code.delx.au - gnu-emacs/blob - lisp/mh-e/mh-junk.el

   1 ;;; mh-junk.el --- Interface to anti-spam measures
   2
   3 ;; Copyright (C) 2003, 2004, 2005, 2006 Free Software Foundation, Inc.
   4
   5 ;; Author: Satyaki Das <satyaki@theforce.stanford.edu>,
   6 ;;         Bill Wohler <wohler@newt.com>
   7 ;; Maintainer: Bill Wohler <wohler@newt.com>
   8 ;; Keywords:  mail, spam
   9
  10 ;; This file is part of GNU Emacs.
  11
  12 ;; GNU Emacs is free software; you can redistribute it and/or modify
  13 ;; it under the terms of the GNU General Public License as published by
  14 ;; the Free Software Foundation; either version 2, or (at your option)
  15 ;; any later version.
  16
  17 ;; GNU Emacs is distributed in the hope that it will be useful,
  18 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
  19 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  20 ;; GNU General Public License for more details.
  21
  22 ;; You should have received a copy of the GNU General Public License
  23 ;; along with GNU Emacs; see the file COPYING.  If not, write to the
  24 ;; Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
  25 ;; Boston, MA 02110-1301, USA.
  26
  27 ;;; Commentary:
  28
  29 ;; Spam handling in MH-E.
  30
  31 ;;; Change Log:
  32
  33 ;;; Code:
  34
  35 (eval-when-compile (require 'mh-acros))
  36 (mh-require-cl)
  37 (require 'mh-e)
  38
  39 ;; Interactive functions callable from the folder buffer
  40 ;;;###mh-autoload
  41 (defun mh-junk-blacklist (range)
  42   "Blacklist RANGE as spam.
  43
  44 This command trains the spam program in use (see the option
  45 `mh-junk-program') with the content of RANGE and then handles the
  46 message(s) as specified by the option `mh-junk-disposition'.
  47
  48 Check the documentation of `mh-interactive-range' to see how RANGE is
  49 read in interactive use.
  50
  51 For more information about using your particular spam fighting
  52 program, see:
  53
  54   - `mh-spamassassin-blacklist'
  55   - `mh-bogofilter-blacklist'
  56   - `mh-spamprobe-blacklist'"
  57   (interactive (list (mh-interactive-range "Blacklist")))
  58   (let ((blacklist-func (nth 1 (assoc mh-junk-choice mh-junk-function-alist))))
  59     (unless blacklist-func
  60       (error "Customize `mh-junk-program' appropriately"))
  61     (let ((dest (cond ((null mh-junk-disposition) nil)
  62                       ((equal mh-junk-disposition "") "+")
  63                       ((eq (aref mh-junk-disposition 0) ?+)
  64                        mh-junk-disposition)
  65                       ((eq (aref mh-junk-disposition 0) ?@)
  66                        (concat mh-current-folder "/"
  67                                (substring mh-junk-disposition 1)))
  68                       (t (concat "+" mh-junk-disposition)))))
  69       (mh-iterate-on-range msg range
  70         (message "Blacklisting message %d..." msg)
  71         (funcall (symbol-function blacklist-func) msg)
  72         (message "Blacklisting message %d...done" msg)
  73         (if (not (memq msg mh-seen-list))
  74             (setq mh-seen-list (cons msg mh-seen-list)))
  75         (if dest
  76             (mh-refile-a-msg nil (intern dest))
  77           (mh-delete-a-msg nil)))
  78       (mh-next-msg))))
  79
  80 ;;;###mh-autoload
  81 (defun mh-junk-whitelist (range)
  82   "Whitelist RANGE as ham.
  83
  84 This command reclassifies the RANGE as ham if it were incorrectly
  85 classified as spam (see the option `mh-junk-program'). It then
  86 refiles the message into the \"+inbox\" folder.
  87
  88 Check the documentation of `mh-interactive-range' to see how
  89 RANGE is read in interactive use."
  90   (interactive (list (mh-interactive-range "Whitelist")))
  91   (let ((whitelist-func (nth 2 (assoc mh-junk-choice mh-junk-function-alist))))
  92     (unless whitelist-func
  93       (error "Customize `mh-junk-program' appropriately"))
  94     (mh-iterate-on-range msg range
  95       (message "Whitelisting message %d..." msg)
  96       (funcall (symbol-function whitelist-func) msg)
  97       (message "Whitelisting message %d...done" msg)
  98       (mh-refile-a-msg nil (intern mh-inbox)))
  99     (mh-next-msg)))
 100
 101 \f
 102
 103 ;; Spamassassin Interface
 104
 105 (defvar mh-spamassassin-executable (executable-find "spamassassin"))
 106 (defvar mh-sa-learn-executable (executable-find "sa-learn"))
 107
 108 (defun mh-spamassassin-blacklist (msg)
 109   "Blacklist MSG with SpamAssassin.
 110
 111 SpamAssassin is one of the more popular spam filtering programs.
 112 Get it from your local distribution or from
 113 http://spamassassin.org/.
 114
 115 To use SpamAssassin, add the following recipes to
 116 \".procmailrc\":
 117
 118     MAILDIR=$HOME/`mhparam Path`
 119
 120     # Fight spam with SpamAssassin.
 121     :0fw
 122     | spamc
 123
 124     # Anything with a spam level of 10 or more is junked immediately.
 125     :0:
 126     * ^X-Spam-Level: ..........
 127     /dev/null
 128
 129     :0:
 130     * ^X-Spam-Status: Yes
 131     spam/.
 132
 133 If you don't use \"spamc\", use \"spamassassin -P -a\".
 134
 135 Note that one of the recipes above throws away messages with a
 136 score greater than or equal to 10. Here's how you can determine a
 137 value that works best for you.
 138
 139 First, run \"spamassassin -t\" on every mail message in your
 140 archive and use Gnumeric to verify that the average plus the
 141 standard deviation of good mail is under 5, the SpamAssassin
 142 default for \"spam\".
 143
 144 Using Gnumeric, sort the messages by score and view the messages
 145 with the highest score. Determine the score which encompasses all
 146 of your interesting messages and add a couple of points to be
 147 conservative. Add that many dots to the \"X-Spam-Level:\" header
 148 field above to send messages with that score down the drain.
 149
 150 In the example above, messages with a score of 5-9 are set aside
 151 in the \"+spam\" folder for later review. The major weakness of
 152 rules-based filters is a plethora of false positives so it is
 153 worthwhile to check.
 154
 155 If SpamAssassin classifies a message incorrectly, or is unsure,
 156 you can use the MH-E commands \\[mh-junk-blacklist] and
 157 \\[mh-junk-whitelist].
 158
 159 The command \\[mh-junk-blacklist] adds a \"blacklist_from\" entry
 160 to \"~/spamassassin/user_prefs\", deletes the message, and sends
 161 the message to the Razor, so that others might not see this spam.
 162 If the \"sa-learn\" command is available, the message is also
 163 recategorized as spam.
 164
 165 The command \\[mh-junk-whitelist] adds a \"whitelist_from\" rule
 166 to the \"~/.spamassassin/user_prefs\" file. If the \"sa-learn\"
 167 command is available, the message is also recategorized as ham.
 168
 169 Over time, you'll observe that the same host or domain occurs
 170 repeatedly in the \"blacklist_from\" entries, so you might think
 171 that you could avoid future spam by blacklisting all mail from a
 172 particular domain. The utility function
 173 `mh-spamassassin-identify-spammers' helps you do precisely that.
 174 This function displays a frequency count of the hosts and domains
 175 in the \"blacklist_from\" entries from the last blank line in
 176 \"~/.spamassassin/user_prefs\" to the end of the file. This
 177 information can be used so that you can replace multiple
 178 \"blacklist_from\" entries with a single wildcard entry such as:
 179
 180     blacklist_from *@*amazingoffersdirect2u.com
 181
 182 In versions of SpamAssassin (2.50 and on) that support a Bayesian
 183 classifier, \\[mh-junk-blacklist] uses the program \"sa-learn\"
 184 to recategorize the message as spam. Neither MH-E, nor
 185 SpamAssassin, rebuilds the database after adding words, so you
 186 will need to run \"sa-learn --rebuild\" periodically. This can be
 187 done by adding the following to your crontab:
 188
 189     0 * * * *   sa-learn --rebuild > /dev/null 2>&1"
 190   (unless mh-spamassassin-executable
 191     (error "Unable to find the spamassassin executable"))
 192   (let ((current-folder mh-current-folder)
 193         (msg-file (mh-msg-filename msg mh-current-folder))
 194         (sender))
 195     (save-excursion
 196       (message "Reporting message %d..." msg)
 197       (mh-truncate-log-buffer)
 198       (call-process mh-spamassassin-executable msg-file mh-log-buffer nil
 199                     ;;"--report" "--remove-from-whitelist"
 200                     "-r" "-R")          ; spamassassin V2.20
 201       (when mh-sa-learn-executable
 202           (message "Recategorizing this message as spam...")
 203           (call-process mh-sa-learn-executable msg-file mh-log-buffer nil
 204                         "--single" "--spam" "--local" "--no-rebuild"))
 205       (message "Blacklisting message %d..." msg)
 206       (set-buffer (get-buffer-create mh-temp-buffer))
 207       (erase-buffer)
 208       (call-process (expand-file-name mh-scan-prog mh-progs)
 209                     nil mh-junk-background nil
 210                     (format "%s" msg) current-folder
 211                     "-format" "%<(mymbox{from})%|%(addr{from})%>")
 212       (goto-char (point-min))
 213       (if (search-forward-regexp "^\\(.+\\)$" nil t)
 214           (progn
 215             (setq sender (match-string 0))
 216             (mh-spamassassin-add-rule "blacklist_from" sender)
 217             (message "Blacklisting message %d...done" msg))
 218         (message "Blacklisting message %d...not done (from my address)" msg)))))
 219
 220 (defun mh-spamassassin-whitelist (msg)
 221   "Whitelist MSG with SpamAssassin.
 222
 223 The \\[mh-junk-whitelist] command adds a \"whitelist_from\" rule to
 224 the \"~/.spamassassin/user_prefs\" file. If the \"sa-learn\" command
 225 is available, the message is also recategorized as ham.
 226
 227 See `mh-spamassassin-blacklist' for more information."
 228   (unless mh-spamassassin-executable
 229     (error "Unable to find the spamassassin executable"))
 230   (let ((msg-file (mh-msg-filename msg mh-current-folder))
 231         (show-buffer (get-buffer mh-show-buffer))
 232         from)
 233     (save-excursion
 234       (set-buffer (get-buffer-create mh-temp-buffer))
 235       (erase-buffer)
 236       (message "Removing spamassassin markup from message...")
 237       (call-process mh-spamassassin-executable msg-file mh-temp-buffer nil
 238                     ;; "--remove-markup"
 239                     "-d")               ; spamassassin V2.20
 240       (if show-buffer
 241           (kill-buffer show-buffer))
 242       (write-file msg-file)
 243       (when mh-sa-learn-executable
 244         (message "Recategorizing this message as ham...")
 245         (call-process mh-sa-learn-executable msg-file mh-temp-buffer nil
 246                       "--single" "--ham" "--local --no-rebuild"))
 247       (message "Whitelisting message %d..." msg)
 248       (setq from
 249             (car (mh-funcall-if-exists
 250                   ietf-drums-parse-address (mh-get-header-field "From:"))))
 251       (kill-buffer nil)
 252       (unless (or (null from) (equal from ""))
 253         (mh-spamassassin-add-rule "whitelist_from" from))
 254       (message "Whitelisting message %d...done" msg))))
 255
 256 (defun mh-spamassassin-add-rule (rule body)
 257   "Add a new rule to \"~/.spamassassin/user_prefs\".
 258 The name of the rule is RULE and its body is BODY."
 259   (save-window-excursion
 260     (let* ((line (format "%s\t%s\n" rule body))
 261            (case-fold-search t)
 262            (file (expand-file-name "~/.spamassassin/user_prefs"))
 263            (buffer-exists (find-buffer-visiting file)))
 264       (find-file file)
 265       (if (not (search-forward (format "\n%s" line) nil t))
 266           (progn
 267             (goto-char (point-max))
 268             (insert (if (bolp) "" "\n") line)
 269             (save-buffer)))
 270       (if (not buffer-exists)
 271           (kill-buffer nil)))))
 272
 273 (defun mh-spamassassin-identify-spammers ()
 274   "Identify spammers who are repeat offenders.
 275
 276 This function displays a frequency count of the hosts and domains
 277 in the \"blacklist_from\" entries from the last blank line in
 278 \"~/.spamassassin/user_prefs\" to the end of the file. This
 279 information can be used so that you can replace multiple
 280 \"blacklist_from\" entries with a single wildcard entry such as:
 281
 282     blacklist_from *@*amazingoffersdirect2u.com"
 283   (interactive)
 284   (let* ((file (expand-file-name "~/.spamassassin/user_prefs"))
 285          (domains (make-hash-table :test 'equal)))
 286     (find-file file)
 287     ;; Only consider entries between last blank line and end of file.
 288     (goto-char (1- (point-max)))
 289     (search-backward-regexp "^$")
 290     ;; Perform frequency count.
 291     (save-excursion
 292       (while (search-forward-regexp "^blacklist_from\\s-*\\(.*\\)@\\(.*\\)$"
 293                                     nil t)
 294         (let ((host (match-string 2))
 295               value)
 296           ;; Remove top-level-domain from hostname.
 297           (setq host (cdr (reverse (split-string host "\\."))))
 298           ;; Add counts for each host and domain part.
 299           (while host
 300             (setq value (gethash (car host) domains))
 301             (setf (gethash (car host) domains) (1+ (if (not value) 0 value)))
 302             (setq host (cdr host))))))
 303
 304     ;; Output
 305     (delete-other-windows)
 306     (pop-to-buffer (get-buffer-create "*MH-E Spammer Frequencies*"))
 307     (erase-buffer)
 308     (maphash '(lambda (key value) ""
 309                 (if (> value 2)
 310                     (insert (format "%s %s\n" key value))))
 311              domains)
 312     (sort-numeric-fields 2 (point-min) (point-max))
 313     (reverse-region (point-min) (point-max))
 314     (goto-char (point-min))))
 315
 316 \f
 317
 318 ;; Bogofilter Interface
 319
 320 (defvar mh-bogofilter-executable (executable-find "bogofilter"))
 321
 322 (defun mh-bogofilter-blacklist (msg)
 323   "Blacklist MSG with bogofilter.
 324
 325 Bogofilter is a Bayesian spam filtering program. Get it from your
 326 local distribution or from http://bogofilter.sourceforge.net/.
 327
 328 Bogofilter is taught by running:
 329
 330     bogofilter -n < good-message
 331
 332 on every good message, and
 333
 334     bogofilter -s < spam-message
 335
 336 on every spam message. This is called a full training; three other
 337 training methods are described in the FAQ that is distributed with
 338 bogofilter. Note that most Bayesian filters need 1000 to 5000 of each
 339 type of message to start doing a good job.
 340
 341 To use bogofilter, add the following recipes to \".procmailrc\":
 342
 343     MAILDIR=$HOME/`mhparam Path`
 344
 345     # Fight spam with bogofilter.
 346     :0fw
 347     | bogofilter -3 -e -p
 348
 349     :0:
 350     * ^X-Bogosity: Yes, tests=bogofilter
 351     spam/.
 352
 353     :0:
 354     * ^X-Bogosity: Unsure, tests=bogofilter
 355     spam/unsure/.
 356
 357 If bogofilter classifies a message incorrectly, or is unsure, you can
 358 use the MH-E commands \\[mh-junk-blacklist] and \\[mh-junk-whitelist]
 359 to update bogofilter's training.
 360
 361 The \"Bogofilter FAQ\" suggests that you run the following
 362 occasionally to shrink the database:
 363
 364     bogoutil -d wordlist.db | bogoutil -l wordlist.db.new
 365     mv wordlist.db wordlist.db.prv
 366     mv wordlist.db.new wordlist.db
 367
 368 The \"Bogofilter tuning HOWTO\" describes how you can fine-tune Bogofilter."
 369   (unless mh-bogofilter-executable
 370     (error "Unable to find the bogofilter executable"))
 371   (let ((msg-file (mh-msg-filename msg mh-current-folder)))
 372     (call-process mh-bogofilter-executable msg-file mh-junk-background
 373                   nil "-s")))
 374
 375 (defun mh-bogofilter-whitelist (msg)
 376   "Whitelist MSG with bogofilter.
 377
 378 See `mh-bogofilter-blacklist' for more information."
 379   (unless mh-bogofilter-executable
 380     (error "Unable to find the bogofilter executable"))
 381   (let ((msg-file (mh-msg-filename msg mh-current-folder)))
 382     (call-process mh-bogofilter-executable msg-file mh-junk-background
 383                   nil "-n")))
 384
 385 \f
 386
 387 ;; Spamprobe Interface
 388
 389 (defvar mh-spamprobe-executable (executable-find "spamprobe"))
 390
 391 (defun mh-spamprobe-blacklist (msg)
 392   "Blacklist MSG with SpamProbe.
 393
 394 SpamProbe is a Bayesian spam filtering program. Get it from your local
 395 distribution or from http://spamprobe.sourceforge.net.
 396
 397 To use SpamProbe, add the following recipes to \".procmailrc\":
 398
 399     MAILDIR=$HOME/`mhparam Path`
 400
 401     # Fight spam with SpamProbe.
 402     :0
 403     SCORE=| spamprobe receive
 404
 405     :0 wf
 406     | formail -I \"X-SpamProbe: $SCORE\"
 407
 408     :0:
 409     *^X-SpamProbe: SPAM
 410     spam/.
 411
 412 If SpamProbe classifies a message incorrectly, you can use the
 413 MH-E commands \\[mh-junk-blacklist] and \\[mh-junk-whitelist] to
 414 update SpamProbe's training."
 415   (unless mh-spamprobe-executable
 416     (error "Unable to find the spamprobe executable"))
 417   (let ((msg-file (mh-msg-filename msg mh-current-folder)))
 418     (call-process mh-spamprobe-executable msg-file mh-junk-background
 419                   nil "spam")))
 420
 421 (defun mh-spamprobe-whitelist (msg)
 422   "Whitelist MSG with SpamProbe.
 423
 424 See `mh-spamprobe-blacklist' for more information."
 425   (unless mh-spamprobe-executable
 426     (error "Unable to find the spamprobe executable"))
 427   (let ((msg-file (mh-msg-filename msg mh-current-folder)))
 428     (call-process mh-spamprobe-executable msg-file mh-junk-background
 429                   nil "good")))
 430
 431 (provide 'mh-junk)
 432
 433 ;; Local Variables:
 434 ;; indent-tabs-mode: nil
 435 ;; sentence-end-double-space: nil
 436 ;; End:
 437
 438 ;; arch-tag: 603335f1-77ff-4306-8828-5d3dad51abe1
 439 ;;; mh-junk.el ends here