miroir de
https://github.com/rspamd/rspamd.git
synchronisé 2024-08-28 05:54:33 +02:00
Rapid spam filtering system.
cgi | ||
compat | ||
conf | ||
contrib | ||
debian | ||
doc | ||
freebsd | ||
lib | ||
linux | ||
perl | ||
src | ||
test | ||
utils | ||
.hgtags | ||
ChangeLog | ||
CMakeLists.txt | ||
config.h.in | ||
LICENSE | ||
Makefile.in | ||
README |
DESCRIPTION =========== Rspamd is a complex spam filter that allows to estimate messages by many rules, statistical data and custom services like URL black lists. Each message is estimated by rspamd and got so called 'spam score'. According to spam score and user's settings rspamd send recommended action for this message to MTA. Rspamd has own unique features among spam filters: * event driven architecture allowing to process many messages at a time * flexible syntax of rules allowing to write rules in lua language * a lot of plugins and rules shipped with rspamd distribution * highly optimized mail processing advanced statistic All these features allow rspamd to process messages fast and make good results in spam filtering. INSTALLATION ============ Rspamd runs on a Unix like operational systems. FreeBSD users can use ports collection (mail/rspamd) for rspamd installation. Users of other OSes should use sources to build and install rspamd. You can obtain rspamd sources at bitbucket download page: https:bitbucket.org/vstakhov/rspamd/downloads. Build requirements ================== Rspamd requires several 3-rd party software to build and run: * libevent - asynchronous event library * glib - common purposes library * gmime - mime parser * lua - extendable scripting language * cmake - advanced software build system You can either install them from sources or (recommended) install using package manager of your system. Build process ============= Building of rspamd is quite simple: $ cmake . $ make # make install After installation binaries, rules, plugins and a sample configuration will be installed in the target directories (prefixed by /usr/local by default). To start working with rspamd you should do several steps: 1. Copy a sample configuration $PREFIX/etc/rspamd.xml.sample to $PREFIX/etc/rspamd.xml 2. Edit rspamd.xml according to your system (described later). 3. Make a directory for rspamd pid file and data (/var/run/rspamd by default) and make rspamd user (nobody by default) as owner of rspamd data directory. 4. Make a directory for rspamd logs (or setup syslog to accept rspamd log messages) 5. Install start script to a proper place (this step is done when installing from FreeBSD ports) 6. Start rspamd using start script If start script is not suitable for your system (now rspamd shipped with start script for FreeBSD, Debian and RedHat like operational systems) you should write a start script based on your system's documentation. FURTHER ACTIONS =============== You can improve quality of rspamd by learning its statistic module. The easiest way to do it is to use rspamc client (you can setup a custom email alias to pipe messages to rspamc) $ rspamc -P 'q1' -c bayes learn_spam [ file1 [file2 [...]]] $ rspamc -P 'q1' -c bayes learn_ham [ file1 [file2 [...]]] Note: you should consider to change default controller's password 'q1' to yours one specified in controller section of configuration. Also a system administrator may want to customize rule's weights or actions thresholds. This can be easily done by editing metric section in the configuration file. For writing new rules you can examine the main rspamd documentation and lua api gude and reference. REFERENCES ========== Home site: https://bitbucket.org/vstakhov/rspamd Downloads: https://bitbucket.org/vstakhov/rspamd/downloads Wiki: https://bitbucket.org/vstakhov/rspamd/wiki/