How to share object between blueprint in Flask - flask

I have two blueprints, such as auth and sprints, the structure is as follows:
auth
__init__.py
views.py
forms.py
sprints
__init__.py
views.py
forms.py
I want to share api object after login successfully
#auth.route('/login', methods=['GET', 'POST'])
def login():
form = LoginForm()
if form.validate_on_submit():
try:
jira_url = app.config["JIRA_URL"]
app_id = app.config["JIRA_APPID"]
api = GreenHopper(options={"server": jira_url, "appid": 37},
basic_auth=(form.username.data, form.password.data))
g.api = api
except JIRAError:
api = None
...
...
return redirect(request.args.get('next') or url_for('scrumworks.index'))
return render_template('auth/login.html', form=form)
On the sprints blueprint, I want to get the api object like codes below:
#scrumworks.route("/sprints")
#login_required
def show_sprints():
if current_user.admin:
sprints = g.api.sprints(37)
return render_template("scrumworks/sprints.html")
else:
flash("Only admin users can access this page")
return redirect(url_for(".index"))
When the codes run, it will complain AttributeError: '_AppCtxGlobals' object has no attribute 'api'.
It seems the life cycle of g is per request, not global, so I can't store api in g
Another choice is session, but the api object cannot be serializable, so it can't be stored in session, what should I do next?
Any ideas? Thanks :-)

The g object is only available between requests.
The quick and dirty way to solve this, is to add your api object into the global space where the Flask app is defined and then importing the api object from there.
Keep in mind that with multiple servers and concurrency, this object can get out of sync fairly quickly and should remain stateless.
See this project on github for an example
The preferred solution for this is to use a database that fits the needs of your object. Redis, MongoDB and PostgreSQL are some of the more common ones.

Related

Is it okay to call django functions through an API?

I'm building a project and now I'm new to VueJS I'm currently learning it. and I found that you can make HTTP Requests on APIs using axios. And to make my project easy, Can I call functions on my views.py thru axios?
Like I'm fetching urls in urls.py to execute some functions on my backend.
Is it okay? I mean for security and best practices. etc.
Thanks
Absolutely ok, that's what Django is for:
urls.py:
urlpatterns = [
...
path('my_view_function/', views.my_view_function, name='my_view_function'),
...
]
views.py:
def my_view_function(request):
# unpack data:
my_value = request.GET['my_key']
# some logic:
...
# pack response:
response = json.dumps({
'some_other_key' : 'some_other_value'
})
return HttpResponse(response)
Another option is that you use signals in django, some time ago I used signals that for when a new record is created without a field, it will be completed with the algorithm you want, for example an order and automatically placed a code, you can apply it only by pointing to your models, when you want something more punctual.
#receiver(pre_save, sender=MyModel)
def my_handler(sender, **kwargs):
...
The my_handler function will only be called when an instance of MyModel is saved.
Here I leave the documentation in case you want to review it
https://docs.djangoproject.com/en/3.1/topics/signals/

How should I be implementing user SSO with AAD in a Django application (using the Django Microsoft Authentication Backend module)?

I'm developing a Django (2.2.3) application with Django Microsoft Auth installed to handle SSO with Azure AD. I've been able to follow the quickstart documentation to allow me to log into the Django Admin panel by either using my Microsoft identity, or a standard username and password I've added to the Django user table. This all works out of the box and is fine.
My question put (really) simply is "What do I do next?". From a user's perspective, I'd like them to:
Navigate to my application (example.com/ or example.com/content) - Django will realise they aren't authenticated, and either
automatically redirect them to the SSO portal in the same window, or
redirect them to example.com/login, which requires them to click a button that will open the SSO
portal in a window (which is what happens in the default admin case)
Allow them to sign in and use MFA with their Microsoft Account
Once successful redirect them to my #login_required pages (example.com/content)
Currently, at the root of my navigation (example.com/), I have this:
def index(request):
if request.user.is_authenticated:
return redirect("/content")
else:
return redirect("/login")
My original idea was to simply change the redirect("/login") to redirect(authorization_url) - and this is where my problems start..
As far as I can tell, there isn't any way to get the current instance(?) of the context processor or backend of the microsoft_auth plugin to call the authorization_url() function and redirect the user from views.py.
Ok... Then I thought I'd just instantiate the MicrosoftClient class that generates the auth URL. This didn't work - not 100% sure why, but it think it may have something to do with the fact that some state variable used by the actual MicrosoftClient instance on the backend/context processor is inconsistent with my instance.
Finally, I tried to mimic what the automatic /admin page does - present an SSO button for the user to click, and open the Azure portal in a separate window. After digging around a bit, I realise that I fundamentally have the same problem - the auth URL is passed into the admin login page template as inline JS, which is later used to create the Azure window asynchronously on the client side.
As a sanity check, I tried to manually navigate to the auth URL as it is presented in the admin login page, and that did work (though the redirect to /content didn't).
At this point, given how difficult I think I'm making it for myself, I'm feel like I'm going about this whole thing completely the wrong way. Sadly, I can't find any documentation on how to complete this part of the process.
So, what am I doing wrong?!
A couple more days at this and I eventually worked out the issues myself, and learned a little more about how Django works too.
The link I was missing was how/where context processors from (third party) Django modules pass their context's through to the page that's eventually rendered. I didn't realise that variables from the microsoft_auth package (such as the authorisation_url used in its template) were accessible to me in any of my templates by default as well. Knowing this, I was able to implement a slightly simpler version of the same JS based login process that the admin panel uses.
Assuming that anyone reading this in the future is going through the same (learning) process I have (with this package in particular), I might be able to guess at the next couple of questions you'll have...
The first one was "I've logged in successfully...how do I do anything on behalf of the user?!". One would assume you'd be given the user's access token to use for future requests, but at the time of writing this package didn't seem to do it in any obvious way by default. The docs for the package only get you as far as logging into the admin panel.
The (in my opinion, not so obvious) answer is that you have to set MICROSOFT_AUTH_AUTHENTICATE_HOOK to a function that can be called on a successful authentication. It will be passed the logged in user (model) and their token JSON object for you to do with as you wish. After some deliberation, I opted to extend my user model using AbstractUser and just keep each user's token with their other data.
models.py
class User(AbstractUser):
access_token = models.CharField(max_length=2048, blank=True, null=True)
id_token = models.CharField(max_length=2048, blank=True, null=True)
token_expires = models.DateTimeField(blank=True, null=True)
aad.py
from datetime import datetime
from django.utils.timezone import make_aware
def store_token(user, token):
user.access_token = token["access_token"]
user.id_token = token["id_token"]
user.token_expires = make_aware(datetime.fromtimestamp(token["expires_at"]))
user.save()
settings.py
MICROSOFT_AUTH_EXTRA_SCOPES = "User.Read"
MICROSOFT_AUTH_AUTHENTICATE_HOOK = "django_app.aad.store_token"
Note the MICROSOFT_AUTH_EXTRA_SCOPES setting, which might be your second/side question - The default scopes set in the package as SCOPE_MICROSOFT = ["openid", "email", "profile"], and how to add more isn't made obvious. I needed to add User.Read at the very least. Keep in mind that the setting expects a string of space separated scopes, not a list.
Once you have the access token, you're free to make requests to the Microsoft Graph API. Their Graph Explorer is extremely useful in helping out with this.
So I made this custom view in Django based on https://github.com/Azure-Samples/ms-identity-python-webapp.
Hopefully, this will help someone.
import logging
import uuid
from os import getenv
import msal
import requests
from django.http import JsonResponse
from django.shortcuts import redirect, render
from rest_framework.generics import ListAPIView
logging.getLogger("msal").setLevel(logging.WARN)
# Application (client) ID of app registration
CLIENT_ID = "<appid of client registered in AD>"
TENANT_ID = "<tenantid of AD>"
CLIENT_SECRET = getenv("CLIENT_SECRET")
AUTHORITY = "https://login.microsoftonline.com/" + TENANT_ID
# This resource requires no admin consent
GRAPH_ENDPOINT = 'https://graph.microsoft.com/v1.0/me'
SCOPE = ["User.Read"]
LOGIN_URI = "https://<your_domain>/login"
# This is registered as a redirect URI in app registrations in AD
REDIRECT_URI = "https://<your_domain>/authorize"
class Login(ListAPIView):
'''initial login
'''
def get(self, request):
session = request.session
id_token_claims = get_token_from_cache(session, SCOPE)
if id_token_claims:
access_token = id_token_claims.get("access_token")
if access_token:
graph_response = microsoft_graph_call(access_token)
if graph_response.get("error"):
resp = JsonResponse(graph_response, status=401)
else:
resp = render(request, 'API_AUTH.html', graph_response)
else:
session["state"] = str(uuid.uuid4())
auth_url = build_auth_url(scopes=SCOPE, state=session["state"])
resp = redirect(auth_url)
else:
session["state"] = str(uuid.uuid4())
auth_url = build_auth_url(scopes=SCOPE, state=session["state"])
resp = redirect(auth_url)
return resp
class Authorize(ListAPIView):
'''authorize after login
'''
def get(self, request):
session = request.session
# If states don't match login again
if request.GET.get('state') != session.get("state"):
return redirect(LOGIN_URI)
# Authentication/Authorization failure
if "error" in request.GET:
return JsonResponse({"error":request.GET.get("error")})
if request.GET.get('code'):
cache = load_cache(session)
result = build_msal_app(cache=cache).acquire_token_by_authorization_code(
request.GET['code'],
# Misspelled scope would cause an HTTP 400 error here
scopes=SCOPE,
redirect_uri=REDIRECT_URI
)
if "error" in result:
resp = JsonResponse({"error":request.GET.get("error")})
else:
access_token = result["access_token"]
session["user"] = result.get("id_token_claims")
save_cache(session, cache)
# Get user details using microsoft graph api call
graph_response = microsoft_graph_call(access_token)
resp = render(request, 'API_AUTH.html', graph_response)
else:
resp = JsonResponse({"login":"failed"}, status=401)
return resp
def load_cache(session):
'''loads from msal cache
'''
cache = msal.SerializableTokenCache()
if session.get("token_cache"):
cache.deserialize(session["token_cache"])
return cache
def save_cache(session,cache):
'''saves to msal cache
'''
if cache.has_state_changed:
session["token_cache"] = cache.serialize()
def build_msal_app(cache=None, authority=None):
'''builds msal cache
'''
return msal.ConfidentialClientApplication(
CLIENT_ID, authority=authority or AUTHORITY,
client_credential=CLIENT_SECRET, token_cache=cache)
def build_auth_url(authority=None, scopes=None, state=None):
'''builds auth url per tenantid
'''
return build_msal_app(authority=authority).get_authorization_request_url(
scopes or [],
state=state or str(uuid.uuid4()),
redirect_uri=REDIRECT_URI)
def get_token_from_cache(session, scope):
'''get accesstoken from cache
'''
# This web app maintains one cache per session
cache = load_cache(session)
cca = build_msal_app(cache=cache)
accounts = cca.get_accounts()
# So all account(s) belong to the current signed-in user
if accounts:
result = cca.acquire_token_silent(scope, account=accounts[0])
save_cache(session, cache)
return result
def microsoft_graph_call(access_token):
'''graph api to microsoft
'''
# Use token to call downstream service
graph_data = requests.get(
url=GRAPH_ENDPOINT,
headers={'Authorization': 'Bearer ' + access_token},
).json()
if "error" not in graph_data:
return {
"Login" : "success",
"UserId" : graph_data.get("id"),
"UserName" : graph_data.get("displayName"),
"AccessToken" : access_token
}
else:
return {"error" : graph_data}

How to safely access request object in Django models

What I am trying to do:
I am trying to access request object in my django models so that I can get the currently logged in user with request.user.
What I have tried:
I found a hack on this site. But someone in the comments pointed out not to do it when in production.
I also tried to override model's __init__ method just like mentioned in this post. But I got an AttributeError: 'RelatedManager' object has no attribute 'request'
Models.py:
class TestManager(models.Manager):
def user_test(self):
return self.filter(user=self.request.user, viewed=False)
class Test(models.Model):
def __init__(self, *args, **kwargs):
self.request = kwargs.pop('request', None)
super(Test, self).__init__(*args, **kwargs)
user = models.ForeignKey(User, related_name='test')
viewed = models.BooleanField(default=False)
objects = TestManager()
I trying to access request object in my Django models so that I can get the currently logged in user with request.user.
Well a problem is that models are not per se used in the context of a request. One for example frequently defines custom commands to do bookkeeping, or one can define an API where for example the user is not present. The idea of the Django approach is that models should not be request-aware. Models define the "business logic" layer: the models define entities and how they interact. By not respecting these layers, one makes the application vulnerable for a lot of problems.
The blog you refer to aims to create what they call a global state (which is a severe anti-patten): you save the request in the middleware when the view makes a call, such that you can then fetch that object in the model layer. There are some problems with this approach: first of all, like already said, not all use cases are views, and thus not all use cases pass through the middleware. It is thus possible that the attribute does not exist when fetching it.
Furthermore it is not guaranteed that the request object is indeed the request object of the view. It is for example possible that we use the model layer with a command that thus does not pass through the middleware, in which case we should use the previous view request (so potentially with a different user). If the server processes multiple requests concurrently, it is also possible that a view will see a request that arrived a few nanoseconds later, and thus again take the wrong user. It is also possible that the authentication middleware is conditional, and thus that not all requests have a user attribute. In short there are more than enough scenario's where this can fail, and the results can be severe: people seeing, editing, or deleting data that they do not "own" (have no permission to view, edit, or delete).
You thus will need to pass the request, or user object to the user_test method. For example with:
from django.http import HttpRequest
class TestManager(models.Manager):
def user_test(self, request_or_user):
if isinstance(request_or_user, HttpRequest):
return self.filter(user=request_or_user.user, viewed=False)
else:
return self.filter(user=request_or_user, viewed=False)
one thus has to pass the request object from the view to the function. Even this is not really pure. A real pure approach would only accept a user object:
class TestManager(models.Manager):
def user_test(self, user):
return self.filter(user=user, viewed=False)
So in a view one can use this as:
def some_view(request):
some_tests = Test.objects.user_test(request.user)
# ...
# return Http response
For example if we want to render a template with this queryset, we can pass it like:
def some_view(request):
some_tests = Test.objects.user_test(request.user)
# ...
return render(request, 'my_template.html', {'some_tests': some_tests})

How does interacting with the django rest api through the url work?

I get that the Django rest framework is for interacting with the Django server programmatically but one thing I still don't understand is how.what i want to do is have my client app (mobile app) send data (somehow) to the Django server in order to create/retrieve data based on variables and obviously this has to be done through the URL since there will be no direct GUI interaction with the API. (unless I'm mistaken, which I probably am) I have gone through the official documentation and followed the tutorial to the end and still don't understand how this is supposed to work.all I ask for is a quick and simple explanation because I have searched everywhere and haven't found a simple enough explanation to grasp the core concept of how this is all supposed to work.
I think what you're looking for is JSONResponse and related objects:
This will allow you to send JSON in response to a request.
from django.http import JsonResponse
def my_view_json(request):
response = JsonResponse({'foo': 'bar'})
return response
If your templates or webpages need to make a request to a view and specify different parameters, they can do so by adding POST variables (examples). These can be parsed in the view like so:
def myView(request):
my_post_var = request.POST.get('variable_name', 'default_value')
my_get_var = request.GET.get('variable_name', 'default_value')
You can then parse what was sent any way you like and decide what you want to do with it.
Basically,
You define the URLS upon which you perform Get/POST/PUT Requests and You can Send Data to that.
Eg:
urls.py
from django.conf.urls import url,include
from app import views
urlpatterns = [
url(r'^(?i)customertype/$',views.CustomerViewSet.as_view()),
url(r'^(?i)profile/$', views.Save_Customer_Profile.as_view()),
url(r'^(?i)customer_image/$', views.Save_Customer_Image.as_view()),
]
Now Whenever User would send a Request to:
example.com/profile ==> This would be received in the Save_Customer_Profile View based on the Method Type, Save_Customer_Profile is as follows:
class Save_Customer_Profile(APIView):
"""Saves and Updates User Profile!"""
def get(self, request, format=None):
return AllImports.Response({"Request":"Method Type is GET Request"})
def post(self, request, format=None):
return AllImports.Response({"Request":"Method Type is Post Request"})
def put(self,request, format=None):
return AllImports.Response({"Request":"Method Type is Put Request"})
I think the OP was referring to how to do GET/POST request programmatically. In that case, it is enough to do (values are dummy):
GET:
import requests
r = requests.get('http://localhost:8000/snippets/')
print(r.json())
print(r.status_code, r.reason)
POST:
data = {'code': 'print(" HELLO !!!")', 'language': 'java','owner': 'testuser'}
r = requests.post('http://localhost:8000/snippets/', data=data, auth=('testuser', 'test'))

Django Rest Framework testing save POST request data

I'm writing some tests for my Django Rest Framework and trying to keep them as simple as possible. Before, I was creating objects using factory boy in order to have saved objects available for GET requests.
Why are my POST requests in the tests not creating an actual object in my test database? Everything works fine using the actual API, but I can't get the POST in the tests to save the object to make it available for GET requests. Is there something I'm missing?
from rest_framework import status
from rest_framework.test import APITestCase
# from .factories import InterestFactory
class APITestMixin(object):
"""
mixin to perform the default API Test functionality
"""
api_root = '/v1/'
model_url = ''
data = {}
def get_endpoint(self):
"""
return the API endpoint
"""
url = self.api_root + self.model_url
return url
def test_create_object(self):
"""
create a new object
"""
response = self.client.post(self.get_endpoint(), self.data)
self.assertEqual(response.status_code, status.HTTP_201_CREATED)
self.assertEqual(response.data, self.data)
# this passes the test and the response says the object was created
def test_get_objects(self):
"""
get a list of objects
"""
response = self.client.get(self.get_endpoint())
self.assertEqual(response.status_code, status.HTTP_200_OK)
self.assertEqual(response.data, self.data)
# this test fails and says the response is empty [] with no objects
class InterestTests(APITestCase, APITestMixin):
def setUp(self):
self.model_url = 'interests/'
self.data = {
'id': 1,
'name': 'hiking',
}
# self.interest = InterestFactory.create(name='travel')
"""
if I create the object with factory boy, the object is
there. But I don't want to have to do this - I want to use
the data that was created in the POST request
"""
You can see the couple lines of commented out code which are the object that I need to create through factory boy because the object does not get created and saved (although the create test does pass and say the object is created).
I didn't post any of the model, serializer or viewsets code because the actual API works, this is a question specific to the test.
First of all, Django TestCase (APITestCase's base class) encloses the test code in a database transaction that is rolled back at the end of the test (refer). That's why test_get_objects cannot see objects which created in test_create_object
Then, from (Django Testing Docs)
Having tests altering each others data, or having tests that depend on another test altering data are inherently fragile.
The first reason came into my mind is that you cannot rely on the execution order of tests. For now, the order within a TestCase seems to be alphabetical. test_create_object just happened to be executed before test_get_objects. If you change the method name to test_z_create_object, test_get_objects will go first. So better to make each test independent
Solution for your case, if you anyway don't want database reset after each test, use APISimpleTestCase
More recommended, group tests. E.g., rename test_create_object, test_get_objects to subtest_create_object, subtest_get_objects. Then create another test method to invoke the two tests as needed