rssparser.lisp

A Web-to-RSS parser in Common Lisp.

What does it do?

This software was written because a disappointing number of websites still does not have an RSS or Atom feed so I could subscribe to their updates, e.g. the KiTTY website. The script tries to find new articles on any website according to given criteria (CSS selectors) and parse them into a valid RSS feed so I can subscribe to them in my usual RSS feed reader.

Screenshot

Syntax

• chmod +x rssparser.lisp, then:
  • ./rssparser.lisp add *Title* *URL* *EntrySelector* *TitleSelector* [*ContentSelector*]
  • ./rssparser.lisp delete *ID*
  • ./rssparser.lisp list
  • ./rssparser.lisp export *ID*

Run a simple web interface on port 5000:

• ./rssparser.lisp webserver

Cronjob or manual feed creation command:

• ./rssparser.lisp parse

Supported selectors are all valid CSS selectors. If you don't specify a ContentSelector when adding a new feed, rssparser.lisp will use an empty item body.

Example

If you want to subscribe to the KiTTY website, you can either use the web interface or perform the following commands:

% ./rssparser.lisp add "KiTTY" "https://www.9bis.net/kitty/?action=news&zone=en" ".news" "h1" ""
Success!

% ./rssparser.lisp parse

% ./rssparser.lisp list
1 feed is set up:

ID: 23  Title:        KiTTY
        URL:          https://www.9bis.net/kitty/?action=news&zone=en
        Last success: Sun, 27 Mar 2016 17:54:18 +0200

By default, the KiTTY website feed will be stored as feeds/feed23.xml then.

Requirements

You'll need the files from this repository and SBCL with Quicklisp set up. SQLite3 should be available. Also, you should create a folder where your feed files should be created (./feeds by default). Hard links are allowed.

Packages

Usually, Quicklisp should install the required packages for you. If you want to install them manually, rssparser.lisp currently requires these:

• datafly
• hunchentoot
• cl-who
• parenscript
• smackjack
• lass
• cl-ppcre
• dexador
• clss
• plump
• plump-sexp
• local-time
• xml-emitter

SQLite schema

The feeds.db file has the following schema:

CREATE TABLE feeds (
  id integer primary key autoincrement,
  feedtitle text not null,
  url text not null,
  entryselector text not null,
  titleselector text not null,
  contentselector text not null,
  lastsuccess integer
);

CREATE TABLE entries (
  id integer primary key autoincrement,
  feedid integer,
  title text not null,
  contents blob,
  url text not null,
  timestamp integer
);

Exporting feeds into a new database

If you want to transfer one or more of your stored feeds into a new database, that's what the export command is for:

% ./rssparser.lisp export 23
Execute this SQL command to add this feed to a new database:
  INSERT INTO feeds ('feedtitle', 'url', 'entryselector', 'titleselector', 'contentselector') VALUES ('KiTTy', 'https://www.9bis.net/kitty/?action=news&zone=en', '.news', 'h1', '');

Configuration

You can set a couple of parameters in the config.lisp file:

• +database-file+: The SQLite database file. (Default: feeds.db.) Note that this file needs to be accessible for the RSS parser to work!
• +feed-folder+: The folder where the feed files should be created. (Default: feeds/.) The script needs to be able to create files there; it checks its permissions automatically and informs you if it needs some help.
• +max-items-per-feed+: The maximum number of items per feed. (Default: 50.)
• +feed-cleanup+: If set to t (which is the default value), the entries table will automatically be purged from old entries (only 2 * +max-items-per-feed+ are kept). Set this to nil if you want to bloat your database.
• +remove-dead-feeds+: If set to t, a website which is not reachable anymore will automatically be removed from your feed list. The parser will inform you of that so if you run rssparser.lisp as a cronjob, you'll see what happened in your logfiles.
• +webserver-port+: The port to run the webserver on when rssparser.lisp webserver is executed. It should be available through your firewall. (Default: 5000.)