rssc

#+TITLE: rssc - rss customizer
#+OPTIONS: num:nil
#+OPTIONS: toc:nil
#+OPTIONS: ^:nil


rssc provides a real-time, self-hostable regex-oriented customization service for atom,
rss and json feeds.

* Usage
rssc uses get request parameters to read user's preferences. following are the
supported parameters:
+ ~src~: the source URL of the feeds. Example: https://hnrss.org/newest
+ ~t~: the type of desired source, can be ~rss~, ~atom~ or ~json~.
+ ~descriptionf~: regular expression filter to filter feeds based on the description property.
+ ~titlef~: regular expression filter to filter feeds based on the title property.
+ ~contentf~: regular expression filter to filter feeds based on the content property.
+ ~net~: boolean (1 or 0) (default to 0 when omitted) whether to use the .NET
  engine regex instead of Go's.

All regex should be valid [[https://github.com/google/re2/wiki/Syntax][RE2 regex syntax]] (in case of ~net~, see [[https://learn.microsoft.com/en-us/dotnet/standard/base-types/regular-expression-language-quick-reference][MS' Regular Expression Language]]).
* Examples
Replace https://rssc.fly.dev/ in the following examples with your instance of
rssc (can be a ~localhost:8080~ if self-hosted).

+ Get Edward Snowden tweets that mention the CIA:

  https://rssc.fly.dev/rss?src=https://nitter.lacontrevoie.fr/Snowden/rss&descriptionf=CIA
+ Get all new feeds that contain the word "emacs" but not the word "vim", and
  not posted by user lr0 (hnrss.org does not provide a creator/author property, so the author filter here is only for a demonstration purpose.):

  [[https://rssc.fly.dev/rss?src=https://hnrss.org/newest?q=emacs&titlef=^(?!.*\bvim\b).*\bEmacs\b.*&net=1&authorf=^(?!.*\blr0\b).*$][https://rssc.fly.dev/rss?src=https://hnrss.org/newest?q=emacs&titlef=^(?!.*\bvim\b).*\bEmacs\b.*&net=1&authorf=^(?!.*\blr0\b).*$]]

  Note that ~net~ is used here because RE2 does not support lookarounds.
+ Get only posts about OpenAI from 4chan's g board:

  [[https://rssc.fly.dev/rss?src=https://boards.4channel.org/g/index.rss&titlef=Open\s?AI|OpenAI][https://rssc.fly.dev/rss?src=https://boards.4channel.org/g/index.rss&titlef=Open\s?AI|OpenAI]]
+ Get geopolitical updates from your favorite hacker and skip all free software
  movement goody propaganda:

  [[https://rssc.fly.dev/rss?src=https://stallman.org/rss/rss.xml&contentf=^(?=.*(?:palestine|syria|egypt|iraq|israel|algeria|morocco))(?!.*(?:linux|gnu|software|programming|program)).*$&net=1][https://rssc.fly.dev/rss?src=https://stallman.org/rss/rss.xml&contentf=^(?=.*(?:palestine|syria|egypt|iraq|israel|algeria|morocco))(?!.*(?:linux|gnu|software|programming|program)).*$&net=1]]
+ Get BBC middle east updates that's only related to Palestine:

  [[https://rssc.fly.dev/rss?src=http://feeds.bbci.co.uk/news/world/middle_east/rss.xml&titlef=(?i)palestine|palestinian|gaza][https://rssc.fly.dev/rss?src=http://feeds.bbci.co.uk/news/world/middle_east/rss.xml&titlef=(?i)palestine|palestinian|gaza]]
* Notes
+ RE2 is much safer that the .NET Regex, that's to say, all the good features
  that .NET engine enables you to use come with a risk cost and ends up
  enabling [[https://github.com/dlclark/regexp2#catastrophic-backtracking-and-timeouts][catastrophic backtracking]], therefore a timeout of 5 seconds is currently
  enabled on fly.dev instance when using ~net=1~.