Skip to content

My replacement for Google Reader, to produce a starred items feed. My copy publishes to http://callanbryant.co.uk/#Recommended Easter 2013.

License

Notifications You must be signed in to change notification settings

dvbportal/megafilter

 
 

Repository files navigation

This is a personal replacement for Google reader. It's simple and suits my workflow.

Articles are added to a circular FIFO queue, so that the latest articles always appear next. Articles can be:

  • Skipped (s) to be viewed later
  • Published (p) to a JSONP 'RSS' feed
  • Discarded (d) from the queue
  • Inspected (i) to view the article from the original website, with comments

Destructive actions can be undone (u).

The current article is synchronised across all devices, so you can continue reading on yor smartphone.

It's quite a nice use for your Raspberry Pi.

State: Very beta but working very well! I have migrated from Google reader to Megafilter.

screenshot

See my blog post about it.

sudo npm install -g megafilter
# copy subscriptions.xml and starred.json from your google reader takeout to here
megafilter -s subscriptions.xml --import-greader-starred starred.json
open http://localhost:8080

The idea is that you run this on your own server. You can specify a PORT via the environment variable or --port. Setcap can be used to run from port 80 without sudo; this is documented at the end of this file.

There will be a subscriptions manager soon. For the mean time, always start the server with -s subscriptions.xml. Note that you only need to import once!

You can specify --password <password> to require auth. A --username can be set, but this defaults to the executing user.

By default megafilter will save the published items to ~/megafilter-published.json. You can change this directory with --store-dir

When megafilter starts, it will check each feed within the first 2 minutes (this is so that servers don't get hammered) and after this, each feed will be checked at the average post interval. Articles that are posted after the initial check are added to the queue, so it may take a while for articles to appear. If you are excited and/or impatient like me, then you may use the new --backfill 5 option, to initially add articles that were new in the last 5 hours. Articles will take 2 minutes to appear.

RESTful JSON API

GET /next Get the next article. Returns article and pending count.

GET /current Get the current article (good for first load) returns article and pending count.

DELETE /queue/:id Discard an article from the queue.

PUT /publish/:id Publish an article by ID from the queue.

GET /published Given a count as parameter, return published articles. 0 Means all articles, unspecified means 50.

GET /pending Gives the current number of articles pending

DELETE /published/:id Delete an article from the published collection

POST /enqueue Remotely adds an article to the queue.

Article format

The same as the node-feedparser format.

  • title
  • description (frequently, the full article content)
  • summary (frequently, an excerpt of the article content)
  • link
  • origlink (when FeedBurner or Pheedo puts a special tracking url in the link property, origlink contains the original link)
  • date (most recent update)
  • pubdate (original published date)
  • author
  • guid (a unique identifier for the article)
  • comments (a link to the article's comments section)
  • image (an Object containing url and title properties)
  • categories (an Array of Strings)
  • source (an Object containing url and title properties pointing to the original source for an article; see the RSS Spec for an explanation of this element)
  • enclosures (an Array of Objects, each representing a podcast or other enclosure and having a url property and possibly type and length properties)
  • meta (an Object containing all the feed meta properties; especially handy when using the EventEmitter interface to listen to article emissions)

Published articles

Currently, megafilter does not create a published RSS feed. If that's needed by someone I'll add it. The reason I did not put it is because RSS is not cross-domain, and is annoying to parse with jQuery.

The /published JSONP feed is simply the latest 50 published articles in the format that node-feedparser produces.

If you want an example on how to integrate the published items, see my website engine.

I'll probably make this available as part of megafilter.

Running without sudo on port 80

If you want to host on port 80 under ubuntu on your account, you can install libcap2-bin then run setcap 'cap_net_bind_service=+ep' /usr/local/bin/node to give node access to port 80.

Your path to the node binary may vary, you can find it with which node

Acknowledgements

Megafilter would not have been possible without the following awesome projects:

About

My replacement for Google Reader, to produce a starred items feed. My copy publishes to http://callanbryant.co.uk/#Recommended Easter 2013.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages

  • JavaScript 99.2%
  • PHP 0.8%