[Rspamd-Users] A question about rspamd learning process

[Gabriele Nencioni]

Hi,
just a question about rspamd learning process.

Are the rspamd headers (or any other spam headers) excluded from the
learing process?

I mean if I run the following commands:
cat message.eml | rspamc learn_spam
or
cat message.eml | rspamc learn_ham


and the message.eml contains something as following:
cat message.eml
...
X-Spam-Status: Yes, rspamdscore=10.15, required=10.00
X-Rspamd-Server: rspamdserver
X-Spamd-Result: default: True [10.27 / 10.00];
 ARC_NA(0.00)[];
 HAS_REPLYTO(0.00)[];
 R_DKIM_ALLOW(-0.20)[];
 BAYES_SPAM(7.17)[99.45%];
 FROM_HAS_DN(0.00)[];
 FROM_EXCESS_BASE64(1.50)[];
 R_SPF_ALLOW(-0.20)[];
 ZERO_FONT(0.20)[2];
 REPLYTO_DN_EQ_FROM_DN(0.00)[];
 PREVIOUSLY_DELIVERED(0.00)[];
 TO_DN_NONE(0.00)[];
 HAS_LIST_UNSUB(-0.50)[]; RCPT_COUNT_ONE(0.00)[1];
 MANY_INVISIBLE_PARTS(1.00)[10];
 RCVD_COUNT_THREE(0.00)[3];
 TO_MATCH_ENVRCPT_ALL(0.00)[];
 DKIM_TRACE(0.00)[];
 DMARC_POLICY_ALLOW(-0.50)[];
 REPLYTO_DOM_EQ_FROM_DOM(0.00)[];
 MIME_HTML_ONLY(0.20)[];
 RCVD_NO_TLS_LAST(0.10)[];
 FROM_EQ_ENVFROM(0.00)[];
 REPLYTO_EXCESS_BASE64(1.50)[]
...


does it contaminate the learning process?

Have I to remove any headers, for example coming from the last
(internal) mta? Would it be better?

Thanks in advance
Regards,
-- 
Gabriele Nencioni