A talk I gave at the Stockholm Python Meetup hosted at Lifesum, covering how to use dynamic plans in Django to drive mobile app behaviour from the backend.

Slides:

... and some answers

Well I haven't conducted any interviews recently but this one has been laying in my drafts for a quite while so it is time to take it out of the dust and finish it. As I have said in Python Interview Question and Answers these are basic questions to establish the basic level of the candidates.

1. Django's request response cycle
   You should be aware of the way Django handles the incoming requests - the execution of the middlewares, the work of the URL dispatcher and what should the views return. It is not necessary to know everything in the tiniest detail but you should be generally aware of the whole picture. For reference you can check "the live of the request" slide from my Introduction to Django presentation.
2. Middlewares - what they are and how they work
   Middlewares are one of the most important parts in Django. Not only because they are quite powerfull and useful but also because the lack of knowledge about their work can lead to hours of debugging. From my experience the process_request and process_response hooks are the most frequently used and those are the one I always ask for. You should also know their execution order and when their execution starts/stop. For reference check the official docs.
3. Do you write tests? How?
   Having well tested code benefits you, the project and the rest of the team that is going to maintain/modify your code. Django's built-in test client, Django Webtest, Factory boy, Faker or whatever you use, I would like to hear about it and how you do it. There is not only one way/tool for it so I would like to know what are you using and how. You should show that you understand how important is to test your code and how to do it right.
4. select_related and prefetch_related
   Django's ORM (as every other) can be both a blessing and a curse. Getting a foreign key property can easily lead you to executing million queries. Using select_related and prefetch_related can come as a saviour so it is important to know how to use them.
5. Forms validation
   I expect from the candidates that they know how to build forms, to add custom validation for one field and how to validate fields that depend on each other.
6. Building a REST API with Django
   The most popular choices out there are Django REST Framework and Django Tastypie. Have you used any of them, why an how. What issues have you faced? Here is the place to show that you understand the concept of REST APIs, the correct request/response types and the serialisation of the data.
7. Templates
   Building website using the built-in template system is often part of your daily tasks. This means that you should understand the template inheritance, how to reuse the block and so on. Having experience with sekizai is a plus.

There are a lot of other things to ask about internationalisation(i18n), localisation(l10n), south/migrations etc. Take a look at the docs and you can see them explained pretty well.

In a few words PyCon Sweden 2015 was awesome. Honestly, this was my first Python conference ever but I really hope it won't be the last.

Outside the awesome talks and great organisation it was really nice to spend some time with similar minded people and talk about technology, the universe and everything else. I have met some old friends and made some new ones but lets get back to the talk. Unfortunately I was not able to see all of them but here is a brief about those I saw and found really interesting:

It all started with Ian Ozsvald and his awesome talk about "Data Science Deployed" (slides / video). The most important point here were:

• log everything
• think about data quality, don't use everything just what you need
• think about turning data into business values
• start using your data

Then Rebecca Meritz talked about "From Explicitness to Convention: A Journey from Django to Rails" (slides / video). Whether the title sounds a bit contradictive this was not the usual Django vs Rails talk. At least to me it was more like a comparison between the two frameworks, showing their differences, weak and strong sides. Whether I am a Django user, I am more and more keen to the position that none of the frameworks is better than the other one, they are just two different approaches for building great web apps.

Flavia Missi and "Test-Driven-Development with Python and Django" (slides / video). TDD will help you have cleaner, better structured and easy to maintian code. If you are not doing it the best moment to start is now. Whether it is hard at the beginning you will pretty soon realise how beneficial it is. Especially if someone pushed a bad commit and the tests saved your ass before the code goes to production.

Later Dain Nilsson talked about "U2F: Phishing-proof two-factor authentication for everyone" (video). Whether I don't use two-factor authentication at the moment I am familiar with the concept and I really like it. The U2F protocol looks like a big step towards making it more applicable over different tools and applications and the key holding devices are more and more accessible nowadays. Maybe it is time for me to get one )))

The second day started with Kate Heddleston who talked about ethics in computer programming (video). About how social networks can be used as a tool for ruining peoples lifes and that we as a developers should take a responsibility and work towards making the internet a safer place for everyone. A place where you can have your privacy and have protection if harassed. It is a big problem which won't be solved in a night, but talking about it is the first step towards solving it.

"Why Django Sucks" by Emil Stenström. Don't rush into raging cause this was one of best talks at the conference. Emil showed us the parts in Django where he sees flaws and that need improvement. The main point was the lack of common template language between Django and Javascript and what possible solutions are available or can be made.

Daniele Sluijters reminded us how easy is to work with "Puppet and Python" (video). No more fighting with Ruby, you can easily use your favorite language to build your own tools and helpers

Dennis Ljungmark and "Embedded Python in Practice". The last time I programmed embedded devices was 15 years ago as a part of short course in the high school. Dennis' work is much more complex than what I did then but his talk reminded me of things that are applicable to general programming. Whether using non-embedded systems we often have much more memory and processing power available that does not mean that we should waste it. So think when you code - RAM is not endless and processors are not that fast as we often wish. Also don't forget that Exceptions wil sooner or later occur so make your code ready to handle them.

"How to Build a Python Web Application with Flask and Neo4j" (video) by Nicole White. Well I have heard about Neo4J, but I have never used it or seen it in action so this one was really exciting. Neo4J offers you a whole new perspective about building databases and relations between objects but it is much far from panacea. Actually I can see it it is more like a special tool then as a general replacement of a relation database but it still worths to be tried. Oh, and the Neo4J browser - totally awesome.

In the lightning talks Tome Cvitan talked about Faker. If you are still not familiar with it now is a good time to try it. Especially if you are writing tests.

At the final Kenneth Reitz told us about "Python for Humans"(video). About the not that obvious things in Python and what solutions are out there. And also about The Hitchhiker’s Guide to Python! a great place for beginners and not only and shared the idea to make Python easier and more welcoming by introducing better tools for higher level of operations.

Finally I want to thank to everyone - the organisers, the speakers, the audience and practically everyone who was a part of the conference. Without you it would be the same (or be at all). Thanks, keep up the good work and hopefully we will sea each other again.

P.S. Have I mentioned that the whole conference was recorded on video so hopefully we will see be able to see all the talks pretty soon. I will try to keep this post updated with the links to the videos and/or slides when they become available. Of course if you know about any published slides from the conference that are not linked here please let me know.

The full set of videos are available at PyCon Sweden's Youtube channel.

Brief: In one of the project I work on we had to convert some old naive datetime objects to timezone aware ones. Converting naive datetime to timezone aware one is usually a straightforward job. In django you even have a nice utility function for this. For example:

import pytz
from django.utils import timezone


timezone.make_aware(datetime.datetime(2012, 3, 25, 3, 52),
                    timezone=pytz.timezone('Europe/Stockholm'))
# returns datetime.datetime(2012, 3, 25, 3, 52, tzinfo=<DstTzInfo 'Europe/Stockholm' CEST+2:00:00 DST>)

Problem: You can use this for quite a long time until one day you end up with something like this:

timezone.make_aware(datetime.datetime(2012, 3, 25, 2, 52),
                    timezone=pytz.timezone('Europe/Stockholm'))

# which leads to
Traceback (most recent call last):
  File "<console>", line 1, in <module>
  File "/home/ilian/venvs/test/lib/python3.4/site-packages/django/utils/timezone.py", line 358, in make_aware
    return timezone.localize(value, is_dst=None)
  File "/home/ilian/venvs/test/lib/python3.4/site-packages/pytz/tzinfo.py", line 327, in localize
    raise NonExistentTimeError(dt)
pytz.exceptions.NonExistentTimeError: 2012-03-25 02:52:00

Or this:

    
timezone.make_aware(datetime.datetime(2012, 10, 28, 2, 52),
                    timezone=pytz.timezone('Europe/Stockholm'))

#throws
Traceback (most recent call last):
  File "<console>", line 1, in <module>
  File "/home/ilian/venvs/test/lib/python3.4/site-packages/django/utils/timezone.py", line 358, in make_aware
    return timezone.localize(value, is_dst=None)
  File "/home/ilian/venvs/test/lib/python3.4/site-packages/pytz/tzinfo.py", line 349, in localize
    raise AmbiguousTimeError(dt)
pytz.exceptions.AmbiguousTimeError: 2012-10-28 02:52:00

Explanation: The reason for the first error is that in the real world this datetime does not exists. Due to the DST change on this date the clock jumps from 01:59 standard time to 03:00 DST. Fortunately (or not) pytz is aware of the fact that this time is invalid and will throw the exception above. The second exception is almost the same but it happens when switching from summer to standard time. From 01:59 DST the clock shifts to 01:00 standard time, so we end with a duplicate time.

Why has this happened(in our case)? Well we couldn't be sure how exactly this one got into our legacy data but the assumption is that at the moment when the record was saved the server has been in different timezone where this has been a valid time.

Solution 1: This fix is quite simple, just add an hour if the exception occurs.

try:
    date = make_aware(
        datetime.fromtimestamp(date_time, timezone=pytz.timezone('Europe/Stockholm'))
    )
except (pytz.NonExistentTimeError, pytz.AmbiguousTimeError):
    date = make_aware(
        datetime.fromtimestamp(date_time) + timedelta(hours=1),
        timezone=pytz.timezone('Europe/Stockholm')
    )

Solution 2: Instead of calling make_aware call timezone.localize directly.

try:
    date = make_aware(
        datetime.fromtimestamp(date_time, timezone=pytz.timezone('Europe/Stockholm'))
    )
except (pytz.NonExistentTimeError, pytz.AmbiguousTimeError):
    timezone = pytz.timezone('Europe/Stockholm')
    date = timezone.localize(datetime.fromtimestamp(date_time), is_dst=False)

The second solution probably needs some explanation. First lets check what make_aware does. The code bellow is take from Django's sourcecode as it is in version 1.7.7

def make_aware(value, timezone):
    """
    Makes a naive datetime.datetime in a given time zone aware.
    """
    if hasattr(timezone, 'localize'):
        # This method is available for pytz time zones.
        return timezone.localize(value, is_dst=None)
    else:
        # Check that we won't overwrite the timezone of an aware datetime.
        if is_aware(value):
            raise ValueError(
                "make_aware expects a naive datetime, got %s" % value)
        # This may be wrong around DST changes!
        return value.replace(tzinfo=timezone)    

To simplify it, what Django does is to use the localize method of the timezone object(if it exists) to convert the datetime. When using pytz this localize method takes two arguments: the datetime value and is_dst. The last argument takes three possible values: None, False and True. When using None and the datetime matches the moment of the DST change pytz does not know how to handle the datetime and you get one of the exceptions shown above. False means that it should convert it to standard time and True that it should convert it to summer time.

Why isn't this fixed in Django? The simple answer is "because this is how it should work". For a bit longer check the respectful ticket.

Reminder: Do not forget that the "fix" above does not actually care whether the original datetime is during DST or not. In our case this was not criticla for our app, but in some other cases it might be, so use it carefully.

Thanks: Special thanks to Joshua who correctly pointed out in the comments that I have missed the AmbiguousTimeError in the original post which made me to look a bit more in the problem, research other solutions and update the article to its current content.

Brief: Working with intervals in Python is really easy, fast and simple. If you want to learn more just keep reading.

Task description: Lets say that the case if the following, you have multiple users and each one of them has achieved different number of points on your website. So you want, to know how many users haven't got any point, how many made between 1 and 50 points, how many between 51 and 100 etc. In addition at 1000 the intervals start increasing by 100 instead of 50.

Preparing the intervals: Working with lists in Python is so awesome, so creating the intervals is quite a simple task.

intervals = [0] + \  # The zero intervals
            [x * 50 for x in range(1, 20)] + \  # The 50 intervals
            [x * 100 for x in range(10, 100)] + \  # The 100 intervals
            [x * 1000 for x in range(10, 102)]  # the 1000 intervals

So after running the code above we will have a list with the maximum number of points for each interval. Now it is time to prepare the different buckets that will store the users count. To ease this we are going to use a defaultdict.

from collections import defaultdict

buckets = defaultdict(lambda: 0)

This way, we can increase the count for each bucket without checking if it exists. Now lets got to counting

for user in users:
    try:
        bucket = intervals[bisect.bisect_left(intervals, user.points)]
    except IndexError:
        # we are over the last bucket, so we put in in it
        bucket = intervals[-1]

buckets[bucket] += 1

How it works: Well it is quite simple. The bisect.bisect_left uses a binary search to estimate the position where an item should be inserted to keep the list, in our case intervals sorted. Using the position we take the value from the invervals that represent the bucker where the specified number should go. And we are ready. The result will looks like:

    {
        1: 10,
        10: 5,
        30: 8,
        1100: 2
    }

Final words: As you see when the default dict is used it does not have values for the empty buckets. This can be good or bad depending from the requirements how to present the data but it can be esily fixed by using the items from the intervals as keys for the buckets.

P.S. Comments and ideas for improvement are always welcome.