Skip to content

An RSS/Atom/JSONFeed reading + storing library for Django

License

Notifications You must be signed in to change notification settings

xurble/django-feed-reader

Repository files navigation

Django Feed Reader

This is a simple Django module to allow you subscribe to RSS (and other) feeds.

This app has no UI, it just reads and stores the feeds for you to use as you see fit.

This app builds on top of the FeedParser library to provide feed management, storage, scheduling etc.

Features

  • Consumes RSS, Atom and JSONFeed feeds.
  • Parses feeds liberally to try and accomodate simple errors.
  • Will attempt to bypass Cloudflare protection of feeds
  • Supports enclosure (podcast) discovery
  • Automatic feed scheduling based on frequency of updates

Installation

django-feed-reader is written in Python 3 and supports Django 2.2+

  • pip install django-feed-reader
  • Add feeds to your INSTALLED_APPS
  • Setup some values in settings.py so that your feed reader politely announces itself to servers
    • Set FEEDS_USER_AGENT to the name and (optionally version) of your service e.g. "ExampleFeeder/1.2"
    • Set FEEDS_SERVER to preferred web address of your service so that feed hosts can locate you if required e.g. https://example.com
  • Setup a mechanism to periodically refresh the feeds (see below)

Optional Settings

  • FEEDS_VERIFY_HTTPS (Default True)
    • Older versions of this library did not verify https connections when fetching feeds. Set this to False to revert to the old behaviour.
  • KEEP_OLD_ENCLOSURES (Default False)
    • Some feeds (particularly podcasts with Dynamic Ad Insertion) will change their enclosure urls between reads. By default, old enclosures are deleted and replaced with new ones. Set this to true, to retain old enclosures - they will have their is_current flag set to False
  • SAVE_JSON (Default False)
    • If set, Sources and Posts will store a JSON representation of the all the data retrieved from the feed so that uncommon or custom attributes can be retrieved. Caution - this will dramatically increase tha amount of space used in your database.
  • DRIPFEED_KEY (Default None)
    • If set to a valid Dripfeed API Key, then feeds that are blocked by Cloudflare will be automatically polled via Dripfeed instead.

Basic Models

A feed is represented by a Source object which has (among other things) a feed_url.

To start reading a feed, simply create a new Source with the desired feed_url

Source objects have Post children which contain the content.

A Post may have Enclosure (or more) which is what podcasts use to send their audio. The app does not download enclosures, if you want to do that you will need to do that in your project using the url provided.

Refreshing feeds

To conserve resources with large feed lists, the module will adjust how often it polls feeds based on how often they are updated. The fastest it will poll a feed is every hour. The slowest it will poll is every 24 hours.

Sources that don't get updated are polled progressively more slowly until the 24 hour limit is reached. When a feed changes, its polling frequency increases.

You will need to decided how and when to run the poller. When the poller runs, it checks all feeds that are currently due. The ideal frequency to run it is every 5 - 10 minutes.

Polling with cron

Set up a job that calls python manage.py refreshfeeds on your desired schedule.

Be careful to ensure you're running out of the correct directory and with the correct python environment.

Polling with celery

Create a new celery task and schedule in your app (see the celery documentation for details). Your tasks.py should look something like this:

from celery import shared_task
from feeds.utils import update_feeds

@shared_task
def get_those_feeds():

  # the number is the max number of feeds to poll in one go
  update_feeds(30)

Tracking read/unread state of feeds

There are two ways to track the read/unread state of feeds depending on your needs.

Single User Installations

If your usage is just for a single user, then there are helper methods on a Source to track your read state.

All posts come in unread. You can get the current number of unread posts from Source.unread_count.

To get a ResultSet of all the unread posts from a feed call Source.get_unread_posts

To mark all posts on a fed as read call Source.mark_read

To get all of the posts in a feed regardless of read status, a page at a time call Source.get_paginated_posts which returns a tuple of (Posts, Paginator)

Multi-User Installations

To allow multiple users to follow the same feed with individual read/unread status, create a new Subscription for that Source and User.

Subscription has the same helper methods for retrieving posts and marking read as Source.

You can also arrange feeds into a folder-like hierarchy using Subscriptions. Every Subscription has an optional parent. Subscriptions with a None parent are considered at the root level. By convention, Subscriptions that are acting as parent folders should have a None source

Subscriptions have a name field which by convention should be a display name if it is a folder or the name of the Source it is tracking. However this can be set to any value if you want to give a personally-meaningful name to a feed who's name is cryptic.

There are two helper methods in the utils module to help manage subscriptions as folders. get_subscription_list_for_user will return all Subscriptions for a User where the parent is None. get_unread_subscription_list_for_user will do the same but only returns Subscriptions that are unread or that have unread children if they are a folder.

Cloudflare Busting

django-feed-reader has Dripfeed support built in. If a feed becomes blocked by Cloudflare it can be polled via Dripfeed instead. This requires a Dripfeed account and API key.

For more details see the full documentation.