Digest Generation

[benniledl]

Hi all :)

I was wondering if Pyzor currently accounts for small personalizations in emails, specifically the recipient's account name (the local part before the @domain.com). If not, I think adjusting how the digest is generated could improve detection accuracy.

To improve detection accuracy, Pyzor should:

Extract the account name (local part of the recipient's email address) before removing headers.
Remove the extracted account name from the message body before generating the digest.

This ensures that minor personalizations in the spam mails do not affect Pyzor’s ability to detect spam messages.

[xrat]

Pyzor ignores the "To:" header.

$ pyzor digest < tmp.eml
6386969e29423a77f085544d78f88abf627a2a3c
$ wc -l tmp.eml
578 tmp.eml
$ grep -Ev 'To: ' tmp.eml | wc -l
577
$ grep -Ev 'To: ' tmp.eml | pyzor digest
6386969e29423a77f085544d78f88abf627a2a3c