lisp/mh-e/mh-junk.el

   1 ;;; mh-junk.el --- MH-E interface to anti-spam measures
   2
   3 ;; Copyright (C) 2003-2014 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 3 of the License, or
  15 ;; (at your option) 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.  If not, see <http://www.gnu.org/licenses/>.
  24
  25 ;;; Commentary:
  26
  27 ;; Spam handling in MH-E.
  28
  29 ;;; Change Log:
  30
  31 ;;; Code:
  32
  33 (require 'mh-e)
  34 (require 'mh-scan)
  35 (mh-require-cl)
  36
  37 ;;;###mh-autoload
  38 (defun mh-junk-blacklist (range)
  39   "Blacklist RANGE as spam.
  40
  41 This command trains the spam program in use (see the option
  42 `mh-junk-program') with the content of RANGE and then handles the
  43 message(s) as specified by the option `mh-junk-disposition'.
  44
  45 Check the documentation of `mh-interactive-range' to see how RANGE is
  46 read in interactive use.
  47
  48 For more information about using your particular spam fighting
  49 program, see:
  50
  51   - `mh-spamassassin-blacklist'
  52   - `mh-bogofilter-blacklist'
  53   - `mh-spamprobe-blacklist'"
  54   (interactive (list (mh-interactive-range "Blacklist")))
  55   (mh-iterate-on-range () range (mh-blacklist-a-msg nil))
  56   (if (looking-at mh-scan-blacklisted-msg-regexp)
  57       (mh-next-msg)))
  58
  59 (defun mh-blacklist-a-msg (message)
  60   "Blacklist MESSAGE.
  61 If MESSAGE is nil then the message at point is blacklisted.
  62 The hook `mh-blacklisted-msg-hook' is called after you mark a message
  63 for blacklisting."
  64   (save-excursion
  65     (if (numberp message)
  66         (mh-goto-msg message nil t)
  67       (beginning-of-line)
  68       (setq message (mh-get-msg-num t)))
  69     (cond ((looking-at mh-scan-refiled-msg-regexp)
  70            (error "Message %d is refiled; undo refile before blacklisting"
  71                   message))
  72           ((looking-at mh-scan-deleted-msg-regexp)
  73            (error "Message %d is deleted; undo delete before blacklisting"
  74                   message))
  75           ((looking-at mh-scan-whitelisted-msg-regexp)
  76            (error "Message %d is whitelisted; undo before blacklisting"
  77                   message))
  78           ((looking-at mh-scan-blacklisted-msg-regexp) nil)
  79           (t
  80            (mh-set-folder-modified-p t)
  81            (setq mh-blacklist (cons message mh-blacklist))
  82            (if (not (memq message mh-seen-list))
  83                (setq mh-seen-list (cons message mh-seen-list)))
  84            (mh-notate nil mh-note-blacklisted mh-cmd-note)
  85            (run-hooks 'mh-blacklist-msg-hook)))))
  86
  87 ;;;###mh-autoload
  88 (defun mh-junk-blacklist-disposition ()
  89   "Determines the fate of the selected spam."
  90   (cond ((null mh-junk-disposition) nil)
  91         ((equal mh-junk-disposition "") "+")
  92         ((eq (aref mh-junk-disposition 0) ?+)
  93          mh-junk-disposition)
  94         ((eq (aref mh-junk-disposition 0) ?@)
  95          (concat mh-current-folder "/"
  96                  (substring mh-junk-disposition 1)))
  97         (t (concat "+" mh-junk-disposition))))
  98
  99 ;;;###mh-autoload
 100 (defun mh-junk-process-blacklist (range)
 101   "Blacklist RANGE as spam.
 102 This command trains the spam program in use (see the option
 103 `mh-junk-program') with the content of RANGE and then handles the
 104 message(s) as specified by the option `mh-junk-disposition'."
 105   (let ((blacklist-func (nth 1 (assoc mh-junk-choice mh-junk-function-alist))))
 106     (unless blacklist-func
 107       (error "Customize `mh-junk-program' appropriately"))
 108     (mh-iterate-on-range msg range
 109       (message "Blacklisting message %d..." msg)
 110       (funcall (symbol-function blacklist-func) msg)
 111       (message "Blacklisting message %d...done" msg))
 112     (mh-next-msg)))
 113
 114 ;;;###mh-autoload
 115 (defun mh-junk-whitelist (range)
 116   "Whitelist RANGE as ham.
 117
 118 This command reclassifies the RANGE as ham if it were incorrectly
 119 classified as spam (see the option `mh-junk-program'). It then
 120 refiles the message into the \"+inbox\" folder.
 121
 122 Check the documentation of `mh-interactive-range' to see how
 123 RANGE is read in interactive use."
 124   (interactive (list (mh-interactive-range "Whitelist")))
 125   (mh-iterate-on-range () range (mh-junk-whitelist-a-msg nil))
 126   (if (looking-at mh-scan-whitelisted-msg-regexp)
 127       (mh-next-msg)))
 128
 129 (defun mh-junk-whitelist-a-msg (message)
 130   "Whitelist MESSAGE.
 131 If MESSAGE is nil then the message at point is whitelisted. The
 132 hook `mh-whitelist-msg-hook' is called after you mark a message
 133 for whitelisting."
 134   (save-excursion
 135     (if (numberp message)
 136         (mh-goto-msg message nil t)
 137       (beginning-of-line)
 138       (setq message (mh-get-msg-num t)))
 139     (cond ((looking-at mh-scan-refiled-msg-regexp)
 140            (error "Message %d is refiled; undo refile before whitelisting"
 141                   message))
 142           ((looking-at mh-scan-deleted-msg-regexp)
 143            (error "Message %d is deleted; undo delete before whitelisting"
 144                   message))
 145           ((looking-at mh-scan-blacklisted-msg-regexp)
 146            (error "Message %d is blacklisted; undo before whitelisting"
 147                   message))
 148           ((looking-at mh-scan-whitelisted-msg-regexp) nil)
 149           (t
 150            (mh-set-folder-modified-p t)
 151            (setq mh-whitelist (cons message mh-whitelist))
 152            (mh-notate nil mh-note-whitelisted mh-cmd-note)
 153            (run-hooks 'mh-whitelist-msg-hook)))))
 154
 155 ;;;###mh-autoload
 156 (defun mh-junk-process-whitelist (range)
 157   "Whitelist RANGE as ham.
 158
 159 This command reclassifies the RANGE as ham if it were incorrectly
 160 classified as spam (see the option `mh-junk-program')."
 161   (let ((whitelist-func (nth 2 (assoc mh-junk-choice mh-junk-function-alist))))
 162     (unless whitelist-func
 163       (error "Customize `mh-junk-program' appropriately"))
 164     (mh-iterate-on-range msg range
 165       (message "Whitelisting message %d..." msg)
 166       (funcall (symbol-function whitelist-func) msg)
 167       (message "Whitelisting message %d...done" msg))
 168     (mh-next-msg)))
 169
 170 \f
 171
 172 ;; Spamassassin Interface
 173
 174 (defvar mh-spamassassin-executable (executable-find "spamassassin"))
 175 (defvar mh-sa-learn-executable (executable-find "sa-learn"))
 176
 177 ;;;###mh-autoload
 178 (defun mh-spamassassin-blacklist (msg)
 179   "Blacklist MSG with SpamAssassin.
 180
 181 SpamAssassin is one of the more popular spam filtering programs.
 182 Get it from your local distribution or from the SpamAssassin web
 183 site at URL `http://spamassassin.org/'.
 184
 185 To use SpamAssassin, add the following recipes to
 186 \".procmailrc\":
 187
 188     PATH=$PATH:/usr/bin/mh
 189     MAILDIR=$HOME/`mhparam Path`
 190
 191     # Fight spam with SpamAssassin.
 192     :0fw
 193     | spamc
 194
 195     # Anything with a spam level of 10 or more is junked immediately.
 196     :0:
 197     * ^X-Spam-Level: ..........
 198     /dev/null
 199
 200     :0:
 201     * ^X-Spam-Status: Yes
 202     spam/.
 203
 204 If you don't use \"spamc\", use \"spamassassin -P -a\".
 205
 206 Note that one of the recipes above throws away messages with a
 207 score greater than or equal to 10. Here's how you can determine a
 208 value that works best for you.
 209
 210 First, run \"spamassassin -t\" on every mail message in your
 211 archive and use Gnumeric to verify that the average plus the
 212 standard deviation of good mail is under 5, the SpamAssassin
 213 default for \"spam\".
 214
 215 Using Gnumeric, sort the messages by score and view the messages
 216 with the highest score. Determine the score which encompasses all
 217 of your interesting messages and add a couple of points to be
 218 conservative. Add that many dots to the \"X-Spam-Level:\" header
 219 field above to send messages with that score down the drain.
 220
 221 In the example above, messages with a score of 5-9 are set aside
 222 in the \"+spam\" folder for later review. The major weakness of
 223 rules-based filters is a plethora of false positives so it is
 224 worthwhile to check.
 225
 226 If SpamAssassin classifies a message incorrectly, or is unsure,
 227 you can use the MH-E commands \\[mh-junk-blacklist] and
 228 \\[mh-junk-whitelist].
 229
 230 The command \\[mh-junk-blacklist] adds a \"blacklist_from\" entry
 231 to \"~/spamassassin/user_prefs\", deletes the message, and sends
 232 the message to the Razor, so that others might not see this spam.
 233 If the \"sa-learn\" command is available, the message is also
 234 recategorized as spam.
 235
 236 The command \\[mh-junk-whitelist] adds a \"whitelist_from\" rule
 237 to the \"~/.spamassassin/user_prefs\" file. If the \"sa-learn\"
 238 command is available, the message is also recategorized as ham.
 239
 240 Over time, you'll observe that the same host or domain occurs
 241 repeatedly in the \"blacklist_from\" entries, so you might think
 242 that you could avoid future spam by blacklisting all mail from a
 243 particular domain. The utility function
 244 `mh-spamassassin-identify-spammers' helps you do precisely that.
 245 This function displays a frequency count of the hosts and domains
 246 in the \"blacklist_from\" entries from the last blank line in
 247 \"~/.spamassassin/user_prefs\" to the end of the file. This
 248 information can be used so that you can replace multiple
 249 \"blacklist_from\" entries with a single wildcard entry such as:
 250
 251     blacklist_from *@*amazingoffersdirect2u.com
 252
 253 In versions of SpamAssassin (2.50 and on) that support a Bayesian
 254 classifier, \\[mh-junk-blacklist] uses the program \"sa-learn\"
 255 to recategorize the message as spam. Neither MH-E, nor
 256 SpamAssassin, rebuilds the database after adding words, so you
 257 will need to run \"sa-learn --rebuild\" periodically. This can be
 258 done by adding the following to your crontab:
 259
 260     0 * * * *   sa-learn --rebuild > /dev/null 2>&1"
 261   (unless mh-spamassassin-executable
 262     (error "Unable to find the spamassassin executable"))
 263   (let ((current-folder mh-current-folder)
 264         (msg-file (mh-msg-filename msg mh-current-folder))
 265         (sender))
 266     (message "Reporting message %d..." msg)
 267     (mh-truncate-log-buffer)
 268     ;; Put call-process output in log buffer if we are saving it
 269     ;; (this happens if mh-junk-background is t).
 270     (with-current-buffer mh-log-buffer
 271       (call-process mh-spamassassin-executable msg-file mh-junk-background nil
 272                     ;;"--report" "--remove-from-whitelist"
 273                     "-r" "-R")          ; spamassassin V2.20
 274     (when mh-sa-learn-executable
 275       (message "Recategorizing message %d as spam..." msg)
 276       (mh-truncate-log-buffer)
 277       (call-process mh-sa-learn-executable msg-file mh-junk-background nil
 278                     "--single" "--spam" "--local" "--no-rebuild")))
 279     (message "Blacklisting sender of message %d..." msg)
 280     (with-current-buffer (get-buffer-create mh-temp-buffer)
 281       (erase-buffer)
 282       (call-process (expand-file-name mh-scan-prog mh-progs)
 283                     nil t nil
 284                     (format "%d" msg) current-folder
 285                     "-format" "%<(mymbox{from})%|%(addr{from})%>")
 286       (goto-char (point-min))
 287       (if (search-forward-regexp "^\\(.+\\)$" nil t)
 288           (progn
 289             (setq sender (match-string 0))
 290             (mh-spamassassin-add-rule "blacklist_from" sender)
 291             (message "Blacklisting sender of message %d...done" msg))
 292         (message "Blacklisting sender of message %d...not done (from my address)" msg)))))
 293
 294 ;;;###mh-autoload
 295 (defun mh-spamassassin-whitelist (msg)
 296   "Whitelist MSG with SpamAssassin.
 297
 298 The \\[mh-junk-whitelist] command adds a \"whitelist_from\" rule to
 299 the \"~/.spamassassin/user_prefs\" file. If the \"sa-learn\" command
 300 is available, the message is also recategorized as ham.
 301
 302 See `mh-spamassassin-blacklist' for more information."
 303   (unless mh-spamassassin-executable
 304     (error "Unable to find the spamassassin executable"))
 305   (let ((msg-file (mh-msg-filename msg mh-current-folder))
 306         (show-buffer (get-buffer mh-show-buffer))
 307         from)
 308     (with-current-buffer (get-buffer-create mh-temp-buffer)
 309       (erase-buffer)
 310       (message "Removing spamassassin markup from message %d..." msg)
 311       (call-process mh-spamassassin-executable msg-file t nil
 312                     ;; "--remove-markup"
 313                     "-d")               ; spamassassin V2.20
 314       (if show-buffer
 315           (kill-buffer show-buffer))
 316       (write-file msg-file)
 317       (when mh-sa-learn-executable
 318         (message "Recategorizing message %d as ham..." msg)
 319         (mh-truncate-log-buffer)
 320         ;; Put call-process output in log buffer if we are saving it
 321         ;; (this happens if mh-junk-background is t).
 322         (with-current-buffer mh-log-buffer
 323           (call-process mh-sa-learn-executable msg-file mh-junk-background nil
 324                         "--single" "--ham" "--local" "--no-rebuild")))
 325       (message "Whitelisting sender of message %d..." msg)
 326       (setq from
 327             (car (mh-funcall-if-exists
 328                   ietf-drums-parse-address (mh-get-header-field "From:"))))
 329       (kill-buffer nil)
 330       (unless (or (null from) (equal from ""))
 331         (mh-spamassassin-add-rule "whitelist_from" from))
 332       (message "Whitelisting sender of message %d...done" msg))))
 333
 334 (defun mh-spamassassin-add-rule (rule body)
 335   "Add a new rule to \"~/.spamassassin/user_prefs\".
 336 The name of the rule is RULE and its body is BODY."
 337   (save-window-excursion
 338     (let* ((line (format "%s\t%s\n" rule body))
 339            (case-fold-search t)
 340            (file (expand-file-name "~/.spamassassin/user_prefs"))
 341            (buffer-exists (find-buffer-visiting file)))
 342       (find-file file)
 343       (if (not (search-forward (format "\n%s" line) nil t))
 344           (progn
 345             (goto-char (point-max))
 346             (insert (if (bolp) "" "\n") line)
 347             (save-buffer)))
 348       (if (not buffer-exists)
 349           (kill-buffer nil)))))
 350
 351 ;;;###mh-autoload
 352 (defun mh-spamassassin-identify-spammers ()
 353   "Identify spammers who are repeat offenders.
 354
 355 This function displays a frequency count of the hosts and domains
 356 in the \"blacklist_from\" entries from the last blank line in
 357 \"~/.spamassassin/user_prefs\" to the end of the file. This
 358 information can be used so that you can replace multiple
 359 \"blacklist_from\" entries with a single wildcard entry such as:
 360
 361     blacklist_from *@*amazingoffersdirect2u.com"
 362   (interactive)
 363   (let* ((file (expand-file-name "~/.spamassassin/user_prefs"))
 364          (domains (make-hash-table :test 'equal)))
 365     (find-file file)
 366     ;; Only consider entries between last blank line and end of file.
 367     (goto-char (1- (point-max)))
 368     (search-backward-regexp "^$")
 369     ;; Perform frequency count.
 370     (save-excursion
 371       (while (search-forward-regexp "^blacklist_from\\s-*\\(.*\\)@\\(.*\\)$"
 372                                     nil t)
 373         (let ((host (match-string 2))
 374               value)
 375           ;; Remove top-level-domain from hostname.
 376           (setq host (cdr (reverse (split-string host "\\."))))
 377           ;; Add counts for each host and domain part.
 378           (while host
 379             (setq value (gethash (car host) domains))
 380             (setf (gethash (car host) domains) (1+ (if (not value) 0 value)))
 381             (setq host (cdr host))))))
 382
 383     ;; Output
 384     (delete-other-windows)
 385     (pop-to-buffer (get-buffer-create "*MH-E Spammer Frequencies*"))
 386     (erase-buffer)
 387     (maphash (lambda (key value) ""
 388                (if (> value 2)
 389                    (insert (format "%s %s\n" key value))))
 390              domains)
 391     (sort-numeric-fields 2 (point-min) (point-max))
 392     (reverse-region (point-min) (point-max))
 393     (goto-char (point-min))))
 394
 395 \f
 396
 397 ;; Bogofilter Interface
 398
 399 (defvar mh-bogofilter-executable (executable-find "bogofilter"))
 400
 401 ;;;###mh-autoload
 402 (defun mh-bogofilter-blacklist (msg)
 403   "Blacklist MSG with bogofilter.
 404
 405 Bogofilter is a Bayesian spam filtering program. Get it from your
 406 local distribution or from the bogofilter web site at URL
 407 `http://bogofilter.sourceforge.net/'.
 408
 409 Bogofilter is taught by running:
 410
 411     bogofilter -n < good-message
 412
 413 on every good message, and
 414
 415     bogofilter -s < spam-message
 416
 417 on every spam message. This is called a full training; three other
 418 training methods are described in the FAQ that is distributed with
 419 bogofilter. Note that most Bayesian filters need 1000 to 5000 of each
 420 type of message to start doing a good job.
 421
 422 To use bogofilter, add the following recipes to \".procmailrc\":
 423
 424     PATH=$PATH:/usr/bin/mh
 425     MAILDIR=$HOME/`mhparam Path`
 426
 427     # Fight spam with bogofilter.
 428     :0fw
 429     | bogofilter -3 -e -p
 430
 431     :0:
 432     * ^X-Bogosity: Yes, tests=bogofilter
 433     spam/.
 434
 435     :0:
 436     * ^X-Bogosity: Unsure, tests=bogofilter
 437     spam/unsure/.
 438
 439 If bogofilter classifies a message incorrectly, or is unsure, you can
 440 use the MH-E commands \\[mh-junk-blacklist] and \\[mh-junk-whitelist]
 441 to update bogofilter's training.
 442
 443 The \"Bogofilter FAQ\" suggests that you run the following
 444 occasionally to shrink the database:
 445
 446     bogoutil -d wordlist.db | bogoutil -l wordlist.db.new
 447     mv wordlist.db wordlist.db.prv
 448     mv wordlist.db.new wordlist.db
 449
 450 The \"Bogofilter tuning HOWTO\" describes how you can fine-tune Bogofilter."
 451   (unless mh-bogofilter-executable
 452     (error "Unable to find the bogofilter executable"))
 453   (let ((msg-file (mh-msg-filename msg mh-current-folder)))
 454     (mh-truncate-log-buffer)
 455     ;; Put call-process output in log buffer if we are saving it
 456     ;; (this happens if mh-junk-background is t).
 457     (with-current-buffer mh-log-buffer
 458       (call-process mh-bogofilter-executable msg-file mh-junk-background
 459                     nil "-s"))))
 460
 461 ;;;###mh-autoload
 462 (defun mh-bogofilter-whitelist (msg)
 463   "Whitelist MSG with bogofilter.
 464
 465 See `mh-bogofilter-blacklist' for more information."
 466   (unless mh-bogofilter-executable
 467     (error "Unable to find the bogofilter executable"))
 468   (let ((msg-file (mh-msg-filename msg mh-current-folder)))
 469     (mh-truncate-log-buffer)
 470     ;; Put call-process output in log buffer if we are saving it
 471     ;; (this happens if mh-junk-background is t).
 472     (with-current-buffer mh-log-buffer
 473       (call-process mh-bogofilter-executable msg-file mh-junk-background
 474                     nil "-n"))))
 475
 476 \f
 477
 478 ;; Spamprobe Interface
 479
 480 (defvar mh-spamprobe-executable (executable-find "spamprobe"))
 481
 482 ;;;###mh-autoload
 483 (defun mh-spamprobe-blacklist (msg)
 484   "Blacklist MSG with SpamProbe.
 485
 486 SpamProbe is a Bayesian spam filtering program. Get it from your
 487 local distribution or from the SpamProbe web site at URL
 488 `http://spamprobe.sourceforge.net'.
 489
 490 To use SpamProbe, add the following recipes to \".procmailrc\":
 491
 492     PATH=$PATH:/usr/bin/mh
 493     MAILDIR=$HOME/`mhparam Path`
 494
 495     # Fight spam with SpamProbe.
 496     :0
 497     SCORE=| spamprobe receive
 498
 499     :0 wf
 500     | formail -I \"X-SpamProbe: $SCORE\"
 501
 502     :0:
 503     *^X-SpamProbe: SPAM
 504     spam/.
 505
 506 If SpamProbe classifies a message incorrectly, you can use the
 507 MH-E commands \\[mh-junk-blacklist] and \\[mh-junk-whitelist] to
 508 update SpamProbe's training."
 509   (unless mh-spamprobe-executable
 510     (error "Unable to find the spamprobe executable"))
 511   (let ((msg-file (mh-msg-filename msg mh-current-folder)))
 512     (mh-truncate-log-buffer)
 513     ;; Put call-process output in log buffer if we are saving it
 514     ;; (this happens if mh-junk-background is t).
 515     (with-current-buffer mh-log-buffer
 516       (call-process mh-spamprobe-executable msg-file mh-junk-background
 517                     nil "spam"))))
 518
 519 ;;;###mh-autoload
 520 (defun mh-spamprobe-whitelist (msg)
 521   "Whitelist MSG with SpamProbe.
 522
 523 See `mh-spamprobe-blacklist' for more information."
 524   (unless mh-spamprobe-executable
 525     (error "Unable to find the spamprobe executable"))
 526   (let ((msg-file (mh-msg-filename msg mh-current-folder)))
 527     (mh-truncate-log-buffer)
 528     ;; Put call-process output in log buffer if we are saving it
 529     ;; (this happens if mh-junk-background is t).
 530     (with-current-buffer mh-log-buffer
 531       (call-process mh-spamprobe-executable msg-file mh-junk-background
 532                     nil "good"))))
 533
 534 (provide 'mh-junk)
 535
 536 ;; Local Variables:
 537 ;; indent-tabs-mode: nil
 538 ;; sentence-end-double-space: nil
 539 ;; End:
 540
 541 ;;; mh-junk.el ends here