Operations on the DSPAM directory. More...

Inheritance diagram for Dspam.DSpamDirectory:

Public Member Functions
def	__init__ (self, userdir)
	Initialize DSpamDirectory. More...

def	dspam_ctx (self, op, flags=0, user=None)
	Create dspam.ctx using configured defaults. More...

def	get_group (self, user)
	Return group user belongs to. More...

def	user_files (self, user)
	Set username and return common pathnames. More...

def	check_spam (self, user, txt, recipients=None, classify=False, quarantine=True, force_result=None)
	Check spaminess of a message. More...

def	write_web_stats (self, totals)
	Update dspam stats stored as text for the web interface. More...

def	add_spam (self, user, txt)
	Report a false negative. More...

def	false_positive (self, user, txt)
	Report a false positive. More...

Public Attributes
	userdir
	DSPAM home. More...

	groupfile
	Path of group definition file. More...

	log
	Logging method.

	headerchange
	Hook for changing mail headers in the MTA to mirror changes by Dspam. More...

	username
	The dspam user current being processed.

	probability
	The spam score from 0.0 to 1.0.

	factors
	The top tokens that determined the spam score.

	algorithms
	Default spam score algorithms to use.

	tokenizer
	Default tokenizer.

	training
	Training mode. More...

	totals
	DSPAM totals for user from last operation.

	result
	Classification from last check_spam.

	mbox
	Path of quarantine mailbox used by pydspam for user. More...

	lock
	Path of lock file used by libdspam. More...

	dspam_dict

	dspam_stats

	sigfile

Private Member Functions
def	_lognull (self, *msg)

def	_add_sig (self, txt, sig, recipients=None)
	Add signature key to message, and quarantine if spammy. More...

def	_feedback (self, user, txt, op, queue=False)

def	_innoc (self, user, sigs, op)

Detailed Description

Operations on the DSPAM directory.

Constructor & Destructor Documentation

◆ init()

def Dspam.DSpamDirectory.__init__	(	self,
		userdir
	)

Initialize DSpamDirectory.

The base directory is something like /var/lib/dspam.

Parameters

userdir the DSPAM base directory

Member Function Documentation

◆ _add_sig()

def Dspam.DSpamDirectory._add_sig	(	self,
		txt,
		sig,
		recipients = `None`
	)

private

Add signature key to message, and quarantine if spammy.

The results of the last check_spam as used.

Parameters

txt	the message bytes
sig	the signature
recipients	list of recipients for later delivery

References Dspam.add_signature_tag(), Dspam.DSpamDirectory.dspam_ctx(), Dspam.DSpamDirectory.factors, Dspam.DSpamDirectory.headerchange, Dspam.DSpamDirectory.log, Dspam.DSpamDirectory.mbox, Dspam.DSpamDirectory.probability, Dspam.put_signature(), Dspam.DSpamDirectory.result, and Dspam.DSpamDirectory.write_web_stats().

Referenced by Dspam.DSpamDirectory.check_spam().

◆ add_spam()

def Dspam.DSpamDirectory.add_spam	(	self,
		user,
		txt
	)

Report a false negative.

Tell DSPAM a message it though was innocent is actually spam. DSPAM looks for signature keys, and looks up stored signatures with them. It trains DSPAM with the signature, setting the source to DSS_ERROR. If no signature is found, it adds the spam as a spam "corpus".

Parameters

user	str - the DSPAM user
txt	bytes - the spam message

References Dspam.DSpamDirectory._feedback(), and Dspam.DSpamDirectory.probability.

◆ check_spam()

def Dspam.DSpamDirectory.check_spam	(	self,
		user,
		txt,
		recipients = `None`,
		classify = `False`,
		quarantine = `True`,
		force_result = `None`
	)

Check spaminess of a message.

Parameters

user	the dspam user (email account)
txt	the message as bytes collected from the MTA
recipients	If provided, a list of recipients to record in quarantined messages to assist later delivery.
classify
quarantine	Add messages classified as spam to mbox quarantine if true.
force_result	train as this result

Returns: tagged message, or None if message was quarantined

References Dspam.DSpamDirectory._add_sig(), Dspam.DSpamDirectory._innoc(), Dspam.convert_eol(), Dspam.DSpamDirectory.dspam_ctx(), Dspam.DSpamDirectory.factors, Dspam.DSpamDirectory.lock, Dspam.DSpamDirectory.probability, Dspam.DSpamDirectory.result, and Dspam.DSpamDirectory.user_files().

◆ dspam_ctx()

def Dspam.DSpamDirectory.dspam_ctx	(	self,
		op,
		flags = `0`,
		user = `None`
	)

Create dspam.ctx using configured defaults.

E.g.

 with self.dspam_ctx(dspam.DSM_CLASSIFY) as ds:
   ds.process(txt)
   self.result = ds.result

Parameters

op	dspam operation mode: one of dspam.DSM_*
flags	dspam operation flags: set of dspam.DSF_*
user	str - a different user, e.g. for innoculation

Returns: context manager for dspam.ctx

References Dspam.DSpamDirectory.algorithms, Dspam.DSpamDirectory.tokenizer, Dspam.DSpamDirectory.totals, Dspam.DSpamDirectory.training, Dspam.DSpamDirectory.userdir, and Dspam.DSpamDirectory.username.

Referenced by Dspam.DSpamDirectory._add_sig(), Dspam.DSpamDirectory.check_spam(), and Dspam.DSpamDirectory.write_web_stats().

◆ false_positive()

def Dspam.DSpamDirectory.false_positive	(	self,
		user,
		txt
	)

Report a false positive.

Tell DSPAM a message it though was spam is actually innocent. DSPAM looks for signature keys, and looks up stored signatures with them. It trains DSPAM with the signature, setting the source to DSS_ERROR. If no signature is found, it adds the message as an innocent "corpus" (DSS_CORPUS).

Parameters