Django Twitter Tag

This is the Pancentric fork of Django Twitter Tag. We added django-cacheback and Celery so that users never have to wait for us to fetch new tweets - fetching is done in a background task queue.

A django template tag to display user's recent tweets / search results. Version 1.0 uses Twitter API 1.1.

Basic features are limiting numbers of displayed tweets, filtering out replies and retweets. Library exposes each tweet json in template, adding extra attributes: html and datetime. First one makes urls, hashtags or twitter usernames clickable, juts like you expect them to be. Last one provides python datetime object to ease output in templates. Urls are expanded by default. Library handles twitter exceptions gracefully, returning last successful response.

Usage

• Load tag in your template like this:

  {% load twitter_tag %}
  

• Get user's (futurecolors in example) most recent tweets and store them in tweets variable:

  {% get_tweets for "futurecolors" as tweets %}
  

• Now you have a list of tweets in your template context, which you can iterate over like this:

  <ul>
  {% for tweet in tweets %}
      <li>{{ tweet.html|safe }}</li>
  {% endfor %}
  </ul>
  

See how it looks like on our site.

Installation

This app works with python 2.6, 2.7 and 3.3, Django 1.3-1.5.

Recommended way to install is pip:

pip install -e git+https://github.com/pancentric/django-twitter-tag.git#egg=django-twitter-tag

We use Redis as our Celery message broker so you probably want to install the following:

pip install django-celery-with-redis==3.0

Add twitter_tag to INSTALLED_APPS in settings.py:

INSTALLED_APPS = (...
                  'cacheback',
                  'djcelery',
                  'twitter_tag',
                  ...
                 )

Configuration

Twitter API 1.1 requires authentication for every request you make, so you have to provide some credentials for oauth dance to work. First, create an application, second, request access token on newly created app page. The process of obtaining a token is explained in detail in docs.

Here is an example of how your config might look like:

# settings.py
# Make sure to replace with your own values, theses are just made up

# Your access token: Access token
TWITTER_OAUTH_TOKEN = '91570701-BQMM5Ix9AJUC5JtM5Ix9DtwNAiaaYIYGN2CyPgduPVZKSX'
# Your access token: Access token secret
TWITTER_OAUTH_SECRET = 'hi1UiXm8rF4essN3HlaqMz7GoUvy3e4DsVkBAVsg4M'
# OAuth settings: Consumer key
TWITTER_CONSUMER_KEY = '3edIOec4uu00IGFxvQcwJe'
# OAuth settings: Consumer secret
TWITTER_CONSUMER_SECRET = 'YBD6GyFpvumNbNA218RAphszFnkifxR8K9h8Rdtq1A'

TWITTER_CACHE_TIMEOUT = 300

BROKER_URL = _get_env_variable('BROKER_URL', default='redis://localhost:6379/0')
CELERY_RESULT_BACKEND = BROKER_URL
CELERY_ACCEPT_CONTENT = ['pickle', 'json', 'msgpack', 'yaml']

Examples

You can specify number of tweets to show:

{% get_tweets for "futurecolors" as tweets limit 10 %}

To filter out tweet replies (that start with @ char):

{% get_tweets for "futurecolors" as tweets exclude "replies" %}

To ignore native retweets:

{% get_tweets for "futurecolors" as tweets exclude "retweets" %}

Or everything from above together:

{% get_tweets for "futurecolors" as tweets exclude "replies, retweets" limit 10 %}

Search tag (experimental)

You can search for tweets:

{% search_tweets for "python 3" as tweets limit 5 %}

Search api arguments are supported via key=value pairs:

{% search_tweets for "python 3" as tweets lang='eu' result_type='popular' %}

Relevant API docs for search.

Caching

There is no need to use any django cache template tags as suggested in the upstream project's documentation. You will want to configure CACHES in settings.py to be something like this so that the task jobs can put the results in the shared location:

CACHES = {
    'default': {
        'BACKEND': 'redis_cache.cache.RedisCache',
        'LOCATION': _get_env_variable('REDIS_LOCATION'),
        'OPTIONS': {
            'CLIENT_CLASS': 'redis_cache.client.DefaultClient',
        }
    }
}

Extra

Tweet's properties

get_tweets returns a list of tweets into context. Each tweets is a json dict, that has exactly the same attributes, as stated in API 1.1 docs, describing tweet json. Tweet's created timestamp is converted to python object and is available in templates:

{{ tweet.datetime|date:"D d M Y" }}

Tweet's html

Tweet also has extra html property, which contains tweet, formatted for html output with all needed links. Note, Twitter has guidelines for developers on how embeded tweets should look like.

Exception handling

Any Twitter API exceptions like 'Over capacity' are silenced and logged. Django cache is used internally to store last successful response in case twitter is down.

Going beyond

Since version 1.0 you can create your own template tags for specific twitter queries, not supported by this library. Simply inherit from twitter_tag.templatetags.twitter_tag.BaseTwitterTag and implement your own get_json method (tag syntax is contolled by django-classy-tags).

Development

To install development version, use pip install django-twitter-tag==dev

Tests

Run:

DJANGO_SETTINGS_MODULE = twitter_tag.test_settings python setup.py test