Mercurial > emacs
annotate lisp/gnus/nnir.el @ 110262:97a40b7935bd
gnus-html.el (gnus-html-wash-tags, gnus-html-put-image): Mark cid and internal images as deletable by `W D D'.
author | Katsumi Yamaoka <yamaoka@jpl.org> |
---|---|
date | Tue, 07 Sep 2010 06:23:16 +0000 |
parents | 5b9f64b04a04 |
children | f2e111723c3a |
rev | line source |
---|---|
95610 | 1 ;;; nnir.el --- search mail with various search engines -*- coding: iso-8859-1 -*- |
2 | |
95616 | 3 ;; Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, |
106815 | 4 ;; 2007, 2008, 2009, 2010 Free Software Foundation, Inc. |
95610 | 5 |
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> | |
95616 | 14 ;; Keywords: news mail searching ir |
15 | |
16 ;; This file is part of GNU Emacs. | |
17 | |
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. | |
22 | |
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. | |
27 | |
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/>. | |
30 | |
31 ;;; Commentary: | |
95610 | 32 |
33 ;; TODO: Documentation in the Gnus manual | |
34 | |
35 ;; From: Reiner Steib | |
36 ;; Subject: Re: Including nnir.el | |
37 ;; Newsgroups: gmane.emacs.gnus.general | |
38 ;; Message-ID: <v9d5dnp6aq.fsf@marauder.physik.uni-ulm.de> | |
39 ;; Date: 2006-06-05 22:49:01 GMT | |
40 ;; | |
41 ;; On Sun, Jun 04 2006, Sascha Wilde wrote: | |
42 ;; | |
43 ;; > The one thing most hackers like to forget: Documentation. By now the | |
44 ;; > documentation is only in the comments at the head of the source, I | |
45 ;; > would use it as basis to cook up some minimal texinfo docs. | |
46 ;; > | |
47 ;; > Where in the existing gnus manual would this fit best? | |
48 | |
49 ;; Maybe (info "(gnus)Combined Groups") for a general description. | |
50 ;; `gnus-group-make-nnir-group' might be described in (info | |
51 ;; "(gnus)Foreign Groups") as well. | |
52 | |
53 | |
54 ;; The most recent version of this can always be fetched from the Gnus | |
107975
7d805250c222
Synch with Gnus trunk:
Katsumi Yamaoka <yamaoka@jpl.org>
parents:
106815
diff
changeset
|
55 ;; repository. See http://www.gnus.org/ for more information. |
95610 | 56 |
57 ;; This code is still in the development stage but I'd like other | |
58 ;; people to have a look at it. Please do not hesitate to contact me | |
59 ;; with your ideas. | |
60 | |
61 ;; What does it do? Well, it allows you to index your mail using some | |
62 ;; search engine (freeWAIS-sf, swish-e and others -- see later), | |
63 ;; then type `G G' in the Group buffer and issue a query to the search | |
64 ;; engine. You will then get a buffer which shows all articles | |
65 ;; matching the query, sorted by Retrieval Status Value (score). | |
66 | |
67 ;; When looking at the retrieval result (in the Summary buffer) you | |
68 ;; can type `G T' (aka M-x gnus-summary-nnir-goto-thread RET) on an | |
69 ;; article. You will be teleported into the group this article came | |
70 ;; from, showing the thread this article is part of. (See below for | |
71 ;; restrictions.) | |
72 | |
73 ;; The Lisp installation is simple: just put this file on your | |
74 ;; load-path, byte-compile it, and load it from ~/.gnus or something. | |
75 ;; This will install a new command `G G' in your Group buffer for | |
76 ;; searching your mail. Note that you also need to configure a number | |
77 ;; of variables, as described below. | |
78 | |
79 ;; Restrictions: | |
80 ;; | |
81 ;; * If you don't use HyREX as your search engine, this expects that | |
82 ;; you use nnml or another one-file-per-message backend, because the | |
83 ;; others doesn't support nnfolder. | |
84 ;; * It can only search the mail backend's which are supported by one | |
85 ;; search engine, because of different query languages. | |
86 ;; * There are restrictions to the Wais setup. | |
87 ;; * There are restrictions to the imap setup. | |
88 ;; * gnus-summary-nnir-goto-thread: Fetches whole group first, before | |
89 ;; limiting to the right articles. This is much too slow, of | |
90 ;; course. May issue a query for number of articles to fetch; you | |
91 ;; must accept the default of all articles at this point or things | |
92 ;; may break. | |
93 | |
94 ;; The Lisp setup involves setting a few variables and setting up the | |
95 ;; search engine. You can define the variables in the server definition | |
96 ;; like this : | |
97 ;; (setq gnus-secondary-select-methods '( | |
98 ;; (nnimap "" (nnimap-address "localhost") | |
99 ;; (nnir-search-engine hyrex) | |
100 ;; (nnir-hyrex-additional-switches ("-d" "ddl-nnimap.xml")) | |
101 ;; ))) | |
102 ;; Or you can define the global ones. The variables set in the mailer- | |
103 ;; definition will be used first. | |
104 ;; The variable to set is `nnir-search-engine'. Choose one of the engines | |
105 ;; listed in `nnir-engines'. (Actually `nnir-engines' is an alist, | |
106 ;; type `C-h v nnir-engines RET' for more information; this includes | |
107 ;; examples for setting `nnir-search-engine', too.) | |
108 ;; | |
109 ;; The variable nnir-mail-backend isn't used anymore. | |
110 ;; | |
111 | |
112 ;; You must also set up a search engine. I'll tell you about the two | |
113 ;; search engines currently supported: | |
114 | |
115 ;; 1. freeWAIS-sf | |
116 ;; | |
117 ;; As always with freeWAIS-sf, you need a so-called `format file'. I | |
118 ;; use the following file: | |
119 ;; | |
120 ;; ,----- | |
121 ;; | # Kai's format file for freeWAIS-sf for indexing mails. | |
122 ;; | # Each mail is in a file, much like the MH format. | |
123 ;; | | |
124 ;; | # Document separator should never match -- each file is a document. | |
125 ;; | record-sep: /^@this regex should never match@$/ | |
126 ;; | | |
127 ;; | # Searchable fields specification. | |
128 ;; | | |
129 ;; | region: /^[sS]ubject:/ /^[sS]ubject: */ | |
130 ;; | subject "Subject header" stemming TEXT BOTH | |
131 ;; | end: /^[^ \t]/ | |
132 ;; | | |
133 ;; | region: /^([tT][oO]|[cC][cC]):/ /^([tT][oO]|[cC][cC]): */ | |
134 ;; | to "To and Cc headers" SOUNDEX BOTH | |
135 ;; | end: /^[^ \t]/ | |
136 ;; | | |
137 ;; | region: /^[fF][rR][oO][mM]:/ /^[fF][rR][oO][mM]: */ | |
138 ;; | from "From header" SOUNDEX BOTH | |
139 ;; | end: /^[^ \t]/ | |
140 ;; | | |
141 ;; | region: /^$/ | |
142 ;; | stemming TEXT GLOBAL | |
143 ;; | end: /^@this regex should never match@$/ | |
144 ;; `----- | |
145 ;; | |
146 ;; 1998-07-22: waisindex would dump core on me for large articles with | |
147 ;; the above settings. I used /^$/ as the end regex for the global | |
148 ;; field. That seemed to work okay. | |
149 | |
150 ;; There is a Perl module called `WAIS.pm' which is available from | |
151 ;; CPAN as well as ls6-ftp.cs.uni-dortmund.de:/pub/wais/Perl. This | |
152 ;; module comes with a nifty tool called `makedb', which I use for | |
153 ;; indexing. Here's my `makedb.conf': | |
154 ;; | |
155 ;; ,----- | |
156 ;; | # Config file for makedb | |
157 ;; | | |
158 ;; | # Global options | |
159 ;; | waisindex = /usr/local/bin/waisindex | |
160 ;; | wais_opt = -stem -t fields | |
161 ;; | # `-stem' option necessary when `stemming' is specified for the | |
162 ;; | # global field in the *.fmt file | |
163 ;; | | |
164 ;; | # Own variables | |
165 ;; | homedir = /home/kai | |
166 ;; | | |
167 ;; | # The mail database. | |
168 ;; | database = mail | |
169 ;; | files = `find $homedir/Mail -name \*[0-9] -print` | |
170 ;; | dbdir = $homedir/.wais | |
171 ;; | limit = 100 | |
172 ;; `----- | |
173 ;; | |
174 ;; The Lisp setup involves the `nnir-wais-*' variables. The most | |
175 ;; difficult to understand variable is probably | |
176 ;; `nnir-wais-remove-prefix'. Here's what it does: the output of | |
177 ;; `waissearch' basically contains the file name and the (full) | |
178 ;; directory name. As Gnus works with group names rather than | |
179 ;; directory names, the directory name is transformed into a group | |
180 ;; name as follows: first, a prefix is removed from the (full) | |
181 ;; directory name, then all `/' are replaced with `.'. The variable | |
182 ;; `nnir-wais-remove-prefix' should contain a regex matching exactly | |
183 ;; this prefix. It defaults to `$HOME/Mail/' (note the trailing | |
184 ;; slash). | |
185 | |
186 ;; 2. Namazu | |
187 ;; | |
188 ;; The Namazu backend requires you to have one directory containing all | |
189 ;; index files, this is controlled by the `nnir-namazu-index-directory' | |
190 ;; variable. To function the `nnir-namazu-remove-prefix' variable must | |
191 ;; also be correct, see the documentation for `nnir-wais-remove-prefix' | |
192 ;; above. | |
193 ;; | |
194 ;; It is particularly important not to pass any any switches to namazu | |
195 ;; that will change the output format. Good switches to use include | |
196 ;; `--sort', `--ascending', `--early' and `--late'. Refer to the Namazu | |
197 ;; documentation for further information on valid switches. | |
198 ;; | |
199 ;; To index my mail with the `mknmz' program I use the following | |
200 ;; configuration file: | |
201 ;; | |
202 ;; ,---- | |
203 ;; | package conf; # Don't remove this line! | |
204 ;; | | |
205 ;; | # Paths which will not be indexed. Don't use `^' or `$' anchors. | |
206 ;; | $EXCLUDE_PATH = "spam|sent"; | |
207 ;; | | |
208 ;; | # Header fields which should be searchable. case-insensitive | |
209 ;; | $REMAIN_HEADER = "from|date|message-id|subject"; | |
210 ;; | | |
211 ;; | # Searchable fields. case-insensitive | |
212 ;; | $SEARCH_FIELD = "from|date|message-id|subject"; | |
213 ;; | | |
214 ;; | # The max length of a word. | |
215 ;; | $WORD_LENG_MAX = 128; | |
216 ;; | | |
217 ;; | # The max length of a field. | |
218 ;; | $MAX_FIELD_LENGTH = 256; | |
219 ;; `---- | |
220 ;; | |
221 ;; My mail is stored in the directories ~/Mail/mail/, ~/Mail/lists/ and | |
222 ;; ~/Mail/archive/, so to index them I go to the directory set in | |
223 ;; `nnir-namazu-index-directory' and issue the following command. | |
224 ;; | |
225 ;; mknmz --mailnews ~/Mail/archive/ ~/Mail/mail/ ~/Mail/lists/ | |
226 ;; | |
227 ;; For maximum searching efficiency I have a cron job set to run this | |
228 ;; command every four hours. | |
229 | |
230 ;; 3. HyREX | |
231 ;; | |
232 ;; The HyREX backend requires you to have one directory from where all | |
233 ;; your relative paths are to, if you use them. This directory must be | |
234 ;; set in the `nnir-hyrex-index-directory' variable, which defaults to | |
235 ;; your home directory. You must also pass the base, class and | |
236 ;; directory options or simply your dll to the `nnir-hyrex-programm' by | |
237 ;; setting the `nnir-hyrex-additional-switches' variable accordently. | |
238 ;; To function the `nnir-hyrex-remove-prefix' variable must also be | |
239 ;; correct, see the documentation for `nnir-wais-remove-prefix' above. | |
240 | |
241 ;; 4. find-grep | |
242 ;; | |
243 ;; The find-grep engine simply runs find(1) to locate eligible | |
244 ;; articles and searches them with grep(1). This, of course, is much | |
245 ;; slower than using a proper search engine but OTOH doesn't require | |
246 ;; maintenance of an index and is still faster than using any built-in | |
247 ;; means for searching. The method specification of the server to | |
248 ;; search must include a directory for this engine to work (E.g., | |
249 ;; `nnml-directory'). The tools must be POSIX compliant. GNU Find | |
250 ;; prior to version 4.2.12 (4.2.26 on Linux due to incorrect ARG_MAX | |
251 ;; handling) does not work. | |
252 ;; ,---- | |
253 ;; | ;; find-grep configuration for searching the Gnus Cache | |
254 ;; | | |
255 ;; | (nnml "cache" | |
256 ;; | (nnml-get-new-mail nil) | |
257 ;; | (nnir-search-engine find-grep) | |
258 ;; | (nnml-directory "~/News/cache/") | |
259 ;; | (nnml-active-file "~/News/cache/active")) | |
260 ;; `---- | |
261 | |
262 ;; Developer information: | |
263 | |
264 ;; I have tried to make the code expandable. Basically, it is divided | |
265 ;; into two layers. The upper layer is somewhat like the `nnvirtual' | |
110100
8f51744211eb
Remove nndb, nnkiboze and related code.
Katsumi Yamaoka <yamaoka@jpl.org>
parents:
107978
diff
changeset
|
266 ;; backend: given a specification of what articles to show from |
8f51744211eb
Remove nndb, nnkiboze and related code.
Katsumi Yamaoka <yamaoka@jpl.org>
parents:
107978
diff
changeset
|
267 ;; another backend, it creates a group containing exactly those |
8f51744211eb
Remove nndb, nnkiboze and related code.
Katsumi Yamaoka <yamaoka@jpl.org>
parents:
107978
diff
changeset
|
268 ;; articles. The lower layer issues a query to a search engine and |
8f51744211eb
Remove nndb, nnkiboze and related code.
Katsumi Yamaoka <yamaoka@jpl.org>
parents:
107978
diff
changeset
|
269 ;; produces such a specification of what articles to show from the |
95610 | 270 ;; other backend. |
271 | |
272 ;; The interface between the two layers consists of the single | |
273 ;; function `nnir-run-query', which just selects the appropriate | |
274 ;; function for the search engine one is using. The input to | |
275 ;; `nnir-run-query' is a string, representing the query as input by | |
276 ;; the user. The output of `nnir-run-query' is supposed to be a | |
277 ;; vector, each element of which should in turn be a three-element | |
278 ;; vector. The first element should be full group name of the article, | |
279 ;; the second element should be the article number, and the third | |
280 ;; element should be the Retrieval Status Value (RSV) as returned from | |
281 ;; the search engine. An RSV is the score assigned to the document by | |
282 ;; the search engine. For Boolean search engines, the | |
283 ;; RSV is always 1000 (or 1 or 100, or whatever you like). | |
284 | |
285 ;; The sorting order of the articles in the summary buffer created by | |
286 ;; nnir is based on the order of the articles in the above mentioned | |
287 ;; vector, so that's where you can do the sorting you'd like. Maybe | |
288 ;; it would be nice to have a way of displaying the search result | |
289 ;; sorted differently? | |
290 | |
291 ;; So what do you need to do when you want to add another search | |
292 ;; engine? You write a function that executes the query. Temporary | |
293 ;; data from the search engine can be put in `nnir-tmp-buffer'. This | |
294 ;; function should return the list of articles as a vector, as | |
295 ;; described above. Then, you need to register this backend in | |
296 ;; `nnir-engines'. Then, users can choose the backend by setting | |
297 ;; `nnir-search-engine'. | |
298 | |
299 ;; Todo, or future ideas: | |
300 | |
301 ;; * It should be possible to restrict search to certain groups. | |
302 ;; | |
303 ;; * There is currently no error checking. | |
304 ;; | |
305 ;; * The summary buffer display is currently really ugly, with all the | |
306 ;; added information in the subjects. How could I make this | |
307 ;; prettier? | |
308 ;; | |
309 ;; * A function which can be called from an nnir summary buffer which | |
310 ;; teleports you into the group the current article came from and | |
311 ;; shows you the whole thread this article is part of. | |
312 ;; Implementation suggestions? | |
313 ;; (1998-07-24: There is now a preliminary implementation, but | |
314 ;; it is much too slow and quite fragile.) | |
315 ;; | |
316 ;; * Support other mail backends. In particular, probably quite a few | |
317 ;; people use nnfolder. How would one go about searching nnfolders | |
318 ;; and producing the right data needed? The group name and the RSV | |
319 ;; are simple, but what about the article number? | |
320 ;; - The article number is encoded in the `X-Gnus-Article-Number' | |
321 ;; header of each mail. | |
322 ;; - The HyREX engine supports nnfolder. | |
323 ;; | |
324 ;; * Support compressed mail files. Probably, just stripping off the | |
325 ;; `.gz' or `.Z' file name extension is sufficient. | |
326 ;; | |
327 ;; * At least for imap, the query is performed twice. | |
328 ;; | |
329 | |
330 ;; Have you got other ideas? | |
331 | |
332 ;;; Setup Code: | |
333 | |
334 (require 'nnoo) | |
335 (require 'gnus-group) | |
336 (require 'gnus-sum) | |
337 (require 'message) | |
338 (require 'gnus-util) | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
339 (eval-when-compile |
95610 | 340 (require 'cl)) |
341 | |
342 (nnoo-declare nnir) | |
343 (nnoo-define-basics nnir) | |
344 | |
345 (gnus-declare-backend "nnir" 'mail) | |
346 | |
347 (defvar nnir-imap-search-field "TEXT" | |
348 "The IMAP search item when doing an nnir search") | |
349 | |
350 (defvar nnir-imap-search-arguments | |
351 '(("Whole message" . "TEXT") | |
352 ("Subject" . "SUBJECT") | |
353 ("To" . "TO") | |
354 ("From" . "FROM") | |
355 (nil . "HEADER \"%s\"")) | |
356 "Mapping from user readable strings to IMAP search items for use in nnir") | |
357 | |
358 (defvar nnir-imap-search-argument-history () | |
359 "The history for querying search options in nnir") | |
360 | |
107473 | 361 (defvar nnir-get-article-nov-override-function nil |
362 "If non-nil, a function that will be passed each search result. This | |
363 should return a message's headers in NOV format. | |
364 | |
365 If this variable is nil, or if the provided function returns nil for a search | |
366 result, `gnus-retrieve-headers' will be called instead.") | |
367 | |
368 | |
95610 | 369 ;;; Developer Extension Variable: |
370 | |
371 (defvar nnir-engines | |
372 `((wais nnir-run-waissearch | |
373 ()) | |
374 (imap nnir-run-imap | |
95616 | 375 ((criteria |
95610 | 376 "Search in: " ; Prompt |
377 ,nnir-imap-search-arguments ; alist for completing | |
378 nil ; no filtering | |
379 nil ; allow any user input | |
380 nil ; initial value | |
381 nnir-imap-search-argument-history ; the history to use | |
382 ,nnir-imap-search-field ; default | |
383 ))) | |
384 (swish++ nnir-run-swish++ | |
385 ((group . "Group spec: "))) | |
386 (swish-e nnir-run-swish-e | |
387 ((group . "Group spec: "))) | |
388 (namazu nnir-run-namazu | |
389 ()) | |
390 (hyrex nnir-run-hyrex | |
391 ((group . "Group spec: "))) | |
392 (find-grep nnir-run-find-grep | |
393 ((grep-options . "Grep options: ")))) | |
394 "Alist of supported search engines. | |
395 Each element in the alist is a three-element list (ENGINE FUNCTION ARGS). | |
396 ENGINE is a symbol designating the searching engine. FUNCTION is also | |
397 a symbol, giving the function that does the search. The third element | |
398 ARGS is a list of cons pairs (PARAM . PROMPT). When issuing a query, | |
399 the FUNCTION will issue a query for each of the PARAMs, using PROMPT. | |
400 | |
401 The value of `nnir-search-engine' must be one of the ENGINE symbols. | |
402 For example, use the following line for searching using freeWAIS-sf: | |
403 (setq nnir-search-engine 'wais) | |
404 Use the following line if you read your mail via IMAP and your IMAP | |
405 server supports searching: | |
406 (setq nnir-search-engine 'imap) | |
407 Note that you have to set additional variables for most backends. For | |
408 example, the `wais' backend needs the variables `nnir-wais-program', | |
409 `nnir-wais-database' and `nnir-wais-remove-prefix'. | |
410 | |
411 Add an entry here when adding a new search engine.") | |
412 | |
413 ;;; User Customizable Variables: | |
414 | |
415 (defgroup nnir nil | |
416 "Search nnmh and nnml groups in Gnus with swish-e, freeWAIS-sf, or EWS." | |
417 :group 'gnus) | |
418 | |
419 ;; Mail backend. | |
420 | |
421 ;; TODO: | |
422 ;; If `nil', use server parameters to find out which server to search. CCC | |
423 ;; | |
424 (defcustom nnir-mail-backend '(nnml "") | |
425 "*Specifies which backend should be searched. | |
426 More precisely, this is used to determine from which backend to fetch the | |
427 messages found. | |
428 | |
429 This must be equal to an existing server, so maybe it is best to use | |
430 something like the following: | |
431 (setq nnir-mail-backend (nth 0 gnus-secondary-select-methods)) | |
432 The above line works fine if the mail backend you want to search is | |
433 the first element of gnus-secondary-select-methods (`nth' starts counting | |
434 at zero)." | |
435 :type '(sexp) | |
436 :group 'nnir) | |
437 | |
438 ;; Search engine to use. | |
439 | |
440 (defcustom nnir-search-engine 'wais | |
441 "*The search engine to use. Must be a symbol. | |
442 See `nnir-engines' for a list of supported engines, and for example | |
443 settings of `nnir-search-engine'." | |
444 :type '(sexp) | |
445 :group 'nnir) | |
446 | |
447 ;; freeWAIS-sf. | |
448 | |
449 (defcustom nnir-wais-program "waissearch" | |
450 "*Name of waissearch executable." | |
451 :type '(string) | |
452 :group 'nnir) | |
453 | |
454 (defcustom nnir-wais-database (expand-file-name "~/.wais/mail") | |
455 "*Name of Wais database containing the mail. | |
456 | |
457 Note that this should be a file name without extension. For example, | |
458 if you have a file /home/john/.wais/mail.fmt, use this: | |
459 (setq nnir-wais-database \"/home/john/.wais/mail\") | |
460 The string given here is passed to `waissearch -d' as-is." | |
461 :type '(file) | |
462 :group 'nnir) | |
463 | |
464 (defcustom nnir-wais-remove-prefix (concat (getenv "HOME") "/Mail/") | |
465 "*The prefix to remove from each directory name returned by waissearch | |
466 in order to get a group name (albeit with / instead of .). This is a | |
467 regular expression. | |
468 | |
469 For example, suppose that Wais returns file names such as | |
470 \"/home/john/Mail/mail/misc/42\". For this example, use the following | |
471 setting: (setq nnir-wais-remove-prefix \"/home/john/Mail/\") | |
472 Note the trailing slash. Removing this prefix gives \"mail/misc/42\". | |
473 `nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to | |
474 arrive at the correct group name, \"mail.misc\"." | |
475 :type '(regexp) | |
476 :group 'nnir) | |
477 | |
478 (defcustom nnir-swish++-configuration-file | |
479 (expand-file-name "~/Mail/swish++.conf") | |
480 "*Configuration file for swish++." | |
481 :type '(file) | |
482 :group 'nnir) | |
483 | |
484 (defcustom nnir-swish++-program "search" | |
485 "*Name of swish++ search executable." | |
486 :type '(string) | |
487 :group 'nnir) | |
488 | |
489 (defcustom nnir-swish++-additional-switches '() | |
490 "*A list of strings, to be given as additional arguments to swish++. | |
491 | |
492 Note that this should be a list. Ie, do NOT use the following: | |
493 (setq nnir-swish++-additional-switches \"-i -w\") ; wrong | |
494 Instead, use this: | |
495 (setq nnir-swish++-additional-switches '(\"-i\" \"-w\"))" | |
496 :type '(repeat (string)) | |
497 :group 'nnir) | |
498 | |
499 (defcustom nnir-swish++-remove-prefix (concat (getenv "HOME") "/Mail/") | |
500 "*The prefix to remove from each file name returned by swish++ | |
501 in order to get a group name (albeit with / instead of .). This is a | |
502 regular expression. | |
503 | |
504 This variable is very similar to `nnir-wais-remove-prefix', except | |
505 that it is for swish++, not Wais." | |
506 :type '(regexp) | |
507 :group 'nnir) | |
508 | |
509 ;; Swish-E. | |
510 ;; URL: http://sunsite.berkeley.edu/SWISH-E/ | |
511 ;; New version: http://www.boe.es/swish-e | |
512 ;; Variables `nnir-swish-e-index-file', `nnir-swish-e-program' and | |
513 ;; `nnir-swish-e-additional-switches' | |
514 | |
515 (make-obsolete-variable 'nnir-swish-e-index-file | |
104700
9866b069527c
* spam.el (spam-ifile-path, spam-bogofilter-path, spam-sa-learn-path)
Stefan Monnier <monnier@iro.umontreal.ca>
parents:
104584
diff
changeset
|
516 'nnir-swish-e-index-files "Emacs 23.1") |
95610 | 517 (defcustom nnir-swish-e-index-file |
518 (expand-file-name "~/Mail/index.swish-e") | |
519 "*Index file for swish-e. | |
520 This could be a server parameter. | |
521 It is never consulted once `nnir-swish-e-index-files', which should be | |
522 used instead, has been customized." | |
523 :type '(file) | |
524 :group 'nnir) | |
525 | |
526 (defcustom nnir-swish-e-index-files | |
527 (list nnir-swish-e-index-file) | |
528 "*List of index files for swish-e. | |
529 This could be a server parameter." | |
530 :type '(repeat (file)) | |
531 :group 'nnir) | |
532 | |
533 (defcustom nnir-swish-e-program "swish-e" | |
534 "*Name of swish-e search executable. | |
535 This cannot be a server parameter." | |
536 :type '(string) | |
537 :group 'nnir) | |
538 | |
539 (defcustom nnir-swish-e-additional-switches '() | |
540 "*A list of strings, to be given as additional arguments to swish-e. | |
541 | |
542 Note that this should be a list. Ie, do NOT use the following: | |
543 (setq nnir-swish-e-additional-switches \"-i -w\") ; wrong | |
544 Instead, use this: | |
545 (setq nnir-swish-e-additional-switches '(\"-i\" \"-w\")) | |
546 | |
547 This could be a server parameter." | |
548 :type '(repeat (string)) | |
549 :group 'nnir) | |
550 | |
551 (defcustom nnir-swish-e-remove-prefix (concat (getenv "HOME") "/Mail/") | |
552 "*The prefix to remove from each file name returned by swish-e | |
553 in order to get a group name (albeit with / instead of .). This is a | |
554 regular expression. | |
555 | |
556 This variable is very similar to `nnir-wais-remove-prefix', except | |
557 that it is for swish-e, not Wais. | |
558 | |
559 This could be a server parameter." | |
560 :type '(regexp) | |
561 :group 'nnir) | |
562 | |
563 ;; HyREX engine, see <URL:http://ls6-www.cs.uni-dortmund.de/> | |
564 | |
565 (defcustom nnir-hyrex-program "nnir-search" | |
566 "*Name of the nnir-search executable." | |
567 :type '(string) | |
568 :group 'nnir) | |
569 | |
570 (defcustom nnir-hyrex-additional-switches '() | |
571 "*A list of strings, to be given as additional arguments for nnir-search. | |
572 Note that this should be a list. Ie, do NOT use the following: | |
573 (setq nnir-hyrex-additional-switches \"-ddl ddl.xml -c nnir\") ; wrong ! | |
574 Instead, use this: | |
575 (setq nnir-hyrex-additional-switches '(\"-ddl\" \"ddl.xml\" \"-c\" \"nnir\"))" | |
576 :type '(repeat (string)) | |
577 :group 'nnir) | |
578 | |
579 (defcustom nnir-hyrex-index-directory (getenv "HOME") | |
580 "*Index directory for HyREX." | |
581 :type '(directory) | |
582 :group 'nnir) | |
583 | |
584 (defcustom nnir-hyrex-remove-prefix (concat (getenv "HOME") "/Mail/") | |
585 "*The prefix to remove from each file name returned by HyREX | |
586 in order to get a group name (albeit with / instead of .). | |
587 | |
588 For example, suppose that HyREX returns file names such as | |
589 \"/home/john/Mail/mail/misc/42\". For this example, use the following | |
590 setting: (setq nnir-hyrex-remove-prefix \"/home/john/Mail/\") | |
591 Note the trailing slash. Removing this prefix gives \"mail/misc/42\". | |
592 `nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to | |
593 arrive at the correct group name, \"mail.misc\"." | |
594 :type '(directory) | |
595 :group 'nnir) | |
596 | |
597 ;; Namazu engine, see <URL:http://ww.namazu.org/> | |
598 | |
599 (defcustom nnir-namazu-program "namazu" | |
600 "*Name of Namazu search executable." | |
601 :type '(string) | |
602 :group 'nnir) | |
603 | |
604 (defcustom nnir-namazu-index-directory (expand-file-name "~/Mail/namazu/") | |
605 "*Index directory for Namazu." | |
606 :type '(directory) | |
607 :group 'nnir) | |
608 | |
609 (defcustom nnir-namazu-additional-switches '() | |
610 "*A list of strings, to be given as additional arguments to namazu. | |
611 The switches `-q', `-a', and `-s' are always used, very few other switches | |
612 make any sense in this context. | |
613 | |
614 Note that this should be a list. Ie, do NOT use the following: | |
615 (setq nnir-namazu-additional-switches \"-i -w\") ; wrong | |
616 Instead, use this: | |
617 (setq nnir-namazu-additional-switches '(\"-i\" \"-w\"))" | |
618 :type '(repeat (string)) | |
619 :group 'nnir) | |
620 | |
621 (defcustom nnir-namazu-remove-prefix (concat (getenv "HOME") "/Mail/") | |
622 "*The prefix to remove from each file name returned by Namazu | |
623 in order to get a group name (albeit with / instead of .). | |
624 | |
625 This variable is very similar to `nnir-wais-remove-prefix', except | |
626 that it is for Namazu, not Wais." | |
627 :type '(directory) | |
628 :group 'nnir) | |
629 | |
630 ;;; Internal Variables: | |
631 | |
632 (defvar nnir-current-query nil | |
633 "Internal: stores current query (= group name).") | |
634 | |
635 (defvar nnir-current-server nil | |
636 "Internal: stores current server (does it ever change?).") | |
637 | |
638 (defvar nnir-current-group-marked nil | |
639 "Internal: stores current list of process-marked groups.") | |
640 | |
641 (defvar nnir-artlist nil | |
642 "Internal: stores search result.") | |
643 | |
644 (defvar nnir-tmp-buffer " *nnir*" | |
645 "Internal: temporary buffer.") | |
646 | |
647 ;;; Code: | |
648 | |
649 ;; Gnus glue. | |
650 | |
651 (defun gnus-group-make-nnir-group (extra-parms query) | |
652 "Create an nnir group. Asks for query." | |
653 (interactive "P\nsQuery: ") | |
654 (setq nnir-current-query nil | |
655 nnir-current-server nil | |
656 nnir-current-group-marked nil | |
657 nnir-artlist nil) | |
658 (let ((parms nil)) | |
659 (if extra-parms | |
660 (setq parms (nnir-read-parms query)) | |
661 (setq parms (list (cons 'query query)))) | |
662 (add-to-list 'parms (cons 'unique-id (message-unique-id)) t) | |
663 (gnus-group-read-ephemeral-group | |
664 (concat "nnir:" (prin1-to-string parms)) '(nnir "") t | |
665 (cons (current-buffer) | |
666 gnus-current-window-configuration) | |
667 nil))) | |
668 | |
95830 | 669 (eval-when-compile |
670 (when (featurep 'xemacs) | |
671 ;; The `kbd' macro requires that the `read-kbd-macro' macro is available. | |
672 (require 'edmacro))) | |
673 | |
95610 | 674 (defun nnir-group-mode-hook () |
675 (define-key gnus-group-mode-map (kbd "G G") | |
676 'gnus-group-make-nnir-group)) | |
677 (add-hook 'gnus-group-mode-hook 'nnir-group-mode-hook) | |
678 | |
679 ;; Why is this needed? Is this for compatibility with old/new gnusae? Using | |
680 ;; gnus-group-server instead works for me. -- Justus Piater | |
681 (defmacro nnir-group-server (group) | |
682 "Return the server for a newsgroup GROUP. | |
683 The returned format is as `gnus-server-to-method' needs it. See | |
684 `gnus-group-real-prefix' and `gnus-group-real-name'." | |
685 `(let ((gname ,group)) | |
686 (if (string-match "^\\([^:]+\\):" gname) | |
687 (progn | |
688 (setq gname (match-string 1 gname)) | |
689 (if (string-match "^\\([^+]+\\)\\+\\(.+\\)$" gname) | |
690 (format "%s:%s" (match-string 1 gname) (match-string 2 gname)) | |
691 (concat gname ":"))) | |
692 (format "%s:%s" (car gnus-select-method) (cadr gnus-select-method))))) | |
693 | |
694 ;; Summary mode commands. | |
695 | |
696 (defun gnus-summary-nnir-goto-thread () | |
697 "Only applies to nnir groups. Go to group this article came from | |
698 and show thread that contains this article." | |
699 (interactive) | |
700 (unless (eq 'nnir (car (gnus-find-method-for-group gnus-newsgroup-name))) | |
105167
e14dfafc8256
* gnus-art.el (gnus-article-encrypt-body):
Juanma Barranquero <lekktu@gmail.com>
parents:
104700
diff
changeset
|
701 (error "Can't execute this command unless in nnir group")) |
95610 | 702 (let* ((cur (gnus-summary-article-number)) |
703 (group (nnir-artlist-artitem-group nnir-artlist cur)) | |
704 (backend-number (nnir-artlist-artitem-number nnir-artlist cur)) | |
705 server backend-group) | |
706 (setq server (nnir-group-server group)) | |
707 (setq backend-group (gnus-group-real-name group)) | |
708 (gnus-group-read-ephemeral-group | |
709 backend-group | |
710 (gnus-server-to-method server) | |
711 t ; activate | |
712 (cons (current-buffer) | |
713 'summary) ; window config | |
714 nil | |
715 (list backend-number)) | |
716 (gnus-summary-limit (list backend-number)) | |
717 (gnus-summary-refer-thread))) | |
718 | |
719 (if (fboundp 'eval-after-load) | |
720 (eval-after-load "gnus-sum" | |
721 '(define-key gnus-summary-goto-map | |
722 "T" 'gnus-summary-nnir-goto-thread)) | |
723 (add-hook 'gnus-summary-mode-hook | |
724 (function (lambda () | |
725 (define-key gnus-summary-goto-map | |
726 "T" 'gnus-summary-nnir-goto-thread))))) | |
727 | |
728 | |
729 | |
730 ;; Gnus backend interface functions. | |
731 | |
732 (deffoo nnir-open-server (server &optional definitions) | |
733 ;; Just set the server variables appropriately. | |
734 (nnoo-change-server 'nnir server definitions)) | |
735 | |
736 (deffoo nnir-request-group (group &optional server fast) | |
737 "GROUP is the query string." | |
738 (nnir-possibly-change-server server) | |
739 ;; Check for cache and return that if appropriate. | |
740 (if (and (equal group nnir-current-query) | |
741 (equal gnus-group-marked nnir-current-group-marked) | |
742 (or (null server) | |
743 (equal server nnir-current-server))) | |
744 nnir-artlist | |
745 ;; Cache miss. | |
746 (setq nnir-artlist (nnir-run-query group))) | |
747 (save-excursion | |
748 (set-buffer nntp-server-buffer) | |
749 (if (zerop (length nnir-artlist)) | |
750 (progn | |
751 (setq nnir-current-query nil | |
752 nnir-current-server nil | |
753 nnir-current-group-marked nil | |
754 nnir-artlist nil) | |
755 (nnheader-report 'nnir "Search produced empty results.")) | |
756 ;; Remember data for cache. | |
757 (setq nnir-current-query group) | |
758 (when server (setq nnir-current-server server)) | |
759 (setq nnir-current-group-marked gnus-group-marked) | |
760 (nnheader-insert "211 %d %d %d %s\n" | |
761 (nnir-artlist-length nnir-artlist) ; total # | |
762 1 ; first # | |
763 (nnir-artlist-length nnir-artlist) ; last # | |
764 group)))) ; group name | |
765 | |
766 (deffoo nnir-retrieve-headers (articles &optional group server fetch-old) | |
767 (save-excursion | |
768 (let ((artlist (copy-sequence articles)) | |
769 art artitem artgroup artno artrsv artfullgroup | |
770 novitem novdata foo server) | |
771 (while (not (null artlist)) | |
772 (setq art (car artlist)) | |
773 (or (numberp art) | |
774 (nnheader-report | |
775 'nnir | |
776 "nnir-retrieve-headers doesn't grok message ids: %s" | |
777 art)) | |
778 (setq artitem (nnir-artlist-article nnir-artlist art)) | |
779 (setq artrsv (nnir-artitem-rsv artitem)) | |
780 (setq artfullgroup (nnir-artitem-group artitem)) | |
781 (setq artno (nnir-artitem-number artitem)) | |
782 (setq artgroup (gnus-group-real-name artfullgroup)) | |
783 (setq server (nnir-group-server artfullgroup)) | |
784 ;; retrieve NOV or HEAD data for this article, transform into | |
785 ;; NOV data and prepend to `novdata' | |
786 (set-buffer nntp-server-buffer) | |
787 (nnir-possibly-change-server server) | |
788 (let ((gnus-override-method | |
789 (gnus-server-to-method server))) | |
107473 | 790 ;; if nnir-get-article-nov-override-function is set, use it |
791 (if nnir-get-article-nov-override-function | |
792 (setq novitem (funcall nnir-get-article-nov-override-function | |
793 artitem)) | |
794 ;; else, set novitem through nnheader-parse-nov/nnheader-parse-head | |
110111
5b9f64b04a04
Delete all trailing white space.
Katsumi Yamaoka <yamaoka@jpl.org>
parents:
110102
diff
changeset
|
795 (case (setq foo (gnus-retrieve-headers (list artno) |
107473 | 796 artfullgroup nil)) |
797 (nov | |
798 (goto-char (point-min)) | |
799 (setq novitem (nnheader-parse-nov)) | |
800 (unless novitem | |
801 (pop-to-buffer nntp-server-buffer) | |
802 (error | |
803 "nnheader-parse-nov returned nil for article %s in group %s" | |
804 artno artfullgroup))) | |
805 (headers | |
806 (goto-char (point-min)) | |
807 (setq novitem (nnheader-parse-head)) | |
808 (unless novitem | |
809 (pop-to-buffer nntp-server-buffer) | |
810 (error | |
811 "nnheader-parse-head returned nil for article %s in group %s" | |
812 artno artfullgroup))) | |
813 (t (error "Unknown header type %s while requesting article %s of group %s" | |
814 foo artno artfullgroup))))) | |
95610 | 815 ;; replace article number in original group with article number |
816 ;; in nnir group | |
817 (mail-header-set-number novitem art) | |
818 (mail-header-set-from novitem | |
819 (mail-header-from novitem)) | |
820 (mail-header-set-subject | |
821 novitem | |
822 (format "[%d: %s/%d] %s" | |
823 artrsv artgroup artno | |
824 (mail-header-subject novitem))) | |
825 ;;-(mail-header-set-extra novitem nil) | |
826 (push novitem novdata) | |
827 (setq artlist (cdr artlist))) | |
828 (setq novdata (nreverse novdata)) | |
829 (set-buffer nntp-server-buffer) (erase-buffer) | |
830 (mapc 'nnheader-insert-nov novdata) | |
831 'nov))) | |
832 | |
833 (deffoo nnir-request-article (article | |
834 &optional group server to-buffer) | |
835 (if (stringp article) | |
836 (nnheader-report | |
837 'nnir | |
838 "nnir-retrieve-headers doesn't grok message ids: %s" | |
839 article) | |
840 (save-excursion | |
841 (let* ((artitem (nnir-artlist-article nnir-artlist | |
842 article)) | |
843 (artfullgroup (nnir-artitem-group artitem)) | |
844 (artno (nnir-artitem-number artitem)) | |
845 ;; Bug? | |
846 ;; Why must we bind nntp-server-buffer here? It won't | |
847 ;; work if `buf' is used, say. (Of course, the set-buffer | |
848 ;; line below must then be updated, too.) | |
849 (nntp-server-buffer (or to-buffer nntp-server-buffer))) | |
850 (set-buffer nntp-server-buffer) | |
851 (erase-buffer) | |
852 (message "Requesting article %d from group %s" | |
853 artno artfullgroup) | |
854 (gnus-request-article artno artfullgroup nntp-server-buffer) | |
855 (cons artfullgroup artno))))) | |
856 | |
857 | |
858 (nnoo-define-skeleton nnir) | |
859 | |
860 | |
861 (defmacro nnir-add-result (dirnam artno score prefix server artlist) | |
95616 | 862 "Ask `nnir-compose-result' to construct a result vector, |
95610 | 863 and if it is non-nil, add it to artlist." |
864 `(let ((result (nnir-compose-result ,dirnam ,artno ,score ,prefix ,server))) | |
865 (when (not (null result)) | |
866 (push result ,artlist)))) | |
867 | |
868 (autoload 'nnmaildir-base-name-to-article-number "nnmaildir") | |
869 | |
870 ;; Helper function currently used by the Swish++ and Namazu backends; | |
871 ;; perhaps useful for other backends as well | |
872 (defun nnir-compose-result (dirnam article score prefix server) | |
873 "Extract the group from dirnam, and create a result vector | |
874 ready to be added to the list of search results." | |
875 | |
876 ;; remove nnir-*-remove-prefix from beginning of dirnam filename | |
877 (when (string-match (concat "^" prefix) dirnam) | |
878 (setq dirnam (replace-match "" t t dirnam))) | |
879 | |
880 (when (file-readable-p (concat prefix dirnam article)) | |
881 ;; remove trailing slash and, for nnmaildir, cur/new/tmp | |
882 (setq dirnam | |
883 (substring dirnam 0 (if (string= server "nnmaildir:") -5 -1))) | |
884 | |
885 ;; Set group to dirnam without any leading dots or slashes, | |
886 ;; and with all subsequent slashes replaced by dots | |
887 (let ((group (gnus-replace-in-string | |
888 (gnus-replace-in-string dirnam "^[./\\]" "" t) | |
889 "[/\\]" "." t))) | |
890 | |
891 (vector (nnir-group-full-name group server) | |
892 (if (string= server "nnmaildir:") | |
893 (nnmaildir-base-name-to-article-number | |
894 (substring article 0 (string-match ":" article)) | |
895 group nil) | |
896 (string-to-number article)) | |
897 (string-to-number score))))) | |
898 | |
899 ;;; Search Engine Interfaces: | |
900 | |
901 ;; freeWAIS-sf interface. | |
902 (defun nnir-run-waissearch (query server &optional group) | |
903 "Run given query agains waissearch. Returns vector of (group name, file name) | |
904 pairs (also vectors, actually)." | |
905 (when group | |
105167
e14dfafc8256
* gnus-art.el (gnus-article-encrypt-body):
Juanma Barranquero <lekktu@gmail.com>
parents:
104700
diff
changeset
|
906 (error "The freeWAIS-sf backend cannot search specific groups")) |
95610 | 907 (save-excursion |
908 (let ((qstring (cdr (assq 'query query))) | |
909 (prefix (nnir-read-server-parm 'nnir-wais-remove-prefix server)) | |
910 artlist score artno dirnam) | |
911 (set-buffer (get-buffer-create nnir-tmp-buffer)) | |
912 (erase-buffer) | |
913 (message "Doing WAIS query %s..." query) | |
914 (call-process nnir-wais-program | |
915 nil ; input from /dev/null | |
916 t ; output to current buffer | |
917 nil ; don't redisplay | |
918 "-d" (nnir-read-server-parm 'nnir-wais-database server) ; database to search | |
919 qstring) | |
920 (message "Massaging waissearch output...") | |
921 ;; remove superfluous lines | |
922 (keep-lines "Score:") | |
923 ;; extract data from result lines | |
924 (goto-char (point-min)) | |
925 (while (re-search-forward | |
926 "Score: +\\([0-9]+\\).*'\\([0-9]+\\) +\\([^']+\\)/'" nil t) | |
927 (setq score (match-string 1) | |
928 artno (match-string 2) | |
929 dirnam (match-string 3)) | |
930 (unless (string-match prefix dirnam) | |
931 (nnheader-report 'nnir "Dir name %s doesn't contain prefix %s" | |
932 dirnam prefix)) | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
933 (setq group (gnus-replace-in-string |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
934 (replace-match "" t t dirnam) "/" ".")) |
95610 | 935 (push (vector (nnir-group-full-name group server) |
936 (string-to-number artno) | |
937 (string-to-number score)) | |
938 artlist)) | |
939 (message "Massaging waissearch output...done") | |
940 (apply 'vector | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
941 (sort artlist |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
942 (function (lambda (x y) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
943 (> (nnir-artitem-rsv x) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
944 (nnir-artitem-rsv y))))))))) |
95610 | 945 |
946 ;; IMAP interface. | |
947 ;; todo: | |
948 ;; nnir invokes this two (2) times???! | |
949 ;; we should not use nnimap at all but open our own server connection | |
950 ;; we should not LIST * but use nnimap-list-pattern from defs | |
951 ;; send queries as literals | |
952 ;; handle errors | |
953 | |
954 (autoload 'nnimap-open-server "nnimap") | |
955 (defvar nnimap-server-buffer) ;; nnimap.el | |
956 (autoload 'imap-mailbox-select "imap") | |
957 (autoload 'imap-search "imap") | |
958 (autoload 'imap-quote-specials "imap") | |
959 | |
960 (defun nnir-run-imap (query srv &optional group-option) | |
961 "Run a search against an IMAP back-end server. | |
962 This uses a custom query language parser; see `nnir-imap-make-query' for | |
963 details on the language and supported extensions" | |
964 (save-excursion | |
965 (let ((qstring (cdr (assq 'query query))) | |
966 (server (cadr (gnus-server-to-method srv))) | |
967 (group (or group-option (gnus-group-group-name))) | |
968 (defs (caddr (gnus-server-to-method srv))) | |
969 (criteria (or (cdr (assq 'criteria query)) | |
970 nnir-imap-search-field)) | |
971 artlist buf) | |
972 (message "Opening server %s" server) | |
973 (condition-case () | |
974 (when (nnimap-open-server server defs) ;; xxx | |
975 (setq buf nnimap-server-buffer) ;; xxx | |
976 (message "Searching %s..." group) | |
977 (let ((arts 0) | |
978 (mbx (gnus-group-real-name group))) | |
979 (when (imap-mailbox-select mbx nil buf) | |
980 (mapc | |
981 (lambda (artnum) | |
982 (push (vector group artnum 1) artlist) | |
983 (setq arts (1+ arts))) | |
984 (imap-search (nnir-imap-make-query criteria qstring) buf)) | |
985 (message "Searching %s... %d matches" mbx arts))) | |
986 (message "Searching %s...done" group)) | |
987 (quit nil)) | |
988 (reverse artlist)))) | |
989 | |
990 (defun nnir-imap-make-query (criteria qstring) | |
991 "Parse the query string and criteria into an appropriate IMAP search | |
992 expression, returning the string query to make. | |
993 | |
994 This implements a little language designed to return the expected results | |
995 to an arbitrary query string to the end user. | |
996 | |
997 The search is always case-insensitive, as defined by RFC2060, and supports | |
95616 | 998 the following features (inspired by the Google search input language): |
95610 | 999 |
1000 Automatic \"and\" queries | |
1001 If you specify multiple words then they will be treated as an \"and\" | |
1002 expression intended to match all components. | |
1003 | |
1004 Phrase searches | |
1005 If you wrap your query in double-quotes then it will be treated as a | |
1006 literal string. | |
1007 | |
1008 Negative terms | |
1009 If you precede a term with \"-\" then it will negate that. | |
1010 | |
1011 \"OR\" queries | |
1012 If you include an upper-case \"OR\" in your search it will cause the | |
1013 term before it and the term after it to be treated as alternatives. | |
1014 | |
1015 In future the following will be added to the language: | |
1016 * support for date matches | |
1017 * support for location of text matching within the query | |
1018 * from/to/etc headers | |
1019 * additional search terms | |
1020 * flag based searching | |
1021 * anything else that the RFC supports, basically." | |
1022 ;; Walk through the query and turn it into an IMAP query string. | |
1023 (nnir-imap-query-to-imap criteria (nnir-imap-parse-query qstring))) | |
1024 | |
1025 | |
1026 (defun nnir-imap-query-to-imap (criteria query) | |
1027 "Turn a s-expression format query into IMAP." | |
1028 (mapconcat | |
1029 ;; Turn the expressions into IMAP text | |
1030 (lambda (item) | |
1031 (nnir-imap-expr-to-imap criteria item)) | |
1032 ;; The query, already in s-expr format. | |
1033 query | |
1034 ;; Append a space between each expression | |
1035 " ")) | |
1036 | |
1037 | |
1038 (defun nnir-imap-expr-to-imap (criteria expr) | |
1039 "Convert EXPR into an IMAP search expression on CRITERIA" | |
1040 ;; What sort of expression is this, eh? | |
1041 (cond | |
1042 ;; Simple string term | |
1043 ((stringp expr) | |
1044 (format "%s \"%s\"" criteria (imap-quote-specials expr))) | |
1045 ;; Trivial term: and | |
1046 ((eq expr 'and) nil) | |
1047 ;; Composite term: or expression | |
1048 ((eq (car-safe expr) 'or) | |
1049 (format "OR %s %s" | |
1050 (nnir-imap-expr-to-imap criteria (second expr)) | |
1051 (nnir-imap-expr-to-imap criteria (third expr)))) | |
1052 ;; Composite term: just the fax, mam | |
1053 ((eq (car-safe expr) 'not) | |
1054 (format "NOT (%s)" (nnir-imap-query-to-imap criteria (rest expr)))) | |
1055 ;; Composite term: just expand it all. | |
1056 ((and (not (null expr)) (listp expr)) | |
1057 (format "(%s)" (nnir-imap-query-to-imap criteria expr))) | |
1058 ;; Complex value, give up for now. | |
1059 (t (error "Unhandled input: %S" expr)))) | |
1060 | |
1061 | |
1062 (defun nnir-imap-parse-query (string) | |
1063 "Turn STRING into an s-expression based query based on the IMAP | |
1064 query language as defined in `nnir-imap-make-query'. | |
1065 | |
1066 This involves turning individual tokens into higher level terms | |
1067 that the search language can then understand and use." | |
1068 (with-temp-buffer | |
1069 ;; Set up the parsing environment. | |
1070 (insert string) | |
1071 (goto-char (point-min)) | |
1072 ;; Now, collect the output terms and return them. | |
1073 (let (out) | |
1074 (while (not (nnir-imap-end-of-input)) | |
1075 (push (nnir-imap-next-expr) out)) | |
1076 (reverse out)))) | |
1077 | |
1078 | |
1079 (defun nnir-imap-next-expr (&optional count) | |
1080 "Return the next expression from the current buffer." | |
1081 (let ((term (nnir-imap-next-term count)) | |
1082 (next (nnir-imap-peek-symbol))) | |
1083 ;; Are we looking at an 'or' expression? | |
1084 (cond | |
1085 ;; Handle 'expr or expr' | |
1086 ((eq next 'or) | |
1087 (list 'or term (nnir-imap-next-expr 2))) | |
1088 ;; Anything else | |
1089 (t term)))) | |
1090 | |
1091 | |
1092 (defun nnir-imap-next-term (&optional count) | |
1093 "Return the next TERM from the current buffer." | |
1094 (let ((term (nnir-imap-next-symbol count))) | |
1095 ;; What sort of term is this? | |
1096 (cond | |
1097 ;; and -- just ignore it | |
1098 ((eq term 'and) 'and) | |
1099 ;; negated term | |
1100 ((eq term 'not) (list 'not (nnir-imap-next-expr))) | |
1101 ;; generic term | |
1102 (t term)))) | |
1103 | |
1104 | |
1105 (defun nnir-imap-peek-symbol () | |
1106 "Return the next symbol from the current buffer, but don't consume it." | |
1107 (save-excursion | |
1108 (nnir-imap-next-symbol))) | |
1109 | |
1110 (defun nnir-imap-next-symbol (&optional count) | |
1111 "Return the next symbol from the current buffer, or nil if we are | |
1112 at the end of the buffer. If supplied COUNT skips some symbols before | |
1113 returning the one at the supplied position." | |
1114 (when (and (numberp count) (> count 1)) | |
1115 (nnir-imap-next-symbol (1- count))) | |
1116 (let ((case-fold-search t)) | |
1117 ;; end of input stream? | |
1118 (unless (nnir-imap-end-of-input) | |
1119 ;; No, return the next symbol from the stream. | |
1120 (cond | |
1121 ;; negated expression -- return it and advance one char. | |
1122 ((looking-at "-") (forward-char 1) 'not) | |
1123 ;; quoted string | |
1124 ((looking-at "\"") (nnir-imap-delimited-string "\"")) | |
1125 ;; list expression -- we parse the content and return this as a list. | |
1126 ((looking-at "(") | |
1127 (nnir-imap-parse-query (nnir-imap-delimited-string ")"))) | |
1128 ;; keyword input -- return a symbol version | |
1129 ((looking-at "\\band\\b") (forward-char 3) 'and) | |
1130 ((looking-at "\\bor\\b") (forward-char 2) 'or) | |
1131 ((looking-at "\\bnot\\b") (forward-char 3) 'not) | |
1132 ;; Simple, boring keyword | |
1133 (t (let ((start (point)) | |
1134 (end (if (search-forward-regexp "[[:blank:]]" nil t) | |
1135 (prog1 | |
1136 (match-beginning 0) | |
1137 ;; unskip if we hit a non-blank terminal character. | |
1138 (when (string-match "[^[:blank:]]" (match-string 0)) | |
1139 (backward-char 1))) | |
1140 (goto-char (point-max))))) | |
1141 (buffer-substring start end))))))) | |
1142 | |
1143 (defun nnir-imap-delimited-string (delimiter) | |
1144 "Return a delimited string from the current buffer." | |
1145 (let ((start (point)) end) | |
1146 (forward-char 1) ; skip the first delimiter. | |
1147 (while (not end) | |
1148 (unless (search-forward delimiter nil t) | |
1149 (error "Unmatched delimited input with %s in query" delimiter)) | |
1150 (let ((here (point))) | |
1151 (unless (equal (buffer-substring (- here 2) (- here 1)) "\\") | |
1152 (setq end (point))))) | |
1153 (buffer-substring (1+ start) (1- end)))) | |
1154 | |
1155 (defun nnir-imap-end-of-input () | |
1156 "Are we at the end of input?" | |
1157 (skip-chars-forward "[[:blank:]]") | |
1158 (looking-at "$")) | |
95616 | 1159 |
95610 | 1160 |
1161 ;; Swish++ interface. | |
1162 ;; -cc- Todo | |
1163 ;; Search by | |
1164 ;; - group | |
1165 ;; Sort by | |
1166 ;; - rank (default) | |
1167 ;; - article number | |
1168 ;; - file size | |
1169 ;; - group | |
1170 (defun nnir-run-swish++ (query server &optional group) | |
1171 "Run QUERY against swish++. | |
1172 Returns a vector of (group name, file name) pairs (also vectors, | |
1173 actually). | |
1174 | |
1175 Tested with swish++ 4.7 on GNU/Linux and with swish++ 5.0b2 on | |
1176 Windows NT 4.0." | |
1177 | |
1178 (when group | |
105167
e14dfafc8256
* gnus-art.el (gnus-article-encrypt-body):
Juanma Barranquero <lekktu@gmail.com>
parents:
104700
diff
changeset
|
1179 (error "The swish++ backend cannot search specific groups")) |
95610 | 1180 |
1181 (save-excursion | |
1182 (let ( (qstring (cdr (assq 'query query))) | |
1183 (groupspec (cdr (assq 'group query))) | |
1184 (prefix (nnir-read-server-parm 'nnir-swish++-remove-prefix server)) | |
1185 artlist | |
1186 ;; nnml-use-compressed-files might be any string, but probably this | |
1187 ;; is sufficient. Note that we can't only use the value of | |
1188 ;; nnml-use-compressed-files because old articles might have been | |
1189 ;; saved with a different value. | |
1190 (article-pattern (if (string= server "nnmaildir:") | |
1191 ":[0-9]+" | |
1192 "^[0-9]+\\(\\.[a-z0-9]+\\)?$")) | |
1193 score artno dirnam filenam) | |
1194 | |
1195 (when (equal "" qstring) | |
105167
e14dfafc8256
* gnus-art.el (gnus-article-encrypt-body):
Juanma Barranquero <lekktu@gmail.com>
parents:
104700
diff
changeset
|
1196 (error "swish++: You didn't enter anything")) |
95610 | 1197 |
1198 (set-buffer (get-buffer-create nnir-tmp-buffer)) | |
1199 (erase-buffer) | |
1200 | |
1201 (if groupspec | |
1202 (message "Doing swish++ query %s on %s..." qstring groupspec) | |
1203 (message "Doing swish++ query %s..." qstring)) | |
1204 | |
1205 (let* ((cp-list `( ,nnir-swish++-program | |
1206 nil ; input from /dev/null | |
1207 t ; output | |
1208 nil ; don't redisplay | |
1209 "--config-file" ,(nnir-read-server-parm 'nnir-swish++-configuration-file server) | |
1210 ,@(nnir-read-server-parm 'nnir-swish++-additional-switches server) | |
1211 ,qstring ; the query, in swish++ format | |
1212 )) | |
1213 (exitstatus | |
1214 (progn | |
1215 (message "%s args: %s" nnir-swish++-program | |
1216 (mapconcat 'identity (cddddr cp-list) " ")) ;; ??? | |
1217 (apply 'call-process cp-list)))) | |
1218 (unless (or (null exitstatus) | |
1219 (zerop exitstatus)) | |
1220 (nnheader-report 'nnir "Couldn't run swish++: %s" exitstatus) | |
1221 ;; swish++ failure reason is in this buffer, show it if | |
1222 ;; the user wants it. | |
1223 (when (> gnus-verbose 6) | |
1224 (display-buffer nnir-tmp-buffer)))) | |
1225 | |
1226 ;; The results are output in the format of: | |
1227 ;; V 4.7 Linux | |
1228 ;; rank relative-path-name file-size file-title | |
1229 ;; V 5.0b2: | |
1230 ;; rank relative-path-name file-size topic?? | |
1231 ;; where rank is an integer from 1 to 100. | |
1232 (goto-char (point-min)) | |
1233 (while (re-search-forward | |
1234 "\\(^[0-9]+\\) \\([^ ]+\\) [0-9]+ \\(.*\\)$" nil t) | |
1235 (setq score (match-string 1) | |
1236 filenam (match-string 2) | |
1237 artno (file-name-nondirectory filenam) | |
1238 dirnam (file-name-directory filenam)) | |
1239 | |
1240 ;; don't match directories | |
1241 (when (string-match article-pattern artno) | |
1242 (when (not (null dirnam)) | |
1243 | |
1244 ;; maybe limit results to matching groups. | |
1245 (when (or (not groupspec) | |
1246 (string-match groupspec dirnam)) | |
1247 (nnir-add-result dirnam artno score prefix server artlist))))) | |
1248 | |
1249 (message "Massaging swish++ output...done") | |
1250 | |
1251 ;; Sort by score | |
1252 (apply 'vector | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1253 (sort artlist |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1254 (function (lambda (x y) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1255 (> (nnir-artitem-rsv x) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1256 (nnir-artitem-rsv y))))))))) |
95610 | 1257 |
1258 ;; Swish-E interface. | |
1259 (defun nnir-run-swish-e (query server &optional group) | |
1260 "Run given query against swish-e. | |
1261 Returns a vector of (group name, file name) pairs (also vectors, | |
1262 actually). | |
1263 | |
1264 Tested with swish-e-2.0.1 on Windows NT 4.0." | |
1265 | |
1266 ;; swish-e crashes with empty parameter to "-w" on commandline... | |
1267 (when group | |
105167
e14dfafc8256
* gnus-art.el (gnus-article-encrypt-body):
Juanma Barranquero <lekktu@gmail.com>
parents:
104700
diff
changeset
|
1268 (error "The swish-e backend cannot search specific groups")) |
95610 | 1269 |
1270 (save-excursion | |
1271 (let ((qstring (cdr (assq 'query query))) | |
1272 (prefix | |
1273 (or (nnir-read-server-parm 'nnir-swish-e-remove-prefix server) | |
1274 (error "Missing parameter `nnir-swish-e-remove-prefix'"))) | |
1275 artlist score artno dirnam group ) | |
1276 | |
1277 (when (equal "" qstring) | |
105167
e14dfafc8256
* gnus-art.el (gnus-article-encrypt-body):
Juanma Barranquero <lekktu@gmail.com>
parents:
104700
diff
changeset
|
1278 (error "swish-e: You didn't enter anything")) |
95610 | 1279 |
1280 (set-buffer (get-buffer-create nnir-tmp-buffer)) | |
1281 (erase-buffer) | |
1282 | |
1283 (message "Doing swish-e query %s..." query) | |
1284 (let* ((index-files | |
1285 (or (nnir-read-server-parm | |
1286 'nnir-swish-e-index-files server) | |
1287 (error "Missing parameter `nnir-swish-e-index-files'"))) | |
1288 (additional-switches | |
1289 (nnir-read-server-parm | |
1290 'nnir-swish-e-additional-switches server)) | |
1291 (cp-list `(,nnir-swish-e-program | |
1292 nil ; input from /dev/null | |
1293 t ; output | |
1294 nil ; don't redisplay | |
1295 "-f" ,@index-files | |
1296 ,@additional-switches | |
1297 "-w" | |
1298 ,qstring ; the query, in swish-e format | |
1299 )) | |
1300 (exitstatus | |
1301 (progn | |
1302 (message "%s args: %s" nnir-swish-e-program | |
1303 (mapconcat 'identity (cddddr cp-list) " ")) | |
1304 (apply 'call-process cp-list)))) | |
1305 (unless (or (null exitstatus) | |
1306 (zerop exitstatus)) | |
1307 (nnheader-report 'nnir "Couldn't run swish-e: %s" exitstatus) | |
1308 ;; swish-e failure reason is in this buffer, show it if | |
1309 ;; the user wants it. | |
1310 (when (> gnus-verbose 6) | |
1311 (display-buffer nnir-tmp-buffer)))) | |
1312 | |
1313 ;; The results are output in the format of: | |
1314 ;; rank path-name file-title file-size | |
1315 (goto-char (point-min)) | |
1316 (while (re-search-forward | |
1317 "\\(^[0-9]+\\) \\([^ ]+\\) \"\\([^\"]+\\)\" [0-9]+$" nil t) | |
1318 (setq score (match-string 1) | |
1319 artno (match-string 3) | |
1320 dirnam (file-name-directory (match-string 2))) | |
1321 | |
1322 ;; don't match directories | |
1323 (when (string-match "^[0-9]+$" artno) | |
1324 (when (not (null dirnam)) | |
1325 | |
1326 ;; remove nnir-swish-e-remove-prefix from beginning of dirname | |
1327 (when (string-match (concat "^" prefix) dirnam) | |
1328 (setq dirnam (replace-match "" t t dirnam))) | |
1329 | |
1330 (setq dirnam (substring dirnam 0 -1)) | |
1331 ;; eliminate all ".", "/", "\" from beginning. Always matches. | |
1332 (string-match "^[./\\]*\\(.*\\)$" dirnam) | |
1333 ;; "/" -> "." | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1334 (setq group (gnus-replace-in-string (match-string 1 dirnam) "/" ".")) |
95610 | 1335 ;; Windows "\\" -> "." |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1336 (setq group (gnus-replace-in-string group "\\\\" ".")) |
95610 | 1337 |
1338 (push (vector (nnir-group-full-name group server) | |
1339 (string-to-number artno) | |
1340 (string-to-number score)) | |
1341 artlist)))) | |
1342 | |
1343 (message "Massaging swish-e output...done") | |
1344 | |
1345 ;; Sort by score | |
1346 (apply 'vector | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1347 (sort artlist |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1348 (function (lambda (x y) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1349 (> (nnir-artitem-rsv x) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1350 (nnir-artitem-rsv y))))))))) |
95610 | 1351 |
1352 ;; HyREX interface | |
1353 (defun nnir-run-hyrex (query server &optional group) | |
1354 (save-excursion | |
1355 (let ((artlist nil) | |
1356 (groupspec (cdr (assq 'group query))) | |
1357 (qstring (cdr (assq 'query query))) | |
1358 (prefix (nnir-read-server-parm 'nnir-hyrex-remove-prefix server)) | |
1359 score artno dirnam) | |
1360 (when (and group groupspec) | |
1361 (error (concat "It does not make sense to use a group spec" | |
1362 " with process-marked groups."))) | |
1363 (when group | |
1364 (setq groupspec (gnus-group-real-name group))) | |
1365 (when (and group (not (equal group (nnir-group-full-name groupspec server)))) | |
1366 (message "%s vs. %s" group (nnir-group-full-name groupspec server)) | |
1367 (error "Server with groupspec doesn't match group !")) | |
1368 (set-buffer (get-buffer-create nnir-tmp-buffer)) | |
1369 (erase-buffer) | |
1370 (if groupspec | |
1371 (message "Doing hyrex-search query %s on %s..." query groupspec) | |
1372 (message "Doing hyrex-search query %s..." query)) | |
1373 (let* ((cp-list | |
1374 `( ,nnir-hyrex-program | |
1375 nil ; input from /dev/null | |
1376 t ; output | |
1377 nil ; don't redisplay | |
1378 "-i",(nnir-read-server-parm 'nnir-hyrex-index-directory server) ; index directory | |
1379 ,@(nnir-read-server-parm 'nnir-hyrex-additional-switches server) | |
1380 ,qstring ; the query, in hyrex-search format | |
1381 )) | |
1382 (exitstatus | |
1383 (progn | |
1384 (message "%s args: %s" nnir-hyrex-program | |
1385 (mapconcat 'identity (cddddr cp-list) " ")) | |
1386 (apply 'call-process cp-list)))) | |
1387 (unless (or (null exitstatus) | |
1388 (zerop exitstatus)) | |
1389 (nnheader-report 'nnir "Couldn't run hyrex-search: %s" exitstatus) | |
1390 ;; nnir-search failure reason is in this buffer, show it if | |
1391 ;; the user wants it. | |
1392 (when (> gnus-verbose 6) | |
1393 (display-buffer nnir-tmp-buffer)))) ;; FIXME: Dont clear buffer ! | |
1394 (if groupspec | |
1395 (message "Doing hyrex-search query \"%s\" on %s...done" qstring groupspec) | |
1396 (message "Doing hyrex-search query \"%s\"...done" qstring)) | |
1397 (sit-for 0) | |
1398 ;; nnir-search returns: | |
1399 ;; for nnml/nnfolder: "filename mailid weigth" | |
1400 ;; for nnimap: "group mailid weigth" | |
1401 (goto-char (point-min)) | |
1402 (delete-non-matching-lines "^\\S + [0-9]+ [0-9]+$") | |
1403 ;; HyREX couldn't search directly in groups -- so filter out here. | |
1404 (when groupspec | |
1405 (keep-lines groupspec)) | |
1406 ;; extract data from result lines | |
1407 (goto-char (point-min)) | |
1408 (while (re-search-forward | |
1409 "\\(\\S +\\) \\([0-9]+\\) \\([0-9]+\\)" nil t) | |
1410 (setq dirnam (match-string 1) | |
1411 artno (match-string 2) | |
1412 score (match-string 3)) | |
1413 (when (string-match prefix dirnam) | |
1414 (setq dirnam (replace-match "" t t dirnam))) | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1415 (push (vector (nnir-group-full-name |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1416 (gnus-replace-in-string dirnam "/" ".") server) |
95610 | 1417 (string-to-number artno) |
1418 (string-to-number score)) | |
1419 artlist)) | |
1420 (message "Massaging hyrex-search output...done.") | |
1421 (apply 'vector | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1422 (sort artlist |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1423 (function (lambda (x y) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1424 (if (string-lessp (nnir-artitem-group x) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1425 (nnir-artitem-group y)) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1426 t |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1427 (< (nnir-artitem-number x) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1428 (nnir-artitem-number y))))))) |
95610 | 1429 ))) |
1430 | |
1431 ;; Namazu interface | |
1432 (defun nnir-run-namazu (query server &optional group) | |
1433 "Run given query against Namazu. Returns a vector of (group name, file name) | |
1434 pairs (also vectors, actually). | |
1435 | |
1436 Tested with Namazu 2.0.6 on a GNU/Linux system." | |
1437 (when group | |
1438 (error "The Namazu backend cannot search specific groups")) | |
1439 (save-excursion | |
1440 (let ((article-pattern (if (string= server "nnmaildir:") | |
1441 ":[0-9]+" | |
1442 "^[0-9]+$")) | |
1443 artlist | |
1444 (qstring (cdr (assq 'query query))) | |
1445 (prefix (nnir-read-server-parm 'nnir-namazu-remove-prefix server)) | |
1446 score group article | |
1447 (process-environment (copy-sequence process-environment))) | |
1448 (setenv "LC_MESSAGES" "C") | |
1449 (set-buffer (get-buffer-create nnir-tmp-buffer)) | |
1450 (erase-buffer) | |
1451 (let* ((cp-list | |
1452 `( ,nnir-namazu-program | |
1453 nil ; input from /dev/null | |
1454 t ; output | |
1455 nil ; don't redisplay | |
1456 "-q" ; don't be verbose | |
1457 "-a" ; show all matches | |
1458 "-s" ; use short format | |
1459 ,@(nnir-read-server-parm 'nnir-namazu-additional-switches server) | |
1460 ,qstring ; the query, in namazu format | |
1461 ,(nnir-read-server-parm 'nnir-namazu-index-directory server) ; index directory | |
1462 )) | |
1463 (exitstatus | |
1464 (progn | |
1465 (message "%s args: %s" nnir-namazu-program | |
1466 (mapconcat 'identity (cddddr cp-list) " ")) | |
1467 (apply 'call-process cp-list)))) | |
1468 (unless (or (null exitstatus) | |
1469 (zerop exitstatus)) | |
1470 (nnheader-report 'nnir "Couldn't run namazu: %s" exitstatus) | |
1471 ;; Namazu failure reason is in this buffer, show it if | |
1472 ;; the user wants it. | |
1473 (when (> gnus-verbose 6) | |
1474 (display-buffer nnir-tmp-buffer)))) | |
1475 | |
1476 ;; Namazu output looks something like this: | |
1477 ;; 2. Re: Gnus agent expire broken (score: 55) | |
1478 ;; /home/henrik/Mail/mail/sent/1310 (4,138 bytes) | |
1479 | |
1480 (goto-char (point-min)) | |
1481 (while (re-search-forward | |
1482 "^\\([0-9]+\\.\\).*\\((score: \\([0-9]+\\)\\))\n\\([^ ]+\\)" | |
1483 nil t) | |
1484 (setq score (match-string 3) | |
1485 group (file-name-directory (match-string 4)) | |
1486 article (file-name-nondirectory (match-string 4))) | |
1487 | |
1488 ;; make sure article and group is sane | |
1489 (when (and (string-match article-pattern article) | |
1490 (not (null group))) | |
1491 (nnir-add-result group article score prefix server artlist))) | |
1492 | |
1493 ;; sort artlist by score | |
1494 (apply 'vector | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1495 (sort artlist |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1496 (function (lambda (x y) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1497 (> (nnir-artitem-rsv x) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1498 (nnir-artitem-rsv y))))))))) |
95610 | 1499 |
1500 (defun nnir-run-find-grep (query server &optional group) | |
1501 "Run find and grep to obtain matching articles." | |
1502 (let* ((method (gnus-server-to-method server)) | |
1503 (sym (intern | |
1504 (concat (symbol-name (car method)) "-directory"))) | |
1505 (directory (cadr (assoc sym (cddr method)))) | |
1506 (regexp (cdr (assoc 'query query))) | |
1507 (grep-options (cdr (assoc 'grep-options query))) | |
1508 artlist) | |
1509 (unless directory | |
1510 (error "No directory found in method specification of server %s" | |
1511 server)) | |
1512 (message "Searching %s using find-grep..." (or group server)) | |
1513 (save-window-excursion | |
1514 (set-buffer (get-buffer-create nnir-tmp-buffer)) | |
1515 (erase-buffer) | |
1516 (if (> gnus-verbose 6) | |
1517 (pop-to-buffer (current-buffer))) | |
1518 (cd directory) ; Using relative paths simplifies postprocessing. | |
1519 (let ((group | |
1520 (if (not group) | |
1521 "." | |
1522 ;; Try accessing the group literally as well as | |
1523 ;; interpreting dots as directory separators so the | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1524 ;; engine works with plain nnml as well as the Gnus Cache. |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1525 (let ((group (gnus-group-real-name group))) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1526 ;; Replace cl-func find-if. |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1527 (if (file-directory-p group) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1528 group |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1529 (if (file-directory-p |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1530 (setq group (gnus-replace-in-string group "\\." "/" t))) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1531 group)))))) |
95610 | 1532 (unless group |
1533 (error "Cannot locate directory for group")) | |
1534 (save-excursion | |
1535 (apply | |
1536 'call-process "find" nil t | |
1537 "find" group "-type" "f" "-name" "[0-9]*" "-exec" | |
1538 "grep" | |
95830 | 1539 `("-l" ,@(and grep-options |
1540 ;; Note: the 3rd arg of `split-string' is not | |
1541 ;; available in Emacs 21. | |
1542 (delete "" (split-string grep-options "\\s-"))) | |
95610 | 1543 "-e" ,regexp "{}" "+")))) |
1544 | |
1545 ;; Translate relative paths to group names. | |
1546 (while (not (eobp)) | |
95830 | 1547 (let* ((path (delete |
1548 "" | |
1549 (split-string | |
1550 (buffer-substring (point) (line-end-position)) "/"))) | |
95610 | 1551 (art (string-to-number (car (last path))))) |
1552 (while (string= "." (car path)) | |
1553 (setq path (cdr path))) | |
104584
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1554 (let ((group (mapconcat 'identity |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1555 ;; Replace cl-func: (subseq path 0 -1) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1556 (let ((end (1- (length path))) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1557 res) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1558 (while (>= (setq end (1- end)) 0) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1559 (push (pop path) res)) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1560 (nreverse res)) |
f540e84a920d
(top-level): Don't require cl at run-time.
Glenn Morris <rgm@gnu.org>
parents:
100908
diff
changeset
|
1561 "."))) |
95610 | 1562 (push (vector (nnir-group-full-name group server) art 0) |
1563 artlist)) | |
1564 (forward-line 1))) | |
1565 (message "Searching %s using find-grep...done" (or group server)) | |
1566 artlist))) | |
1567 | |
1568 ;;; Util Code: | |
1569 | |
1570 (defun nnir-read-parms (query) | |
1571 "Reads additional search parameters according to `nnir-engines'." | |
1572 (let ((parmspec (caddr (assoc nnir-search-engine nnir-engines)))) | |
1573 (cons (cons 'query query) | |
1574 (mapcar 'nnir-read-parm parmspec)))) | |
1575 | |
1576 (defun nnir-read-parm (parmspec) | |
1577 "Reads a single search parameter. | |
1578 `parmspec' is a cons cell, the car is a symbol, the cdr is a prompt." | |
1579 (let ((sym (car parmspec)) | |
1580 (prompt (cdr parmspec))) | |
1581 (if (listp prompt) | |
1582 (let* ((result (apply 'completing-read prompt)) | |
1583 (mapping (or (assoc result nnir-imap-search-arguments) | |
1584 (assoc nil nnir-imap-search-arguments)))) | |
1585 (cons sym (format (cdr mapping) result))) | |
1586 (cons sym (read-string prompt))))) | |
1587 | |
1588 (defun nnir-run-query (query) | |
1589 "Invoke appropriate search engine function (see `nnir-engines'). | |
1590 If some groups were process-marked, run the query for each of the groups | |
1591 and concat the results." | |
1592 (let ((q (car (read-from-string query)))) | |
1593 (if gnus-group-marked | |
1594 (apply 'vconcat | |
1595 (mapcar (lambda (x) | |
1596 (let ((server (nnir-group-server x)) | |
1597 search-func) | |
1598 (setq search-func (cadr | |
1599 (assoc | |
1600 (nnir-read-server-parm 'nnir-search-engine server) nnir-engines))) | |
1601 (if search-func | |
1602 (funcall search-func q server x) | |
1603 nil))) | |
1604 gnus-group-marked) | |
1605 ) | |
1606 (apply 'vconcat | |
1607 (mapcar (lambda (x) | |
1608 (if (and (equal (cadr x) 'ok) (not (equal (cadar x) "-ephemeral"))) | |
1609 (let ((server (format "%s:%s" (caar x) (cadar x))) | |
1610 search-func) | |
1611 (setq search-func (cadr | |
1612 (assoc | |
1613 (nnir-read-server-parm 'nnir-search-engine server) nnir-engines))) | |
1614 (if search-func | |
1615 (funcall search-func q server nil) | |
1616 nil)) | |
1617 nil)) | |
1618 gnus-opened-servers) | |
1619 )) | |
1620 )) | |
1621 | |
1622 (defun nnir-read-server-parm (key server) | |
1623 "Returns the parameter value of for the given server, where server is of | |
1624 form 'backend:name'." | |
1625 (let ((method (gnus-server-to-method server))) | |
1626 (cond ((and method (assq key (cddr method))) | |
1627 (nth 1 (assq key (cddr method)))) | |
1628 ((and nnir-mail-backend | |
1629 (gnus-server-equal method nnir-mail-backend)) | |
1630 (symbol-value key)) | |
1631 (t nil)))) | |
1632 ;; (if method | |
1633 ;; (if (assq key (cddr method)) | |
1634 ;; (nth 1 (assq key (cddr method))) | |
1635 ;; (symbol-value key)) | |
1636 ;; (symbol-value key)) | |
1637 ;; )) | |
1638 | |
1639 (defun nnir-group-full-name (shortname server) | |
1640 "For the given group name, return a full Gnus group name. | |
1641 The Gnus backend/server information is added." | |
1642 (gnus-group-prefixed-name shortname (gnus-server-to-method server))) | |
1643 | |
1644 (defun nnir-possibly-change-server (server) | |
1645 (unless (and server (nnir-server-opened server)) | |
1646 (nnir-open-server server))) | |
1647 | |
1648 | |
1649 ;; Data type article list. | |
1650 | |
1651 (defun nnir-artlist-length (artlist) | |
1652 "Returns number of articles in artlist." | |
1653 (length artlist)) | |
1654 | |
1655 (defun nnir-artlist-article (artlist n) | |
1656 "Returns from ARTLIST the Nth artitem (counting starting at 1)." | |
1657 (elt artlist (1- n))) | |
1658 | |
1659 (defun nnir-artitem-group (artitem) | |
1660 "Returns the group from the ARTITEM." | |
1661 (elt artitem 0)) | |
1662 | |
1663 (defun nnir-artlist-artitem-group (artlist n) | |
1664 "Returns from ARTLIST the group of the Nth artitem (counting from 1)." | |
1665 (nnir-artitem-group (nnir-artlist-article artlist n))) | |
1666 | |
1667 (defun nnir-artitem-number (artitem) | |
1668 "Returns the number from the ARTITEM." | |
1669 (elt artitem 1)) | |
1670 | |
1671 (defun nnir-artlist-artitem-number (artlist n) | |
1672 "Returns from ARTLIST the number of the Nth artitem (counting from 1)." | |
1673 (nnir-artitem-number (nnir-artlist-article artlist n))) | |
1674 | |
1675 (defun nnir-artitem-rsv (artitem) | |
1676 "Returns the Retrieval Status Value (RSV, score) from the ARTITEM." | |
1677 (elt artitem 2)) | |
1678 | |
1679 (defun nnir-artlist-artitem-rsv (artlist n) | |
1680 "Returns from ARTLIST the Retrieval Status Value of the Nth artitem | |
1681 \(counting from 1)." | |
1682 (nnir-artitem-rsv (nnir-artlist-article artlist n))) | |
1683 | |
1684 ;; unused? | |
1685 (defun nnir-artlist-groups (artlist) | |
1686 "Returns a list of all groups in the given ARTLIST." | |
1687 (let ((res nil) | |
1688 (with-dups nil)) | |
1689 ;; from each artitem, extract group component | |
1690 (setq with-dups (mapcar 'nnir-artitem-group artlist)) | |
1691 ;; remove duplicates from above | |
1692 (mapc (function (lambda (x) (add-to-list 'res x))) | |
1693 with-dups) | |
1694 res)) | |
1695 | |
1696 | |
1697 ;; The end. | |
1698 (provide 'nnir) | |
1699 | |
95616 | 1700 ;;; nnir.el ends here |