This post outlines that bare minimum that needs to be done to set up a post_save signal using Django. One rationale behind using a signal is that a model instance can be saved without the user having to wait for additional code to process (e.g. consuming an external API). This lets the user quickly save a form and be on their merry way while in the background the model save() method triggers our signal which executes additional code. This all happens without incuring any wait time for the user.

Create a new file to hold our signals:

# polls/signals.py
from django.db.models.signals import post_save
from django.dispatch import receiver
from .models import Polls


@receiver(post_save, sender=Polls)
def email_poll_author(sender, **kwargs):
    # Code to email poll author

""" https://docs.djangoproject.com/en/stable/topics/signals/#connecting-to-signals-sent-by-specific-senders """

Initialize our signals when Django loads:

# polls/apps.py
from django.apps import AppConfig


class PollsConfig(AppConfig):
    def ready(self):
        from . import signals

""" https://docs.djangoproject.com/en/stable/ref/applications/#django.apps.AppConfig.ready """

Lastly, tell Django to use our AppConfig (polls/apps.py):

# polls/__init__.py
default_app_config = 'polls.apps.PollsConfig'

""" https://docs.djangoproject.com/en/stable/ref/applications/#for-application-authors """

Signals: https://docs.djangoproject.com/en/stable/topics/signals/
Additional types of signals: https://docs.djangoproject.com/en/stable/ref/signals/

Here's the generalized snippet:

any(dict.get('key') == search_value for dict in list)

A full example:

authors = [
    {'author': 'Michael Crichton', title: 'Jurassic Park'},
    {'author': 'C.S. Lewis', title: 'Til We Have Faces'}
]

if any(book.get('author') == 'Michael Crichton' for book in books):
    print('Found Michael Crichton!')

Or less specifically:

if any('Lewis' in book.get('author') for book in books):
    print('Found Lewis!')

Using the ManifestStaticFilesStorage storage backend alters the filename of our static files. For example, style.css becomes something like style.3d94ea84cd8a.css. When collectstatic is run and finds a file that has changed, the MD5 portion of the filename will be updated (style.3d94ea84cd8a.css -> style.1d74ea7349df.css). This prevents browsers and other caching technologies (e.g. Cloudflare) from referring to an outdated static file.

Here are the minimum settings / steps to get this going:

# settings.py
STATIC_ROOT = os.path.join(BASE_DIR, 'static'),

STATICFILES_STORAGE = 'django.contrib.staticfiles.storage.ManifestStaticFilesStorage'

STATICFILES_DIRS = [
    os.path.join(BASE_DIR, 'staticfiles'),
]

After setup (and after updates to any files contained within STATICFILES_DIRS), make sure to run ./manage.py collectstatic. This is easy to forget!

More:
https://docs.djangoproject.com/en/stable/ref/contrib/staticfiles/#manifeststaticfilesstorage
https://docs.djangoproject.com/en/stable/ref/settings/#staticfiles-storage

Here's how I set up logging where:

• Log files rotate nightly (at midnight).
• We keep 10 total log files (as history).

import logging
import os

# Set up logging
logger = logging.getLogger(__file__)
handler_kwargs = {
    'filename': os.path.join(base_dir, 'file.log'),
    'when': 'midnight',
    'backupCount': 10
}
handler = handlers.TimedRotatingFileHandler(**handler_kwargs)
formatter = logging.Formatter('%(asctime)s %(levelname)s %(message)s')
handler.setFormatter(formatter)
logger.addHandler(handler)
logger.setLevel(logging.INFO)

# Now log something
logger.info('Log something informative.')
logger.warning('Log a warning.')
logger.error('Log an error.')
logger.critical('Log a critical error.')

More examples and configurations: https://docs.python.org/2.7/library/logging.html

Python 2.7 is not provided on CentOS 6 because Yum depends on Python 2.6. Simply installing 2.7 over 2.6 will break Yum. Thankfully, Red Hat provides us with an alternate installation method through their SCL project.

First, get Python 2.7 installed:
https://www.softwarecollections.org/en/scls/rhscl/python27/

Next, add the project path to the top of the script:

import sys
sys.path.insert(0,'/path/to/project')
sys.path.insert(0,'/path/to/project/env/lib/python2.7/site-packages')

Lastly, call the script via Cron:

cat /path/to/project/script.py | scl enable python27 python

Here's a custom manager for a class (Change) that gets all approved changes where "today's date" falls between the Change.start_date and Change.end_date.

# models.py
import datetime
...
# Custom Change Manager that only pulls active, approved changes
class ApprovedChanges(models.Manager):
    def get_queryset(self):
        today = datetime.date.today()
        return super(ApprovedChanges, self).get_queryset().filter(
            approved=True,
            start_date__lte=today,
            end_date__gte=today
        )

class Change(models.Model):
    ...
    approved_changes = ApprovedChanges()

In our view, we could then call all changes for "today" that have been approved:

# views.py
def todays_changes(self):
   changes = Changes.approved_changes.all()
   ...

Source: https://docs.djangoproject.com/en/stable/topics/db/managers/

# Simple List
', '.join([author.name for author in authors])

# Django
', '.join([author.name for author in book.author_set.all()])

Both will output "Author Name, Author 2 Name, ..., Author x Name"

I just started my first Django project with Python 3 in earnest. One of the first things I have noticed is that reload isn't available by default from the Django shell. It needs to be imported:

>>> from importlib import reload
>>> reload(<module>)

This became a pretty tedious process and didn't always reload the module as expected. So, instead of reloading the shell and the module to run a command, I moved the commands that I was running into a custom Django management command. No more reloading =)

The following code allows us to monkey patch the json() method that is returned by requests.post().

import requests
from mock import MagicMock


@mock.patch.object(requests, 'post')
def my_function(self, post):
    attrs = {'json.return_value': {'id': 1}}
    post.return_value = MagicMock(**attrs)

    r = requests.post()
    r.json()['id']

Source: https://docs.python.org/dev/library/unittest.mock.htmll#unittest.mock.NonCallableMock

from sqlalchemy.sql import text
from sqlalchemy import create_engine

import secrets


engine = create_engine(
    'mssql+pyodbc://{}:{}@MSSQL'.format(
        secrets.username, secrets.password
    )
)
conn = engine.connect()

s = text("SELECT * FROM users WHERE name = :name")
result = conn.execute(s, name=name).fetchall()
print result

For further information on installing and configuring unixODBC (requirement for pyodbc) on Linux:
https://github.com/mkleehammer/pyodbc/wiki/Connecting-to-SQL-Server-from-RHEL-or-Centos

Connecting to Microsoft SQL Server with pyodbc:
http://docs.sqlalchemy.org/en/latest/dialects/mssql.html#dialect-mssql-pyodbc-connect