend/frontend.
Lets say, you want to serve the Django application in one server, frontend from other server, but with Django tightly coupled with Frontend, you can't do this setup.
You can easily connect external RESTful API services without bothering about Django. Even you can use any other frameworks like Tornado, Flask etc(but DRF+Django ORM is awesome) to develop APIs.
There are some more generic advantages of having backend and frontend separated.
There is a fantastic tutorial which you can read on medium about setting up separate Django + ReactJs app.
You can use GraphQL, it has several advantages over REST, f.e.:
only one endpoint for entire app;
ability to fetch data with relations between them;
easy data structure modifications on both sides;
advanced client with cache/normalization/subscriptions/optimistic updates (I prefer apollo to relay);
can be used as datasource for static site generation (SEO);
you can stich other services/APIs;
... many more.
Using react Server Side Rendering you can get pages without additional requests - 'prefilled'/rehydrated, ready for interactions - better Time to Interactive.
Tutorial/demo: django-graphql-apollo-react-demo

How to run Django views from a script?

I am writing a Django management command that visits a few pages, logged in as a superuser, and saves the results to a set of .html files.
Right now I'm just using the requests library and running the command with the development server running. Is there an easy way to generate the HTML from a view response so I do without actual HTTP requests?
I could create a request object from scratch but that seems like more overhead than the current solution. I was hoping for something simple.
Django has a RequestFactory which seems to suit your needs.
While it's not exactly meant for this purpose, an option would be to use the testing framework's Client to fake a request to the url - be sure to use client.login() before making your requests, to ensure you have superuser capabilities.