Enabling Foreign Language Support

[thebiggeek]

Hi

In India we have started receiving a new set of Spam, that uses badly formatted Hindi. I notice from https://www.lifewire.com/spamassassin-foreign-language-spam-1166250 that SA can do Hindi Spam too, and I wrote some rules to detect the Spam - but am not sure if this will cause the system to get overloaded, and wanted to check with you all on

(A) How do I enable Hindi support
(B) Since Hindi is not straight forward Ascii, how should I write REGEX Rules

Here are some examples

body S3BMS_BODY_133 /संलग्न फॉर्म की जांच करें और जानकारी से भी संपर्क करें/
describe S3BMS_BODY_133 Hindi Spam
score S3BMS_BODY_133 5

body S3BMS_BODY_134 /अटैच की गई फाइल देखें/
describe S3BMS_BODY_134 Hindi Spam
score S3BMS_BODY_134 5

 

[Stoiko Ivanov]

rules matching the complete mail's body can be quite expensive.
You could try adding the rules and watch the logs of how the processing time changes (pmg-smtp-filter logs for each mail how long the processing took) - additionally there you'll see if the rules work as expected and if SpamAssassin can directly work with UTF-8 characters inside rules
(haven't actively tested this myself so am not sure)

Else check the SpamAssassin documentation on writing rules;
https://cwiki.apache.org/confluence/display/SPAMASSASSIN/WritingRules

I hope this helps!

 

[thebiggeek]

Thank you @Stoiko Ivanov - Let me see and revert. I understand that Rules matching body are expensive, but they need to be done as spam is changing every day and with local language spam that is what we need to fix most for our network at the moment.