PIP has a habit of re-downloading same packages every time you hit pip install package_name. That's not so cool, especially, when you deploy packages from requirements.txt and one of the packages failed to build, restarting pip install -r requirements.txt would lead pip to redownload all the packages again.

Thankfully there's a way to fix this easily, create a configuration file named ~/.pip/pip.conf, and add the following contents:

In Django, if your model has Filefield / Imagefield, when it's get deleted it doesn't delete attached files by default. To fix that you need to do two steps:

1) Add filecleanup function, that'll take care of deleting files attached to a model:

I had a case recently, where I needed to add custom data to the node display and wanted this data to behave like a field, however the data itself didn't belong to a field. By "behaving like a field" I mean you can that field at node display settings and able to control it's visibility, label and weight by dragging and dropping that field.

So, as you may have undestood, hook_preprocess_node / node_view_alter approach alone wasn't enough.

Very often we can have multiple Django apps each running celery and all this is installed on a single server. How do we isolate these separate app's celery workers? The answer is: Simple, if you use Rabbitmq as your broker.

Rabbitmq lets us to add multiple virtualhosts, so we can easily separate celery queues and the use different broker_url per our app, let's start:

Today while moving Django/PostgreSQL site between servers and importing an SQL dump into the new location I got a surprising error "ERROR: must be owner of extension plpgsql". I'm already used to PostgreSQL quirks so after a quick search I came to the following solution:

Login as postgres user and do:

If you want to access data from request object in your model form these two mixins will help you to do it easily:

Current version on Haystack (2.10) combined with Elasticsearch produces weird results when you want to sort search results on charfield, the issue has been reported here https://github.com/toastdriven/django-haystack/issues/866 and here: https://github.com/toastdriven/django-haystack/issues/569

I'm a big fan of fighting with Drupal's inefficiencies and bottlenecks. Most of these come from contrib modules. Everytime we install a contrib module we should be ready for surprises which come on board with the module.

To make haystack search pager place nicelly with other get parameters, we can do something like this:

To make Django messages look nice in your bootstrap 3 template you need two things:


Subscribe to TimOnWeb RSS