Merge changes made in Gnus trunk.
[emacs/old-mirror.git] / lisp / gnus / nnir.el
blob750218d4a9a825626ceee125616515109f310bc8
1 ;;; nnir.el --- search mail with various search engines -*- coding: iso-8859-1 -*-
3 ;; Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006,
4 ;; 2007, 2008, 2009, 2010 Free Software Foundation, Inc.
6 ;; Author: Kai Großjohann <grossjohann@ls6.cs.uni-dortmund.de>
7 ;; Swish-e and Swish++ backends by:
8 ;; Christoph Conrad <christoph.conrad@gmx.de>.
9 ;; IMAP backend by: Simon Josefsson <jas@pdc.kth.se>.
10 ;; IMAP search by: Torsten Hilbrich <torsten.hilbrich <at> gmx.net>
11 ;; IMAP search improved by Daniel Pittman <daniel@rimspace.net>.
12 ;; nnmaildir support for Swish++ and Namazu backends by:
13 ;; Justus Piater <Justus <at> Piater.name>
14 ;; Keywords: news mail searching ir
16 ;; This file is part of GNU Emacs.
18 ;; GNU Emacs is free software: you can redistribute it and/or modify
19 ;; it under the terms of the GNU General Public License as published by
20 ;; the Free Software Foundation, either version 3 of the License, or
21 ;; (at your option) any later version.
23 ;; GNU Emacs is distributed in the hope that it will be useful,
24 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
25 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
26 ;; GNU General Public License for more details.
28 ;; You should have received a copy of the GNU General Public License
29 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
31 ;;; Commentary:
33 ;; TODO: Documentation in the Gnus manual
35 ;; Where in the existing gnus manual would this fit best?
37 ;; What does it do? Well, it allows you to search your mail using
38 ;; some search engine (imap, namazu, swish-e, gmane and others -- see
39 ;; later) by typing `G G' in the Group buffer. You will then get a
40 ;; buffer which shows all articles matching the query, sorted by
41 ;; Retrieval Status Value (score).
43 ;; When looking at the retrieval result (in the Summary buffer) you
44 ;; can type `A W' (aka M-x gnus-warp-to-article RET) on an article. You
45 ;; will be warped into the group this article came from. Typing `A T'
46 ;; (aka M-x gnus-summary-refer-thread RET) will warp to the group and
47 ;; also show the thread this article is part of.
49 ;; The Lisp setup may involve setting a few variables and setting up the
50 ;; search engine. You can define the variables in the server definition
51 ;; like this :
52 ;; (setq gnus-secondary-select-methods '(
53 ;; (nnimap "" (nnimap-address "localhost")
54 ;; (nnir-search-engine namazu)
55 ;; )))
56 ;; The main variable to set is `nnir-search-engine'. Choose one of
57 ;; the engines listed in `nnir-engines'. (Actually `nnir-engines' is
58 ;; an alist, type `C-h v nnir-engines RET' for more information; this
59 ;; includes examples for setting `nnir-search-engine', too.)
61 ;; If you use one of the local indices (namazu, find-grep, swish) you
62 ;; must also set up a search engine backend.
64 ;; 1. Namazu
66 ;; The Namazu backend requires you to have one directory containing all
67 ;; index files, this is controlled by the `nnir-namazu-index-directory'
68 ;; variable. To function the `nnir-namazu-remove-prefix' variable must
69 ;; also be correct, see the documentation for `nnir-namazu-remove-prefix'
70 ;; above.
72 ;; It is particularly important not to pass any any switches to namazu
73 ;; that will change the output format. Good switches to use include
74 ;; `--sort', `--ascending', `--early' and `--late'. Refer to the Namazu
75 ;; documentation for further information on valid switches.
77 ;; To index my mail with the `mknmz' program I use the following
78 ;; configuration file:
80 ;; ,----
81 ;; | package conf; # Don't remove this line!
82 ;; |
83 ;; | # Paths which will not be indexed. Don't use `^' or `$' anchors.
84 ;; | $EXCLUDE_PATH = "spam|sent";
85 ;; |
86 ;; | # Header fields which should be searchable. case-insensitive
87 ;; | $REMAIN_HEADER = "from|date|message-id|subject";
88 ;; |
89 ;; | # Searchable fields. case-insensitive
90 ;; | $SEARCH_FIELD = "from|date|message-id|subject";
91 ;; |
92 ;; | # The max length of a word.
93 ;; | $WORD_LENG_MAX = 128;
94 ;; |
95 ;; | # The max length of a field.
96 ;; | $MAX_FIELD_LENGTH = 256;
97 ;; `----
99 ;; My mail is stored in the directories ~/Mail/mail/, ~/Mail/lists/ and
100 ;; ~/Mail/archive/, so to index them I go to the directory set in
101 ;; `nnir-namazu-index-directory' and issue the following command.
103 ;; mknmz --mailnews ~/Mail/archive/ ~/Mail/mail/ ~/Mail/lists/
105 ;; For maximum searching efficiency I have a cron job set to run this
106 ;; command every four hours.
108 ;; 2. find-grep
110 ;; The find-grep engine simply runs find(1) to locate eligible
111 ;; articles and searches them with grep(1). This, of course, is much
112 ;; slower than using a proper search engine but OTOH doesn't require
113 ;; maintenance of an index and is still faster than using any built-in
114 ;; means for searching. The method specification of the server to
115 ;; search must include a directory for this engine to work (E.g.,
116 ;; `nnml-directory'). The tools must be POSIX compliant. GNU Find
117 ;; prior to version 4.2.12 (4.2.26 on Linux due to incorrect ARG_MAX
118 ;; handling) does not work.
119 ;; ,----
120 ;; | ;; find-grep configuration for searching the Gnus Cache
121 ;; |
122 ;; | (nnml "cache"
123 ;; | (nnml-get-new-mail nil)
124 ;; | (nnir-search-engine find-grep)
125 ;; | (nnml-directory "~/News/cache/")
126 ;; | (nnml-active-file "~/News/cache/active"))
127 ;; `----
129 ;; Developer information:
131 ;; I have tried to make the code expandable. Basically, it is divided
132 ;; into two layers. The upper layer is somewhat like the `nnvirtual'
133 ;; backend: given a specification of what articles to show from
134 ;; another backend, it creates a group containing exactly those
135 ;; articles. The lower layer issues a query to a search engine and
136 ;; produces such a specification of what articles to show from the
137 ;; other backend.
139 ;; The interface between the two layers consists of the single
140 ;; function `nnir-run-query', which just selects the appropriate
141 ;; function for the search engine one is using. The input to
142 ;; `nnir-run-query' is a string, representing the query as input by
143 ;; the user. The output of `nnir-run-query' is supposed to be a
144 ;; vector, each element of which should in turn be a three-element
145 ;; vector. The first element should be full group name of the article,
146 ;; the second element should be the article number, and the third
147 ;; element should be the Retrieval Status Value (RSV) as returned from
148 ;; the search engine. An RSV is the score assigned to the document by
149 ;; the search engine. For Boolean search engines, the
150 ;; RSV is always 1000 (or 1 or 100, or whatever you like).
152 ;; The sorting order of the articles in the summary buffer created by
153 ;; nnir is based on the order of the articles in the above mentioned
154 ;; vector, so that's where you can do the sorting you'd like. Maybe
155 ;; it would be nice to have a way of displaying the search result
156 ;; sorted differently?
158 ;; So what do you need to do when you want to add another search
159 ;; engine? You write a function that executes the query. Temporary
160 ;; data from the search engine can be put in `nnir-tmp-buffer'. This
161 ;; function should return the list of articles as a vector, as
162 ;; described above. Then, you need to register this backend in
163 ;; `nnir-engines'. Then, users can choose the backend by setting
164 ;; `nnir-search-engine' as a server variable.
166 ;;; Code:
168 ;;; Setup:
170 ;; For Emacs <22.2 and XEmacs.
171 (eval-and-compile
172 (unless (fboundp 'declare-function) (defmacro declare-function (&rest r))))
174 (require 'nnoo)
175 (require 'gnus-group)
176 (require 'message)
177 (require 'gnus-util)
178 (eval-when-compile
179 (require 'cl))
181 ;;; Internal Variables:
183 (defvar nnir-current-query nil
184 "Internal: stores current query (= group name).")
186 (defvar nnir-current-server nil
187 "Internal: stores current server (does it ever change?).")
189 (defvar nnir-current-group-marked nil
190 "Internal: stores current list of process-marked groups.")
192 (defvar nnir-artlist nil
193 "Internal: stores search result.")
195 (defvar nnir-tmp-buffer " *nnir*"
196 "Internal: temporary buffer.")
198 (defvar nnir-search-history ()
199 "Internal: the history for querying search options in nnir")
201 (defvar nnir-extra-parms nil
202 "Internal: stores request for extra search parms")
204 ;; Imap variables
206 (defvar nnir-imap-search-arguments
207 '(("Whole message" . "TEXT")
208 ("Subject" . "SUBJECT")
209 ("To" . "TO")
210 ("From" . "FROM")
211 ("Imap" . ""))
212 "Mapping from user readable keys to IMAP search items for use in nnir")
214 (defvar nnir-imap-search-other "HEADER %S"
215 "The IMAP search item to use for anything other than
216 `nnir-imap-search-arguments'. By default this is the name of an
217 email header field")
219 (defvar nnir-imap-search-argument-history ()
220 "The history for querying search options in nnir")
222 ;;; Helper macros
224 ;; Data type article list.
226 (defmacro nnir-artlist-length (artlist)
227 "Returns number of articles in artlist."
228 `(length ,artlist))
230 (defmacro nnir-artlist-article (artlist n)
231 "Returns from ARTLIST the Nth artitem (counting starting at 1)."
232 `(when (> ,n 0)
233 (elt ,artlist (1- ,n))))
235 (defmacro nnir-artitem-group (artitem)
236 "Returns the group from the ARTITEM."
237 `(elt ,artitem 0))
239 (defmacro nnir-artitem-number (artitem)
240 "Returns the number from the ARTITEM."
241 `(elt ,artitem 1))
243 (defmacro nnir-artitem-rsv (artitem)
244 "Returns the Retrieval Status Value (RSV, score) from the ARTITEM."
245 `(elt ,artitem 2))
247 (defmacro nnir-article-group (article)
248 "Returns the group for ARTICLE"
249 `(nnir-artitem-group (nnir-artlist-article nnir-artlist ,article)))
251 (defmacro nnir-article-number (article)
252 "Returns the number for ARTICLE"
253 `(nnir-artitem-number (nnir-artlist-article nnir-artlist ,article)))
255 (defmacro nnir-article-rsv (article)
256 "Returns the rsv for ARTICLE"
257 `(nnir-artitem-rsv (nnir-artlist-article nnir-artlist ,article)))
259 (defsubst nnir-article-ids (article)
260 "Returns the pair `(nnir id . real id)' of ARTICLE"
261 (cons article (nnir-article-number article)))
263 (defmacro nnir-categorize (sequence keyfunc &optional valuefunc)
264 "Sorts a sequence into categories and returns a list of the form
265 `((key1 (element11 element12)) (key2 (element21 element22))'.
266 The category key for a member of the sequence is obtained
267 as `(keyfunc member)' and the corresponding element is just
268 `member'. If `valuefunc' is non-nil, the element of the list
269 is `(valuefunc member)'."
270 `(unless (null ,sequence)
271 (let (value)
272 (mapc
273 (lambda (member)
274 (let ((y (,keyfunc member))
275 (x ,(if valuefunc
276 `(,valuefunc member)
277 'member)))
278 (if (assoc y value)
279 (push x (cadr (assoc y value)))
280 (push (list y (list x)) value))))
281 ,sequence)
282 value)))
284 ;;; Finish setup:
286 (require 'gnus-sum)
288 (eval-when-compile
289 (autoload 'nnimap-buffer "nnimap")
290 (autoload 'nnimap-command "nnimap")
291 (autoload 'nnimap-possibly-change-group "nnimap")
292 (autoload 'gnus-registry-action "gnus-registry"))
294 (nnoo-declare nnir)
295 (nnoo-define-basics nnir)
297 (gnus-declare-backend "nnir" 'mail)
300 ;;; User Customizable Variables:
302 (defgroup nnir nil
303 "Search groups in Gnus with assorted seach engines."
304 :group 'gnus)
306 (defcustom nnir-method-default-engines
307 '((nnimap . imap)
308 (nntp . gmane))
309 "*Alist of default search engines keyed by server method."
310 :type '(alist)
311 :group 'nnir)
313 (defcustom nnir-ignored-newsgroups ""
314 "*A regexp to match newsgroups in the active file that should
315 be skipped when searching."
316 :type '(regexp)
317 :group 'nnir)
319 (defcustom nnir-summary-line-format nil
320 "*The format specification of the lines in an nnir summary buffer.
322 All the items from `gnus-summary-line-format' are available, along
323 with three items unique to nnir summary buffers:
325 %Z Search retrieval score value (integer)
326 %G Article original full group name (string)
327 %g Article original short group name (string)
329 If nil this will use `gnus-summary-line-format'."
330 :type '(regexp)
331 :group 'nnir)
333 (defcustom nnir-retrieve-headers-override-function nil
334 "*If non-nil, a function that accepts an article list and group
335 and populates the `nntp-server-buffer' with the retrieved
336 headers. Must return either 'nov or 'headers indicating the
337 retrieved header format.
339 If this variable is nil, or if the provided function returns nil for a search
340 result, `gnus-retrieve-headers' will be called instead."
341 :type '(function)
342 :group 'nnir)
344 (defcustom nnir-imap-default-search-key "Whole message"
345 "*The default IMAP search key for an nnir search. Must be one of
346 the keys in `nnir-imap-search-arguments'. To use raw imap queries
347 by default set this to \"Imap\"."
348 :type '(string)
349 :group 'nnir)
351 (defcustom nnir-swish++-configuration-file
352 (expand-file-name "~/Mail/swish++.conf")
353 "*Configuration file for swish++."
354 :type '(file)
355 :group 'nnir)
357 (defcustom nnir-swish++-program "search"
358 "*Name of swish++ search executable."
359 :type '(string)
360 :group 'nnir)
362 (defcustom nnir-swish++-additional-switches '()
363 "*A list of strings, to be given as additional arguments to swish++.
365 Note that this should be a list. Ie, do NOT use the following:
366 (setq nnir-swish++-additional-switches \"-i -w\") ; wrong
367 Instead, use this:
368 (setq nnir-swish++-additional-switches '(\"-i\" \"-w\"))"
369 :type '(repeat (string))
370 :group 'nnir)
372 (defcustom nnir-swish++-remove-prefix (concat (getenv "HOME") "/Mail/")
373 "*The prefix to remove from each file name returned by swish++
374 in order to get a group name (albeit with / instead of .). This is a
375 regular expression.
377 This variable is very similar to `nnir-namazu-remove-prefix', except
378 that it is for swish++, not Namazu."
379 :type '(regexp)
380 :group 'nnir)
382 ;; Swish-E.
383 ;; URL: http://swish-e.org/
384 ;; Variables `nnir-swish-e-index-file', `nnir-swish-e-program' and
385 ;; `nnir-swish-e-additional-switches'
387 (make-obsolete-variable 'nnir-swish-e-index-file
388 'nnir-swish-e-index-files "Emacs 23.1")
389 (defcustom nnir-swish-e-index-file
390 (expand-file-name "~/Mail/index.swish-e")
391 "*Index file for swish-e.
392 This could be a server parameter.
393 It is never consulted once `nnir-swish-e-index-files', which should be
394 used instead, has been customized."
395 :type '(file)
396 :group 'nnir)
398 (defcustom nnir-swish-e-index-files
399 (list nnir-swish-e-index-file)
400 "*List of index files for swish-e.
401 This could be a server parameter."
402 :type '(repeat (file))
403 :group 'nnir)
405 (defcustom nnir-swish-e-program "swish-e"
406 "*Name of swish-e search executable.
407 This cannot be a server parameter."
408 :type '(string)
409 :group 'nnir)
411 (defcustom nnir-swish-e-additional-switches '()
412 "*A list of strings, to be given as additional arguments to swish-e.
414 Note that this should be a list. Ie, do NOT use the following:
415 (setq nnir-swish-e-additional-switches \"-i -w\") ; wrong
416 Instead, use this:
417 (setq nnir-swish-e-additional-switches '(\"-i\" \"-w\"))
419 This could be a server parameter."
420 :type '(repeat (string))
421 :group 'nnir)
423 (defcustom nnir-swish-e-remove-prefix (concat (getenv "HOME") "/Mail/")
424 "*The prefix to remove from each file name returned by swish-e
425 in order to get a group name (albeit with / instead of .). This is a
426 regular expression.
428 This variable is very similar to `nnir-namazu-remove-prefix', except
429 that it is for swish-e, not Namazu.
431 This could be a server parameter."
432 :type '(regexp)
433 :group 'nnir)
435 ;; HyREX engine, see <URL:http://ls6-www.cs.uni-dortmund.de/>
437 (defcustom nnir-hyrex-program "nnir-search"
438 "*Name of the nnir-search executable."
439 :type '(string)
440 :group 'nnir)
442 (defcustom nnir-hyrex-additional-switches '()
443 "*A list of strings, to be given as additional arguments for nnir-search.
444 Note that this should be a list. Ie, do NOT use the following:
445 (setq nnir-hyrex-additional-switches \"-ddl ddl.xml -c nnir\") ; wrong !
446 Instead, use this:
447 (setq nnir-hyrex-additional-switches '(\"-ddl\" \"ddl.xml\" \"-c\" \"nnir\"))"
448 :type '(repeat (string))
449 :group 'nnir)
451 (defcustom nnir-hyrex-index-directory (getenv "HOME")
452 "*Index directory for HyREX."
453 :type '(directory)
454 :group 'nnir)
456 (defcustom nnir-hyrex-remove-prefix (concat (getenv "HOME") "/Mail/")
457 "*The prefix to remove from each file name returned by HyREX
458 in order to get a group name (albeit with / instead of .).
460 For example, suppose that HyREX returns file names such as
461 \"/home/john/Mail/mail/misc/42\". For this example, use the following
462 setting: (setq nnir-hyrex-remove-prefix \"/home/john/Mail/\")
463 Note the trailing slash. Removing this prefix gives \"mail/misc/42\".
464 `nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to
465 arrive at the correct group name, \"mail.misc\"."
466 :type '(directory)
467 :group 'nnir)
469 ;; Namazu engine, see <URL:http://www.namazu.org/>
471 (defcustom nnir-namazu-program "namazu"
472 "*Name of Namazu search executable."
473 :type '(string)
474 :group 'nnir)
476 (defcustom nnir-namazu-index-directory (expand-file-name "~/Mail/namazu/")
477 "*Index directory for Namazu."
478 :type '(directory)
479 :group 'nnir)
481 (defcustom nnir-namazu-additional-switches '()
482 "*A list of strings, to be given as additional arguments to namazu.
483 The switches `-q', `-a', and `-s' are always used, very few other switches
484 make any sense in this context.
486 Note that this should be a list. Ie, do NOT use the following:
487 (setq nnir-namazu-additional-switches \"-i -w\") ; wrong
488 Instead, use this:
489 (setq nnir-namazu-additional-switches '(\"-i\" \"-w\"))"
490 :type '(repeat (string))
491 :group 'nnir)
493 (defcustom nnir-namazu-remove-prefix (concat (getenv "HOME") "/Mail/")
494 "*The prefix to remove from each file name returned by Namazu
495 in order to get a group name (albeit with / instead of .).
497 For example, suppose that Namazu returns file names such as
498 \"/home/john/Mail/mail/misc/42\". For this example, use the following
499 setting: (setq nnir-namazu-remove-prefix \"/home/john/Mail/\")
500 Note the trailing slash. Removing this prefix gives \"mail/misc/42\".
501 `nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to
502 arrive at the correct group name, \"mail.misc\"."
503 :type '(directory)
504 :group 'nnir)
506 ;;; Developer Extension Variable:
508 (defvar nnir-engines
509 `((imap nnir-run-imap
510 ((criteria
511 "Imap Search in" ; Prompt
512 ,(mapcar 'car nnir-imap-search-arguments) ; alist for completing
513 nil ; allow any user input
514 nil ; initial value
515 nnir-imap-search-argument-history ; the history to use
516 ,nnir-imap-default-search-key ; default
518 (gmane nnir-run-gmane
519 ((author . "Gmane Author: ")))
520 (swish++ nnir-run-swish++
521 ((group . "Swish++ Group spec: ")))
522 (swish-e nnir-run-swish-e
523 ((group . "Swish-e Group spec: ")))
524 (namazu nnir-run-namazu
526 (hyrex nnir-run-hyrex
527 ((group . "Hyrex Group spec: ")))
528 (find-grep nnir-run-find-grep
529 ((grep-options . "Grep options: "))))
530 "Alist of supported search engines.
531 Each element in the alist is a three-element list (ENGINE FUNCTION ARGS).
532 ENGINE is a symbol designating the searching engine. FUNCTION is also
533 a symbol, giving the function that does the search. The third element
534 ARGS is a list of cons pairs (PARAM . PROMPT). When issuing a query,
535 the FUNCTION will issue a query for each of the PARAMs, using PROMPT.
537 The value of `nnir-search-engine' must be one of the ENGINE symbols.
538 For example, for searching a server using namazu include
539 (nnir-search-engine namazu)
540 in the server definition. Note that you have to set additional
541 variables for most backends. For example, the `namazu' backend
542 needs the variables `nnir-namazu-program',
543 `nnir-namazu-index-directory' and `nnir-namazu-remove-prefix'.
545 Add an entry here when adding a new search engine.")
548 ;; Gnus glue.
550 (defun gnus-group-make-nnir-group (nnir-extra-parms)
551 "Create an nnir group. Asks for query."
552 (interactive "P")
553 (setq nnir-current-query nil
554 nnir-current-server nil
555 nnir-current-group-marked nil
556 nnir-artlist nil)
557 (let* ((query (read-string "Query: " nil 'nnir-search-history))
558 (parms (list (cons 'query query)))
559 (srv (if (gnus-server-server-name)
560 "all" "")))
561 (add-to-list 'parms (cons 'unique-id (message-unique-id)) t)
562 (gnus-group-read-ephemeral-group
563 (concat "nnir:" (prin1-to-string parms)) (list 'nnir srv) t
564 (cons (current-buffer) gnus-current-window-configuration)
565 nil)))
568 ;; Gnus backend interface functions.
570 (deffoo nnir-open-server (server &optional definitions)
571 ;; Just set the server variables appropriately.
572 (add-hook 'gnus-summary-mode-hook 'nnir-mode)
573 (nnoo-change-server 'nnir server definitions))
575 (deffoo nnir-request-group (group &optional server fast info)
576 "GROUP is the query string."
577 (nnir-possibly-change-server server)
578 ;; Check for cache and return that if appropriate.
579 (if (and (equal group nnir-current-query)
580 (equal gnus-group-marked nnir-current-group-marked)
581 (or (null server)
582 (equal server nnir-current-server)))
583 nnir-artlist
584 ;; Cache miss.
585 (setq nnir-artlist (nnir-run-query group server)))
586 (with-current-buffer nntp-server-buffer
587 (setq nnir-current-query group)
588 (when server (setq nnir-current-server server))
589 (setq nnir-current-group-marked gnus-group-marked)
590 (if (zerop (length nnir-artlist))
591 (nnheader-report 'nnir "Search produced empty results.")
592 ;; Remember data for cache.
593 (nnheader-insert "211 %d %d %d %s\n"
594 (nnir-artlist-length nnir-artlist) ; total #
595 1 ; first #
596 (nnir-artlist-length nnir-artlist) ; last #
597 group)))) ; group name
599 (deffoo nnir-retrieve-headers (articles &optional group server fetch-old)
600 (with-current-buffer nntp-server-buffer
601 (let ((gnus-inhibit-demon t)
602 (articles-by-group (nnir-categorize
603 articles nnir-article-group nnir-article-ids))
604 headers)
605 (while (not (null articles-by-group))
606 (let* ((group-articles (pop articles-by-group))
607 (artgroup (car group-articles))
608 (articleids (cadr group-articles))
609 (artlist (sort (mapcar 'cdr articleids) '<))
610 (server (gnus-group-server artgroup))
611 (gnus-override-method (gnus-server-to-method server))
612 parsefunc)
613 ;; (or (numberp art)
614 ;; (nnheader-report
615 ;; 'nnir
616 ;; "nnir-retrieve-headers doesn't grok message ids: %s"
617 ;; art))
618 (nnir-possibly-change-server server)
619 ;; is this needed?
620 (erase-buffer)
621 (case (setq gnus-headers-retrieved-by
623 (and
624 nnir-retrieve-headers-override-function
625 (funcall nnir-retrieve-headers-override-function
626 artlist artgroup))
627 (gnus-retrieve-headers artlist artgroup nil)))
628 (nov
629 (setq parsefunc 'nnheader-parse-nov))
630 (headers
631 (setq parsefunc 'nnheader-parse-head))
632 (t (error "Unknown header type %s while requesting articles \
633 of group %s" gnus-headers-retrieved-by artgroup)))
634 (goto-char (point-min))
635 (while (not (eobp))
636 (let* ((novitem (funcall parsefunc))
637 (artno (mail-header-number novitem))
638 (art (car (rassoc artno articleids))))
639 (when art
640 (mail-header-set-number novitem art)
641 (push novitem headers))
642 (forward-line 1)))))
643 (setq headers
644 (sort headers
645 (lambda (x y)
646 (< (mail-header-number x) (mail-header-number y)))))
647 (erase-buffer)
648 (mapc 'nnheader-insert-nov headers)
649 'nov)))
651 (deffoo nnir-request-article (article &optional group server to-buffer)
652 (if (stringp article)
653 (nnheader-report
654 'nnir
655 "nnir-retrieve-headers doesn't grok message ids: %s"
656 article)
657 (save-excursion
658 (let ((artfullgroup (nnir-article-group article))
659 (artno (nnir-article-number article))
660 ;; Bug?
661 ;; Why must we bind nntp-server-buffer here? It won't
662 ;; work if `buf' is used, say. (Of course, the set-buffer
663 ;; line below must then be updated, too.)
664 (nntp-server-buffer (or to-buffer nntp-server-buffer)))
665 (set-buffer nntp-server-buffer)
666 (erase-buffer)
667 (message "Requesting article %d from group %s"
668 artno artfullgroup)
669 (gnus-request-article artno artfullgroup nntp-server-buffer)
670 (cons artfullgroup artno)))))
672 (deffoo nnir-request-move-article (article group server accept-form
673 &optional last internal-move-group)
674 (let* ((artfullgroup (nnir-article-group article))
675 (artno (nnir-article-number article))
676 (to-newsgroup (nth 1 accept-form))
677 (to-method (gnus-find-method-for-group to-newsgroup))
678 (from-method (gnus-find-method-for-group artfullgroup))
679 (move-is-internal (gnus-server-equal from-method to-method)))
680 (unless (gnus-check-backend-function
681 'request-move-article artfullgroup)
682 (error "The group %s does not support article moving" artfullgroup))
683 (gnus-request-move-article
684 artno
685 artfullgroup
686 (nth 1 from-method)
687 accept-form
688 last
689 (and move-is-internal
690 to-newsgroup ; Not respooling
691 (gnus-group-real-name to-newsgroup)))))
693 (deffoo nnir-warp-to-article ()
694 (let* ((cur (if (> (gnus-summary-article-number) 0)
695 (gnus-summary-article-number)
696 (error "This is not a real article.")))
697 (gnus-newsgroup-name (nnir-article-group cur))
698 (backend-number (nnir-article-number cur)))
699 (gnus-summary-read-group-1 gnus-newsgroup-name t t gnus-summary-buffer
700 nil (list backend-number))))
702 (nnoo-define-skeleton nnir)
705 (defmacro nnir-add-result (dirnam artno score prefix server artlist)
706 "Ask `nnir-compose-result' to construct a result vector,
707 and if it is non-nil, add it to artlist."
708 `(let ((result (nnir-compose-result ,dirnam ,artno ,score ,prefix ,server)))
709 (when (not (null result))
710 (push result ,artlist))))
712 (autoload 'nnmaildir-base-name-to-article-number "nnmaildir")
714 ;; Helper function currently used by the Swish++ and Namazu backends;
715 ;; perhaps useful for other backends as well
716 (defun nnir-compose-result (dirnam article score prefix server)
717 "Extract the group from dirnam, and create a result vector
718 ready to be added to the list of search results."
720 ;; remove nnir-*-remove-prefix from beginning of dirnam filename
721 (when (string-match (concat "^" prefix) dirnam)
722 (setq dirnam (replace-match "" t t dirnam)))
724 (when (file-readable-p (concat prefix dirnam article))
725 ;; remove trailing slash and, for nnmaildir, cur/new/tmp
726 (setq dirnam
727 (substring dirnam 0
728 (if (string= (gnus-group-server server) "nnmaildir")
729 -5 -1)))
731 ;; Set group to dirnam without any leading dots or slashes,
732 ;; and with all subsequent slashes replaced by dots
733 (let ((group (gnus-replace-in-string
734 (gnus-replace-in-string dirnam "^[./\\]" "" t)
735 "[/\\]" "." t)))
737 (vector (gnus-group-full-name group server)
738 (if (string= (gnus-group-server server) "nnmaildir")
739 (nnmaildir-base-name-to-article-number
740 (substring article 0 (string-match ":" article))
741 group nil)
742 (string-to-number article))
743 (string-to-number score)))))
745 ;;; Search Engine Interfaces:
747 ;; imap interface
748 (defun nnir-run-imap (query srv &optional groups)
749 "Run a search against an IMAP back-end server.
750 This uses a custom query language parser; see `nnir-imap-make-query' for
751 details on the language and supported extensions"
752 (save-excursion
753 (let ((qstring (cdr (assq 'query query)))
754 (server (cadr (gnus-server-to-method srv)))
755 (defs (caddr (gnus-server-to-method srv)))
756 (criteria (or (cdr (assq 'criteria query))
757 (cdr (assoc nnir-imap-default-search-key
758 nnir-imap-search-arguments))))
759 (gnus-inhibit-demon t)
760 (groups (or groups (nnir-get-active srv))))
761 (message "Opening server %s" server)
762 (apply
763 'vconcat
764 (mapcar
765 (lambda (group)
766 (let (artlist)
767 (condition-case ()
768 (when (nnimap-possibly-change-group
769 (gnus-group-short-name group) server)
770 (with-current-buffer (nnimap-buffer)
771 (message "Searching %s..." group)
772 (let ((arts 0)
773 (result (nnimap-command "UID SEARCH %s"
774 (if (string= criteria "")
775 qstring
776 (nnir-imap-make-query
777 criteria qstring)))))
778 (mapc
779 (lambda (artnum) (push (vector group artnum 100) artlist)
780 (setq arts (1+ arts)))
781 (and (car result)
782 (delete 0 (mapcar #'string-to-number
783 (cdr (assoc "SEARCH"
784 (cdr result)))))))
785 (message "Searching %s... %d matches" group arts)))
786 (message "Searching %s...done" group))
787 (quit nil))
788 artlist))
789 groups)))))
791 (defun nnir-imap-make-query (criteria qstring)
792 "Parse the query string and criteria into an appropriate IMAP search
793 expression, returning the string query to make.
795 This implements a little language designed to return the expected results
796 to an arbitrary query string to the end user.
798 The search is always case-insensitive, as defined by RFC2060, and supports
799 the following features (inspired by the Google search input language):
801 Automatic \"and\" queries
802 If you specify multiple words then they will be treated as an \"and\"
803 expression intended to match all components.
805 Phrase searches
806 If you wrap your query in double-quotes then it will be treated as a
807 literal string.
809 Negative terms
810 If you precede a term with \"-\" then it will negate that.
812 \"OR\" queries
813 If you include an upper-case \"OR\" in your search it will cause the
814 term before it and the term after it to be treated as alternatives.
816 In future the following will be added to the language:
817 * support for date matches
818 * support for location of text matching within the query
819 * from/to/etc headers
820 * additional search terms
821 * flag based searching
822 * anything else that the RFC supports, basically."
823 ;; Walk through the query and turn it into an IMAP query string.
824 (nnir-imap-query-to-imap criteria (nnir-imap-parse-query qstring)))
827 (defun nnir-imap-query-to-imap (criteria query)
828 "Turn a s-expression format query into IMAP."
829 (mapconcat
830 ;; Turn the expressions into IMAP text
831 (lambda (item)
832 (nnir-imap-expr-to-imap criteria item))
833 ;; The query, already in s-expr format.
834 query
835 ;; Append a space between each expression
836 " "))
839 (defun nnir-imap-expr-to-imap (criteria expr)
840 "Convert EXPR into an IMAP search expression on CRITERIA"
841 ;; What sort of expression is this, eh?
842 (cond
843 ;; Simple string term
844 ((stringp expr)
845 (format "%s %S" criteria expr))
846 ;; Trivial term: and
847 ((eq expr 'and) nil)
848 ;; Composite term: or expression
849 ((eq (car-safe expr) 'or)
850 (format "OR %s %s"
851 (nnir-imap-expr-to-imap criteria (second expr))
852 (nnir-imap-expr-to-imap criteria (third expr))))
853 ;; Composite term: just the fax, mam
854 ((eq (car-safe expr) 'not)
855 (format "NOT (%s)" (nnir-imap-query-to-imap criteria (rest expr))))
856 ;; Composite term: just expand it all.
857 ((and (not (null expr)) (listp expr))
858 (format "(%s)" (nnir-imap-query-to-imap criteria expr)))
859 ;; Complex value, give up for now.
860 (t (error "Unhandled input: %S" expr))))
863 (defun nnir-imap-parse-query (string)
864 "Turn STRING into an s-expression based query based on the IMAP
865 query language as defined in `nnir-imap-make-query'.
867 This involves turning individual tokens into higher level terms
868 that the search language can then understand and use."
869 (with-temp-buffer
870 ;; Set up the parsing environment.
871 (insert string)
872 (goto-char (point-min))
873 ;; Now, collect the output terms and return them.
874 (let (out)
875 (while (not (nnir-imap-end-of-input))
876 (push (nnir-imap-next-expr) out))
877 (reverse out))))
880 (defun nnir-imap-next-expr (&optional count)
881 "Return the next expression from the current buffer."
882 (let ((term (nnir-imap-next-term count))
883 (next (nnir-imap-peek-symbol)))
884 ;; Are we looking at an 'or' expression?
885 (cond
886 ;; Handle 'expr or expr'
887 ((eq next 'or)
888 (list 'or term (nnir-imap-next-expr 2)))
889 ;; Anything else
890 (t term))))
893 (defun nnir-imap-next-term (&optional count)
894 "Return the next TERM from the current buffer."
895 (let ((term (nnir-imap-next-symbol count)))
896 ;; What sort of term is this?
897 (cond
898 ;; and -- just ignore it
899 ((eq term 'and) 'and)
900 ;; negated term
901 ((eq term 'not) (list 'not (nnir-imap-next-expr)))
902 ;; generic term
903 (t term))))
906 (defun nnir-imap-peek-symbol ()
907 "Return the next symbol from the current buffer, but don't consume it."
908 (save-excursion
909 (nnir-imap-next-symbol)))
911 (defun nnir-imap-next-symbol (&optional count)
912 "Return the next symbol from the current buffer, or nil if we are
913 at the end of the buffer. If supplied COUNT skips some symbols before
914 returning the one at the supplied position."
915 (when (and (numberp count) (> count 1))
916 (nnir-imap-next-symbol (1- count)))
917 (let ((case-fold-search t))
918 ;; end of input stream?
919 (unless (nnir-imap-end-of-input)
920 ;; No, return the next symbol from the stream.
921 (cond
922 ;; negated expression -- return it and advance one char.
923 ((looking-at "-") (forward-char 1) 'not)
924 ;; quoted string
925 ((looking-at "\"") (nnir-imap-delimited-string "\""))
926 ;; list expression -- we parse the content and return this as a list.
927 ((looking-at "(")
928 (nnir-imap-parse-query (nnir-imap-delimited-string ")")))
929 ;; keyword input -- return a symbol version
930 ((looking-at "\\band\\b") (forward-char 3) 'and)
931 ((looking-at "\\bor\\b") (forward-char 2) 'or)
932 ((looking-at "\\bnot\\b") (forward-char 3) 'not)
933 ;; Simple, boring keyword
934 (t (let ((start (point))
935 (end (if (search-forward-regexp "[[:blank:]]" nil t)
936 (prog1
937 (match-beginning 0)
938 ;; unskip if we hit a non-blank terminal character.
939 (when (string-match "[^[:blank:]]" (match-string 0))
940 (backward-char 1)))
941 (goto-char (point-max)))))
942 (buffer-substring start end)))))))
944 (defun nnir-imap-delimited-string (delimiter)
945 "Return a delimited string from the current buffer."
946 (let ((start (point)) end)
947 (forward-char 1) ; skip the first delimiter.
948 (while (not end)
949 (unless (search-forward delimiter nil t)
950 (error "Unmatched delimited input with %s in query" delimiter))
951 (let ((here (point)))
952 (unless (equal (buffer-substring (- here 2) (- here 1)) "\\")
953 (setq end (point)))))
954 (buffer-substring (1+ start) (1- end))))
956 (defun nnir-imap-end-of-input ()
957 "Are we at the end of input?"
958 (skip-chars-forward "[[:blank:]]")
959 (looking-at "$"))
962 ;; Swish++ interface.
963 ;; -cc- Todo
964 ;; Search by
965 ;; - group
966 ;; Sort by
967 ;; - rank (default)
968 ;; - article number
969 ;; - file size
970 ;; - group
971 (defun nnir-run-swish++ (query server &optional group)
972 "Run QUERY against swish++.
973 Returns a vector of (group name, file name) pairs (also vectors,
974 actually).
976 Tested with swish++ 4.7 on GNU/Linux and with swish++ 5.0b2 on
977 Windows NT 4.0."
979 ;; (when group
980 ;; (error "The swish++ backend cannot search specific groups"))
982 (save-excursion
983 (let ( (qstring (cdr (assq 'query query)))
984 (groupspec (cdr (assq 'group query)))
985 (prefix (nnir-read-server-parm 'nnir-swish++-remove-prefix server))
986 artlist
987 ;; nnml-use-compressed-files might be any string, but probably this
988 ;; is sufficient. Note that we can't only use the value of
989 ;; nnml-use-compressed-files because old articles might have been
990 ;; saved with a different value.
991 (article-pattern (if (string= (gnus-group-server server) "nnmaildir")
992 ":[0-9]+"
993 "^[0-9]+\\(\\.[a-z0-9]+\\)?$"))
994 score artno dirnam filenam)
996 (when (equal "" qstring)
997 (error "swish++: You didn't enter anything"))
999 (set-buffer (get-buffer-create nnir-tmp-buffer))
1000 (erase-buffer)
1002 (if groupspec
1003 (message "Doing swish++ query %s on %s..." qstring groupspec)
1004 (message "Doing swish++ query %s..." qstring))
1006 (let* ((cp-list `( ,nnir-swish++-program
1007 nil ; input from /dev/null
1008 t ; output
1009 nil ; don't redisplay
1010 "--config-file" ,(nnir-read-server-parm 'nnir-swish++-configuration-file server)
1011 ,@(nnir-read-server-parm 'nnir-swish++-additional-switches server)
1012 ,qstring ; the query, in swish++ format
1014 (exitstatus
1015 (progn
1016 (message "%s args: %s" nnir-swish++-program
1017 (mapconcat 'identity (cddddr cp-list) " ")) ;; ???
1018 (apply 'call-process cp-list))))
1019 (unless (or (null exitstatus)
1020 (zerop exitstatus))
1021 (nnheader-report 'nnir "Couldn't run swish++: %s" exitstatus)
1022 ;; swish++ failure reason is in this buffer, show it if
1023 ;; the user wants it.
1024 (when (> gnus-verbose 6)
1025 (display-buffer nnir-tmp-buffer))))
1027 ;; The results are output in the format of:
1028 ;; V 4.7 Linux
1029 ;; rank relative-path-name file-size file-title
1030 ;; V 5.0b2:
1031 ;; rank relative-path-name file-size topic??
1032 ;; where rank is an integer from 1 to 100.
1033 (goto-char (point-min))
1034 (while (re-search-forward
1035 "\\(^[0-9]+\\) \\([^ ]+\\) [0-9]+ \\(.*\\)$" nil t)
1036 (setq score (match-string 1)
1037 filenam (match-string 2)
1038 artno (file-name-nondirectory filenam)
1039 dirnam (file-name-directory filenam))
1041 ;; don't match directories
1042 (when (string-match article-pattern artno)
1043 (when (not (null dirnam))
1045 ;; maybe limit results to matching groups.
1046 (when (or (not groupspec)
1047 (string-match groupspec dirnam))
1048 (nnir-add-result dirnam artno score prefix server artlist)))))
1050 (message "Massaging swish++ output...done")
1052 ;; Sort by score
1053 (apply 'vector
1054 (sort artlist
1055 (function (lambda (x y)
1056 (> (nnir-artitem-rsv x)
1057 (nnir-artitem-rsv y)))))))))
1059 ;; Swish-E interface.
1060 (defun nnir-run-swish-e (query server &optional group)
1061 "Run given query against swish-e.
1062 Returns a vector of (group name, file name) pairs (also vectors,
1063 actually).
1065 Tested with swish-e-2.0.1 on Windows NT 4.0."
1067 ;; swish-e crashes with empty parameter to "-w" on commandline...
1068 ;; (when group
1069 ;; (error "The swish-e backend cannot search specific groups"))
1071 (save-excursion
1072 (let ((qstring (cdr (assq 'query query)))
1073 (prefix
1074 (or (nnir-read-server-parm 'nnir-swish-e-remove-prefix server)
1075 (error "Missing parameter `nnir-swish-e-remove-prefix'")))
1076 artlist score artno dirnam group )
1078 (when (equal "" qstring)
1079 (error "swish-e: You didn't enter anything"))
1081 (set-buffer (get-buffer-create nnir-tmp-buffer))
1082 (erase-buffer)
1084 (message "Doing swish-e query %s..." query)
1085 (let* ((index-files
1086 (or (nnir-read-server-parm
1087 'nnir-swish-e-index-files server)
1088 (error "Missing parameter `nnir-swish-e-index-files'")))
1089 (additional-switches
1090 (nnir-read-server-parm
1091 'nnir-swish-e-additional-switches server))
1092 (cp-list `(,nnir-swish-e-program
1093 nil ; input from /dev/null
1094 t ; output
1095 nil ; don't redisplay
1096 "-f" ,@index-files
1097 ,@additional-switches
1098 "-w"
1099 ,qstring ; the query, in swish-e format
1101 (exitstatus
1102 (progn
1103 (message "%s args: %s" nnir-swish-e-program
1104 (mapconcat 'identity (cddddr cp-list) " "))
1105 (apply 'call-process cp-list))))
1106 (unless (or (null exitstatus)
1107 (zerop exitstatus))
1108 (nnheader-report 'nnir "Couldn't run swish-e: %s" exitstatus)
1109 ;; swish-e failure reason is in this buffer, show it if
1110 ;; the user wants it.
1111 (when (> gnus-verbose 6)
1112 (display-buffer nnir-tmp-buffer))))
1114 ;; The results are output in the format of:
1115 ;; rank path-name file-title file-size
1116 (goto-char (point-min))
1117 (while (re-search-forward
1118 "\\(^[0-9]+\\) \\([^ ]+\\) \"\\([^\"]+\\)\" [0-9]+$" nil t)
1119 (setq score (match-string 1)
1120 artno (match-string 3)
1121 dirnam (file-name-directory (match-string 2)))
1123 ;; don't match directories
1124 (when (string-match "^[0-9]+$" artno)
1125 (when (not (null dirnam))
1127 ;; remove nnir-swish-e-remove-prefix from beginning of dirname
1128 (when (string-match (concat "^" prefix) dirnam)
1129 (setq dirnam (replace-match "" t t dirnam)))
1131 (setq dirnam (substring dirnam 0 -1))
1132 ;; eliminate all ".", "/", "\" from beginning. Always matches.
1133 (string-match "^[./\\]*\\(.*\\)$" dirnam)
1134 ;; "/" -> "."
1135 (setq group (gnus-replace-in-string (match-string 1 dirnam) "/" "."))
1136 ;; Windows "\\" -> "."
1137 (setq group (gnus-replace-in-string group "\\\\" "."))
1139 (push (vector (gnus-group-full-name group server)
1140 (string-to-number artno)
1141 (string-to-number score))
1142 artlist))))
1144 (message "Massaging swish-e output...done")
1146 ;; Sort by score
1147 (apply 'vector
1148 (sort artlist
1149 (function (lambda (x y)
1150 (> (nnir-artitem-rsv x)
1151 (nnir-artitem-rsv y)))))))))
1153 ;; HyREX interface
1154 (defun nnir-run-hyrex (query server &optional group)
1155 (save-excursion
1156 (let ((artlist nil)
1157 (groupspec (cdr (assq 'group query)))
1158 (qstring (cdr (assq 'query query)))
1159 (prefix (nnir-read-server-parm 'nnir-hyrex-remove-prefix server))
1160 score artno dirnam)
1161 (when (and (not groupspec) group)
1162 (setq groupspec
1163 (regexp-opt
1164 (mapcar (lambda (x) (gnus-group-real-name x)) group))))
1165 (set-buffer (get-buffer-create nnir-tmp-buffer))
1166 (erase-buffer)
1167 (message "Doing hyrex-search query %s..." query)
1168 (let* ((cp-list
1169 `( ,nnir-hyrex-program
1170 nil ; input from /dev/null
1171 t ; output
1172 nil ; don't redisplay
1173 "-i",(nnir-read-server-parm 'nnir-hyrex-index-directory server) ; index directory
1174 ,@(nnir-read-server-parm 'nnir-hyrex-additional-switches server)
1175 ,qstring ; the query, in hyrex-search format
1177 (exitstatus
1178 (progn
1179 (message "%s args: %s" nnir-hyrex-program
1180 (mapconcat 'identity (cddddr cp-list) " "))
1181 (apply 'call-process cp-list))))
1182 (unless (or (null exitstatus)
1183 (zerop exitstatus))
1184 (nnheader-report 'nnir "Couldn't run hyrex-search: %s" exitstatus)
1185 ;; nnir-search failure reason is in this buffer, show it if
1186 ;; the user wants it.
1187 (when (> gnus-verbose 6)
1188 (display-buffer nnir-tmp-buffer)))) ;; FIXME: Dont clear buffer !
1189 (message "Doing hyrex-search query \"%s\"...done" qstring)
1190 (sit-for 0)
1191 ;; nnir-search returns:
1192 ;; for nnml/nnfolder: "filename mailid weigth"
1193 ;; for nnimap: "group mailid weigth"
1194 (goto-char (point-min))
1195 (delete-non-matching-lines "^\\S + [0-9]+ [0-9]+$")
1196 ;; HyREX doesn't search directly in groups -- so filter out here.
1197 (when groupspec
1198 (keep-lines groupspec))
1199 ;; extract data from result lines
1200 (goto-char (point-min))
1201 (while (re-search-forward
1202 "\\(\\S +\\) \\([0-9]+\\) \\([0-9]+\\)" nil t)
1203 (setq dirnam (match-string 1)
1204 artno (match-string 2)
1205 score (match-string 3))
1206 (when (string-match prefix dirnam)
1207 (setq dirnam (replace-match "" t t dirnam)))
1208 (push (vector (gnus-group-full-name
1209 (gnus-replace-in-string dirnam "/" ".") server)
1210 (string-to-number artno)
1211 (string-to-number score))
1212 artlist))
1213 (message "Massaging hyrex-search output...done.")
1214 (apply 'vector
1215 (sort artlist
1216 (function (lambda (x y)
1217 (if (string-lessp (nnir-artitem-group x)
1218 (nnir-artitem-group y))
1220 (< (nnir-artitem-number x)
1221 (nnir-artitem-number y)))))))
1224 ;; Namazu interface
1225 (defun nnir-run-namazu (query server &optional group)
1226 "Run given query against Namazu. Returns a vector of (group name, file name)
1227 pairs (also vectors, actually).
1229 Tested with Namazu 2.0.6 on a GNU/Linux system."
1230 ;; (when group
1231 ;; (error "The Namazu backend cannot search specific groups"))
1232 (save-excursion
1233 (let ((article-pattern (if (string= (gnus-group-server server) "nnmaildir")
1234 ":[0-9]+"
1235 "^[0-9]+$"))
1236 artlist
1237 (qstring (cdr (assq 'query query)))
1238 (prefix (nnir-read-server-parm 'nnir-namazu-remove-prefix server))
1239 score group article
1240 (process-environment (copy-sequence process-environment)))
1241 (setenv "LC_MESSAGES" "C")
1242 (set-buffer (get-buffer-create nnir-tmp-buffer))
1243 (erase-buffer)
1244 (let* ((cp-list
1245 `( ,nnir-namazu-program
1246 nil ; input from /dev/null
1247 t ; output
1248 nil ; don't redisplay
1249 "-q" ; don't be verbose
1250 "-a" ; show all matches
1251 "-s" ; use short format
1252 ,@(nnir-read-server-parm 'nnir-namazu-additional-switches server)
1253 ,qstring ; the query, in namazu format
1254 ,(nnir-read-server-parm 'nnir-namazu-index-directory server) ; index directory
1256 (exitstatus
1257 (progn
1258 (message "%s args: %s" nnir-namazu-program
1259 (mapconcat 'identity (cddddr cp-list) " "))
1260 (apply 'call-process cp-list))))
1261 (unless (or (null exitstatus)
1262 (zerop exitstatus))
1263 (nnheader-report 'nnir "Couldn't run namazu: %s" exitstatus)
1264 ;; Namazu failure reason is in this buffer, show it if
1265 ;; the user wants it.
1266 (when (> gnus-verbose 6)
1267 (display-buffer nnir-tmp-buffer))))
1269 ;; Namazu output looks something like this:
1270 ;; 2. Re: Gnus agent expire broken (score: 55)
1271 ;; /home/henrik/Mail/mail/sent/1310 (4,138 bytes)
1273 (goto-char (point-min))
1274 (while (re-search-forward
1275 "^\\([0-9]+\\.\\).*\\((score: \\([0-9]+\\)\\))\n\\([^ ]+\\)"
1276 nil t)
1277 (setq score (match-string 3)
1278 group (file-name-directory (match-string 4))
1279 article (file-name-nondirectory (match-string 4)))
1281 ;; make sure article and group is sane
1282 (when (and (string-match article-pattern article)
1283 (not (null group)))
1284 (nnir-add-result group article score prefix server artlist)))
1286 ;; sort artlist by score
1287 (apply 'vector
1288 (sort artlist
1289 (function (lambda (x y)
1290 (> (nnir-artitem-rsv x)
1291 (nnir-artitem-rsv y)))))))))
1293 (defun nnir-run-find-grep (query server &optional grouplist)
1294 "Run find and grep to obtain matching articles."
1295 (let* ((method (gnus-server-to-method server))
1296 (sym (intern
1297 (concat (symbol-name (car method)) "-directory")))
1298 (directory (cadr (assoc sym (cddr method))))
1299 (regexp (cdr (assoc 'query query)))
1300 (grep-options (cdr (assoc 'grep-options query)))
1301 (grouplist (or grouplist (nnir-get-active server)))
1302 artlist)
1303 (unless directory
1304 (error "No directory found in method specification of server %s"
1305 server))
1306 (apply
1307 'vconcat
1308 (mapcar (lambda (x)
1309 (let ((group x))
1310 (message "Searching %s using find-grep..."
1311 (or group server))
1312 (save-window-excursion
1313 (set-buffer (get-buffer-create nnir-tmp-buffer))
1314 (erase-buffer)
1315 (if (> gnus-verbose 6)
1316 (pop-to-buffer (current-buffer)))
1317 (cd directory) ; Using relative paths simplifies
1318 ; postprocessing.
1319 (let ((group
1320 (if (not group)
1322 ;; Try accessing the group literally as
1323 ;; well as interpreting dots as directory
1324 ;; separators so the engine works with
1325 ;; plain nnml as well as the Gnus Cache.
1326 (let ((group (gnus-group-real-name group)))
1327 ;; Replace cl-func find-if.
1328 (if (file-directory-p group)
1329 group
1330 (if (file-directory-p
1331 (setq group
1332 (gnus-replace-in-string
1333 group
1334 "\\." "/" t)))
1335 group))))))
1336 (unless group
1337 (error "Cannot locate directory for group"))
1338 (save-excursion
1339 (apply
1340 'call-process "find" nil t
1341 "find" group "-type" "f" "-name" "[0-9]*" "-exec"
1342 "grep"
1343 `("-l" ,@(and grep-options
1344 (split-string grep-options "\\s-" t))
1345 "-e" ,regexp "{}" "+"))))
1347 ;; Translate relative paths to group names.
1348 (while (not (eobp))
1349 (let* ((path (split-string
1350 (buffer-substring
1351 (point)
1352 (line-end-position)) "/" t))
1353 (art (string-to-number (car (last path)))))
1354 (while (string= "." (car path))
1355 (setq path (cdr path)))
1356 (let ((group (mapconcat 'identity
1357 ;; Replace cl-func:
1358 ;; (subseq path 0 -1)
1359 (let ((end (1- (length path)))
1360 res)
1361 (while
1362 (>= (setq end (1- end)) 0)
1363 (push (pop path) res))
1364 (nreverse res))
1365 ".")))
1366 (push
1367 (vector (gnus-group-full-name group server) art 0)
1368 artlist))
1369 (forward-line 1)))
1370 (message "Searching %s using find-grep...done"
1371 (or group server))
1372 artlist)))
1373 grouplist))))
1375 (declare-function mm-url-insert "mm-url" (url &optional follow-refresh))
1376 (declare-function mm-url-encode-www-form-urlencoded "mm-url" (pairs))
1378 ;; gmane interface
1379 (defun nnir-run-gmane (query srv &optional groups)
1380 "Run a search against a gmane back-end server."
1381 (if (gnus-string-match-p
1382 "gmane.org$"
1383 (or (cadr (assoc 'nntp-address (cddr (gnus-server-to-method srv))))
1384 ""))
1385 (let* ((case-fold-search t)
1386 (qstring (cdr (assq 'query query)))
1387 (server (cadr (gnus-server-to-method srv)))
1388 (groupspec (if groups
1389 (mapconcat
1390 (lambda (x)
1391 (format "group:%s" (gnus-group-short-name x)))
1392 groups " ") ""))
1393 (authorspec
1394 (if (assq 'author query)
1395 (format "author:%s" (cdr (assq 'author query))) ""))
1396 (search (format "%s %s %s"
1397 qstring groupspec authorspec))
1398 (gnus-inhibit-demon t)
1399 artlist)
1400 (require 'mm-url)
1401 (with-current-buffer (get-buffer-create nnir-tmp-buffer)
1402 (erase-buffer)
1403 (mm-url-insert
1404 (concat
1405 "http://search.gmane.org/nov.php"
1407 (mm-url-encode-www-form-urlencoded
1408 `(("query" . ,search)
1409 ("HITSPERPAGE" . "999")))))
1410 (unless (featurep 'xemacs) (set-buffer-multibyte t))
1411 (mm-decode-coding-region (point-min) (point-max) 'utf-8)
1412 (goto-char (point-min))
1413 (forward-line 1)
1414 (while (not (eobp))
1415 (unless (or (eolp) (looking-at "\x0d"))
1416 (let ((header (nnheader-parse-nov)))
1417 (let ((xref (mail-header-xref header))
1418 (xscore (string-to-number (cdr (assoc 'X-Score
1419 (mail-header-extra header))))))
1420 (when (string-match " \\([^:]+\\)[:/]\\([0-9]+\\)" xref)
1421 (push
1422 (vector
1423 (gnus-group-prefixed-name (match-string 1 xref) srv)
1424 (string-to-number (match-string 2 xref)) xscore)
1425 artlist)))))
1426 (forward-line 1)))
1427 (apply 'vector (nreverse (mm-delete-duplicates artlist))))
1428 (message "Can't search non-gmane nntp groups")
1429 nil))
1431 ;;; Util Code:
1433 (defun nnir-read-parms (query nnir-search-engine)
1434 "Reads additional search parameters according to `nnir-engines'."
1435 (let ((parmspec (caddr (assoc nnir-search-engine nnir-engines))))
1436 (append query
1437 (mapcar 'nnir-read-parm parmspec))))
1439 (defun nnir-read-parm (parmspec)
1440 "Reads a single search parameter.
1441 `parmspec' is a cons cell, the car is a symbol, the cdr is a prompt."
1442 (let ((sym (car parmspec))
1443 (prompt (cdr parmspec)))
1444 (if (listp prompt)
1445 (let* ((result (apply 'gnus-completing-read prompt))
1446 (mapping (or (assoc result nnir-imap-search-arguments)
1447 (cons nil nnir-imap-search-other))))
1448 (cons sym (format (cdr mapping) result)))
1449 (cons sym (read-string prompt)))))
1451 (autoload 'gnus-group-topic-name "gnus-topic")
1453 (defun nnir-run-query (query nserver)
1454 "Invoke appropriate search engine function (see `nnir-engines').
1455 If some groups were process-marked, run the query for each of the groups
1456 and concat the results."
1457 (let ((q (car (read-from-string query)))
1458 (groups (if (string= "all-ephemeral" nserver)
1459 (with-current-buffer gnus-server-buffer
1460 (list (list (gnus-server-server-name))))
1461 (nnir-categorize
1462 (or gnus-group-marked
1463 (if (gnus-group-group-name)
1464 (list (gnus-group-group-name))
1465 (cdr (assoc (gnus-group-topic-name)
1466 gnus-topic-alist))))
1467 gnus-group-server))))
1468 (apply 'vconcat
1469 (mapcar
1470 (lambda (x)
1471 (let* ((server (car x))
1472 (nnir-search-engine
1473 (or (nnir-read-server-parm 'nnir-search-engine
1474 server)
1475 (cdr (assoc (car
1476 (gnus-server-to-method server))
1477 nnir-method-default-engines))))
1478 search-func)
1479 (setq search-func (cadr (assoc nnir-search-engine
1480 nnir-engines)))
1481 (if search-func
1482 (funcall search-func
1483 (if nnir-extra-parms
1484 (nnir-read-parms q nnir-search-engine)
1486 server (cadr x))
1487 nil)))
1488 groups))))
1490 (defun nnir-read-server-parm (key server)
1491 "Returns the parameter value of key for the given server, where
1492 server is of form 'backend:name'."
1493 (let ((method (gnus-server-to-method server)))
1494 (cond ((and method (assq key (cddr method)))
1495 (nth 1 (assq key (cddr method))))
1496 (t nil))))
1498 (defun nnir-possibly-change-server (server)
1499 (unless (and server (nnir-server-opened server))
1500 (nnir-open-server server)))
1504 ;; unused?
1505 (defun nnir-artlist-groups (artlist)
1506 "Returns a list of all groups in the given ARTLIST."
1507 (let ((res nil)
1508 (with-dups nil))
1509 ;; from each artitem, extract group component
1510 (setq with-dups (mapcar 'nnir-artitem-group artlist))
1511 ;; remove duplicates from above
1512 (mapc (function (lambda (x) (add-to-list 'res x)))
1513 with-dups)
1514 res))
1516 (defun nnir-get-active (srv)
1517 (let ((method (gnus-server-to-method srv))
1518 groups)
1519 (gnus-request-list method)
1520 (with-current-buffer nntp-server-buffer
1521 (let ((cur (current-buffer))
1522 name)
1523 (goto-char (point-min))
1524 (unless (string= nnir-ignored-newsgroups "")
1525 (delete-matching-lines nnir-ignored-newsgroups))
1526 (if (eq (car method) 'nntp)
1527 (while (not (eobp))
1528 (ignore-errors
1529 (push (mm-string-as-unibyte
1530 (gnus-group-full-name
1531 (buffer-substring
1532 (point)
1533 (progn
1534 (skip-chars-forward "^ \t")
1535 (point))) method))
1536 groups))
1537 (forward-line))
1538 (while (not (eobp))
1539 (ignore-errors
1540 (push (mm-string-as-unibyte
1541 (if (eq (char-after) ?\")
1542 (gnus-group-full-name (read cur) method)
1543 (let ((p (point)) (name ""))
1544 (skip-chars-forward "^ \t\\\\")
1545 (setq name (buffer-substring p (point)))
1546 (while (eq (char-after) ?\\)
1547 (setq p (1+ (point)))
1548 (forward-char 2)
1549 (skip-chars-forward "^ \t\\\\")
1550 (setq name (concat name (buffer-substring
1551 p (point)))))
1552 (gnus-group-full-name name method))))
1553 groups))
1554 (forward-line)))))
1555 groups))
1557 (defun nnir-registry-action (action data-header from &optional to method)
1558 "Call `gnus-registry-action' with the original article group."
1559 (gnus-registry-action
1560 action
1561 data-header
1562 (nnir-article-group (mail-header-number data-header))
1564 method))
1566 (defun nnir-mode ()
1567 (when (eq (car (gnus-find-method-for-group gnus-newsgroup-name)) 'nnir)
1568 (setq gnus-summary-line-format
1569 (or nnir-summary-line-format gnus-summary-line-format))
1570 (remove-hook 'gnus-summary-article-delete-hook 'gnus-registry-action t)
1571 (remove-hook 'gnus-summary-article-move-hook 'gnus-registry-action t)
1572 (add-hook 'gnus-summary-article-delete-hook 'nnir-registry-action t t)
1573 (add-hook 'gnus-summary-article-move-hook 'nnir-registry-action t t)))
1577 ;; The end.
1578 (provide 'nnir)
1580 ;;; nnir.el ends here