I really like that this app comes with a sample survey. I might add some sample data to the call-tracking-django repo also.

When reviewing tutorial code, I'm always looking for ways to reduce the number of commands a developer has to run to use our project.

The load_survey management command is useful, but right now a developer has to run it as a separate command after they migrate the database. It would be better if we could load that data for them automatically as part of a data migration.

See the Django docs for an example: https://docs.djangoproject.com/en/1.8/topics/migrations/#data-migrations

By adding a data migration after your other migrations, we can automatically load the sample survey into the database for our developers.

If you like, you can just have your data migration invoke your load_survey management command using call_command: https://docs.djangoproject.com/en/1.8/ref/django-admin/#running-management-commands-from-your-code

django-environ is a good library for managing difference between development environments, but I find it overkill for our sample apps.

Please remove it and just use normal os.environ.get() when you need to find environment variables.

You can reference the call-tracking-django settings for more examples: https://github.com/TwilioDevEd/call-tracking-django/tree/master/twilio_sample_project/settings

Right now when a developer sets up the app locally, they'll go to http://localhost:8000, get redirected to /automated_survey, and get a 500 error because there are no results in the system.

We want developers to know when they try to access the site for the first time that they configured the app correctly. To fix this, I would adjust the /automated_survey view and template to display some content that says "No survey data yet".

Hey I just started implementing this, however after sniffing the request a little I discovered that there is no is_sms attribute on the twilio request coming through causing it to crash at line 38 of surveys.py

if request.is_sms:

This one's easy - just add an admin.py module (like this one) so that developers can easily change the data that's in the system.

You might need to add a __str__ method to your Survey and QuestionResponse models to get them to show up.

Dependabot couldn't authenticate with https://pypi.python.org/simple/.

You can provide authentication details in your Dependabot dashboard by clicking into the account menu (in the top right) and selecting 'Config variables'.

View the update logs.

When I set up the app for the first time, my Twilio calls would error because I had configured my number to point to /first_survey instead of /first_survey/.

Django error'd because by default it tries to append a slash to inbound requests, but it can't do that on a POST request without the browser dropping the data.

So we need to either update the instructions to include the trailing slash, or change the APPEND_SLASH setting to False.

https://docs.djangoproject.com/en/1.8/ref/settings/#append-slash

I set up this project on a windows 7 server running Python 3.4. It took me a little while to get the packages set up but when I try to run migrate it won't load the sample questions (although it did set up the tables correctly). Instead it gives the error attached.

This should be our last issue:

Testing the app on Heroku today, I spotted the kind field on the Question model. Right now it's a CharField but it looks like you're expecting only a small set of values:

• voice
• numeric
• yes-no

We should modify Question.kind to use a choices attribute instead. That way when devs are playing around in the admin, we'll save them from making a typo.

Choices docs: https://docs.djangoproject.com/en/1.8/ref/models/fields/#choices

I had a hard time familiarizing myself with this Django app because it wasn't laid out in the way I expected. It felt like it was configured more like a Flask app than a Django application, but they actually have fairly different code structures.

My main concern is that there's no clear separation between the Django project directory and your automated_survey Django app directory.

I just recently finished the Call Tracking Django tutorial. I think we should structure all our Django tutorials like this going forward: https://github.com/TwilioDevEd/call-tracking-django

Here's a tree output of the call-tracking-django project structure with some comments:

├── LICENSE
├── Procfile
├── README.md
├── app.json
├── call_tracking <- This is your Django app - in your case, automated_survey
│   ├── __init__.py
│   ├── admin.py
│   ├── forms.py
│   ├── migrations
│   ├── models.py
│   ├── static <- This is where the static files your app needs go
│   ├── templatetags
│   ├── test_models.py
│   ├── test_utils.py <- Test modules can be like this or in a separate 'tests' subdirectory if you prefer
│   ├── test_views.py
│   ├── tests.pyc
│   ├── urls.py <- These are the URLs for views within your automated_survey app
│   ├── utils.py
│   ├── views.py <- You can make a views directory instead if you don't want to put them all in one module
├── manage.py
├── requirements.txt
├── runtime.txt
├── templates <- This contains all the templates you use in your app
│   ├── base.html
│   ├── call_tracking
│   └── index.html
└── twilio_sample_project <- you should also name your Django project folder twilio_sample_project
    ├── __init__.py
    ├── settings <- This is the same as your settings directory
    ├── staticfiles <- This is the directory that's created by 'manage.py collectstatic' - not the directory where static files are committed to the repo
    ├── urls.py <- This is the root URLs for your project
    ├── wsgi.py

I wish I had finished the call-tracking-django app before you started on this one. I'm going to spend time this week making a template Django starter project for us all to use so we don't need to worry about this in the future.

Mostly because Google says so, URLs should have hypens instead of underscores in their patterns.

Because the URL names in Django are Python strings, people usually use underscores for those.

So in this example: https://github.com/TwilioDevEd/automated-survey-django/blob/master/automated_survey/urls.py#L39-L41

It should look like this:

    url(r'^first-survey/',
        csrf_exempt(redirect_to_first_survey),
        name='first_survey'),

Unlike with Flask, few people use class-based views in Django, instead opting for functional views.

Your QuestionResponseView is a big view, so I think it's good that it's a class-based view with helper methods.

Some other views are simpler, though, and should just be normal function-based views that Django developers will understand more easily:

• QuestionView
• SurveyView
• SurveyResultsView

These views are better off looking like redirect_to_first_survey - functions that return their response.

Is the image directory on the master branch only used in the tutorial content? If so, that directory should only be in the gh-pages branch.

I forked this Tutorial and am trying to get it to run locally. The current behavior is that, I send a text to the Twilio number and then it correctly texts me the first question of the survey. However, then when I text back, I don't get a response (should be the 2nd question of the survey) and I see a 500 error in Django. I can see that the second text message is being received by Django and, from the following function, it prints HELLO. However, it doesn't make it to YEP, so it appears to be erroring on answering_question = request.session.get('answering_question_id'), which seems like it is a problem with the session. I can see that there is a new session in the Django_session table created from my first text. I'm stuck on what is going wrong. request.session['answering_question_id'] is being set elsewhere in the code, but, my understanding, is if wasn't set, then answering_question would just be false, not an error. I'm stuck - any thoughts?

def redirects_twilio_request_to_proper_endpoint(request):
    print("HELLO")
    answering_question = request.session.get('answering_question_id')
    print("YEP")
    if not answering_question:
        print('1')
        first_survey = Survey.objects.first()
        redirect_url = reverse('survey',
        kwargs={'survey_id': first_survey.id})
   else:
       print('2')
       question = Question.objects.get(id=answering_question)
       redirect_url = reverse('save_response',
       kwargs={'survey_id': question.survey.id,
               'question_id': question.id})
  return HttpResponseRedirect(redirect_url)

Example error
[03/Jun/2022 03:34:36]"GET /automated-survey/first-survey/?ToCountry=US&ToState=CA&SmsMessageSid=SM9f799005f09308ae910a7c5af722bf76&NumMedia=0&ToCity=LOS+ANGELES&FromZip=95747&SmsSid=SM9f799005f09308ae910a7c5af722bf76&FromState=CA&SmsStatus=received&FromCity=ROSEVILLE&Body=Pink&FromCountry=US&To=%2B17472454008&ToZip=&NumSegments=1&ReferralNumMedia=0&MessageSid=SM9f799005f09308ae910a7c5af722bf76&AccountSid=ACef9b14863560f9a5b30961d21029d03c&From=%2B19164327401&ApiVersion=2010-04-01 HTTP/1.1" 500 139520