lisp/gnus/nnir.el

   1 ;;; nnir.el --- search mail with various search engines -*- coding: iso-8859-1 -*-
   2
   3 ;; Copyright (C) 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006,
   4 ;;   2007, 2008, 2009, 2010 Free Software Foundation, Inc.
   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>
  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:
  32
  33 ;; TODO: Documentation in the Gnus manual
  34
  35 ;; Where in the existing gnus manual would this fit best?
  36
  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).
  42
  43 ;; When looking at the retrieval result (in the Summary buffer) you
  44 ;; can type `G T' (aka M-x gnus-summary-nnir-goto-thread RET) on an
  45 ;; article.  You will be teleported into the group this article came
  46 ;; from, showing the thread this article is part of.
  47
  48 ;; The Lisp setup may involve setting a few variables and setting up the
  49 ;; search engine. You can define the variables in the server definition
  50 ;; like this :
  51 ;;   (setq gnus-secondary-select-methods '(
  52 ;;       (nnimap "" (nnimap-address "localhost")
  53 ;;                  (nnir-search-engine namazu)
  54 ;;       )))
  55 ;; The main variable to set is `nnir-search-engine'.  Choose one of
  56 ;; the engines listed in `nnir-engines'.  (Actually `nnir-engines' is
  57 ;; an alist, type `C-h v nnir-engines RET' for more information; this
  58 ;; includes examples for setting `nnir-search-engine', too.)
  59
  60 ;; If you use one of the local indices (namazu, find-grep, swish) you
  61 ;; must also set up a search engine backend.
  62
  63 ;; 1. Namazu
  64 ;;
  65 ;; The Namazu backend requires you to have one directory containing all
  66 ;; index files, this is controlled by the `nnir-namazu-index-directory'
  67 ;; variable.  To function the `nnir-namazu-remove-prefix' variable must
  68 ;; also be correct, see the documentation for `nnir-namazu-remove-prefix'
  69 ;; above.
  70 ;;
  71 ;; It is particularly important not to pass any any switches to namazu
  72 ;; that will change the output format.  Good switches to use include
  73 ;; `--sort', `--ascending', `--early' and `--late'.  Refer to the Namazu
  74 ;; documentation for further information on valid switches.
  75 ;;
  76 ;; To index my mail with the `mknmz' program I use the following
  77 ;; configuration file:
  78 ;;
  79 ;; ,----
  80 ;; | package conf;  # Don't remove this line!
  81 ;; |
  82 ;; | # Paths which will not be indexed. Don't use `^' or `$' anchors.
  83 ;; | $EXCLUDE_PATH = "spam|sent";
  84 ;; |
  85 ;; | # Header fields which should be searchable. case-insensitive
  86 ;; | $REMAIN_HEADER = "from|date|message-id|subject";
  87 ;; |
  88 ;; | # Searchable fields. case-insensitive
  89 ;; | $SEARCH_FIELD = "from|date|message-id|subject";
  90 ;; |
  91 ;; | # The max length of a word.
  92 ;; | $WORD_LENG_MAX = 128;
  93 ;; |
  94 ;; | # The max length of a field.
  95 ;; | $MAX_FIELD_LENGTH = 256;
  96 ;; `----
  97 ;;
  98 ;; My mail is stored in the directories ~/Mail/mail/, ~/Mail/lists/ and
  99 ;; ~/Mail/archive/, so to index them I go to the directory set in
 100 ;; `nnir-namazu-index-directory' and issue the following command.
 101 ;;
 102 ;;      mknmz --mailnews ~/Mail/archive/ ~/Mail/mail/ ~/Mail/lists/
 103 ;;
 104 ;; For maximum searching efficiency I have a cron job set to run this
 105 ;; command every four hours.
 106
 107 ;; 2. find-grep
 108 ;;
 109 ;; The find-grep engine simply runs find(1) to locate eligible
 110 ;; articles and searches them with grep(1).  This, of course, is much
 111 ;; slower than using a proper search engine but OTOH doesn't require
 112 ;; maintenance of an index and is still faster than using any built-in
 113 ;; means for searching.  The method specification of the server to
 114 ;; search must include a directory for this engine to work (E.g.,
 115 ;; `nnml-directory').  The tools must be POSIX compliant.  GNU Find
 116 ;; prior to version 4.2.12 (4.2.26 on Linux due to incorrect ARG_MAX
 117 ;; handling) does not work.
 118 ;; ,----
 119 ;; |    ;; find-grep configuration for searching the Gnus Cache
 120 ;; |
 121 ;; |    (nnml "cache"
 122 ;; |          (nnml-get-new-mail nil)
 123 ;; |          (nnir-search-engine find-grep)
 124 ;; |          (nnml-directory "~/News/cache/")
 125 ;; |          (nnml-active-file "~/News/cache/active"))
 126 ;; `----
 127
 128 ;; Developer information:
 129
 130 ;; I have tried to make the code expandable.  Basically, it is divided
 131 ;; into two layers.  The upper layer is somewhat like the `nnvirtual'
 132 ;; backend: given a specification of what articles to show from
 133 ;; another backend, it creates a group containing exactly those
 134 ;; articles.  The lower layer issues a query to a search engine and
 135 ;; produces such a specification of what articles to show from the
 136 ;; other backend.
 137
 138 ;; The interface between the two layers consists of the single
 139 ;; function `nnir-run-query', which just selects the appropriate
 140 ;; function for the search engine one is using.  The input to
 141 ;; `nnir-run-query' is a string, representing the query as input by
 142 ;; the user.  The output of `nnir-run-query' is supposed to be a
 143 ;; vector, each element of which should in turn be a three-element
 144 ;; vector.  The first element should be full group name of the article,
 145 ;; the second element should be the article number, and the third
 146 ;; element should be the Retrieval Status Value (RSV) as returned from
 147 ;; the search engine.  An RSV is the score assigned to the document by
 148 ;; the search engine.  For Boolean search engines, the
 149 ;; RSV is always 1000 (or 1 or 100, or whatever you like).
 150
 151 ;; The sorting order of the articles in the summary buffer created by
 152 ;; nnir is based on the order of the articles in the above mentioned
 153 ;; vector, so that's where you can do the sorting you'd like.  Maybe
 154 ;; it would be nice to have a way of displaying the search result
 155 ;; sorted differently?
 156
 157 ;; So what do you need to do when you want to add another search
 158 ;; engine?  You write a function that executes the query.  Temporary
 159 ;; data from the search engine can be put in `nnir-tmp-buffer'.  This
 160 ;; function should return the list of articles as a vector, as
 161 ;; described above.  Then, you need to register this backend in
 162 ;; `nnir-engines'.  Then, users can choose the backend by setting
 163 ;; `nnir-search-engine' as a server variable.
 164
 165 ;;; Setup Code:
 166
 167 ;; For Emacs <22.2 and XEmacs.
 168 (eval-and-compile
 169   (unless (fboundp 'declare-function) (defmacro declare-function (&rest r))))
 170
 171 (require 'nnoo)
 172 (require 'gnus-group)
 173 (require 'gnus-sum)
 174 (require 'message)
 175 (require 'gnus-util)
 176 (eval-when-compile
 177   (require 'cl))
 178
 179
 180 (eval-when-compile
 181   (autoload 'nnimap-buffer "nnimap")
 182   (autoload 'nnimap-command "nnimap")
 183   (autoload 'nnimap-possibly-change-group "nnimap"))
 184
 185 (nnoo-declare nnir)
 186 (nnoo-define-basics nnir)
 187
 188 (gnus-declare-backend "nnir" 'mail)
 189
 190
 191 ;;; User Customizable Variables:
 192
 193 (defgroup nnir nil
 194   "Search groups in Gnus with assorted seach engines."
 195   :group 'gnus)
 196
 197 (defcustom nnir-method-default-engines
 198   '((nnimap . imap)
 199     (nntp . gmane))
 200   "*Alist of default search engines keyed by server method"
 201   :type '(alist)
 202   :group 'nnir)
 203
 204 (defcustom nnir-imap-default-search-key "Whole message"
 205   "*The default IMAP search key for an nnir search. Must be one of
 206   the keys in `nnir-imap-search-arguments'. To use raw imap queries
 207   by default set this to \"Imap\""
 208   :type '(string)
 209   :group 'nnir)
 210
 211 (defcustom nnir-wais-program "waissearch"
 212   "*Name of waissearch executable."
 213   :type '(string)
 214   :group 'nnir)
 215
 216 (defcustom nnir-wais-database (expand-file-name "~/.wais/mail")
 217   "*Name of Wais database containing the mail.
 218
 219 Note that this should be a file name without extension.  For example,
 220 if you have a file /home/john/.wais/mail.fmt, use this:
 221     (setq nnir-wais-database \"/home/john/.wais/mail\")
 222 The string given here is passed to `waissearch -d' as-is."
 223   :type '(file)
 224   :group 'nnir)
 225
 226 (defcustom nnir-wais-remove-prefix (concat (getenv "HOME") "/Mail/")
 227   "*The prefix to remove from each directory name returned by waissearch
 228 in order to get a group name (albeit with / instead of .).  This is a
 229 regular expression.
 230
 231 For example, suppose that Wais returns file names such as
 232 \"/home/john/Mail/mail/misc/42\".  For this example, use the following
 233 setting:  (setq nnir-wais-remove-prefix \"/home/john/Mail/\")
 234 Note the trailing slash.  Removing this prefix gives \"mail/misc/42\".
 235 `nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to
 236 arrive at the correct group name, \"mail.misc\"."
 237   :type '(regexp)
 238   :group 'nnir)
 239
 240 (defcustom nnir-swish++-configuration-file
 241   (expand-file-name "~/Mail/swish++.conf")
 242   "*Configuration file for swish++."
 243   :type '(file)
 244   :group 'nnir)
 245
 246 (defcustom nnir-swish++-program "search"
 247   "*Name of swish++ search executable."
 248   :type '(string)
 249   :group 'nnir)
 250
 251 (defcustom nnir-swish++-additional-switches '()
 252   "*A list of strings, to be given as additional arguments to swish++.
 253
 254 Note that this should be a list.  Ie, do NOT use the following:
 255     (setq nnir-swish++-additional-switches \"-i -w\") ; wrong
 256 Instead, use this:
 257     (setq nnir-swish++-additional-switches '(\"-i\" \"-w\"))"
 258   :type '(repeat (string))
 259   :group 'nnir)
 260
 261 (defcustom nnir-swish++-remove-prefix (concat (getenv "HOME") "/Mail/")
 262   "*The prefix to remove from each file name returned by swish++
 263 in order to get a group name (albeit with / instead of .).  This is a
 264 regular expression.
 265
 266 This variable is very similar to `nnir-namazu-remove-prefix', except
 267 that it is for swish++, not Namazu."
 268   :type '(regexp)
 269   :group 'nnir)
 270
 271 ;; Swish-E.
 272 ;; URL: http://swish-e.org/
 273 ;; Variables `nnir-swish-e-index-file', `nnir-swish-e-program' and
 274 ;; `nnir-swish-e-additional-switches'
 275
 276 (make-obsolete-variable 'nnir-swish-e-index-file
 277                         'nnir-swish-e-index-files "Emacs 23.1")
 278 (defcustom nnir-swish-e-index-file
 279   (expand-file-name "~/Mail/index.swish-e")
 280   "*Index file for swish-e.
 281 This could be a server parameter.
 282 It is never consulted once `nnir-swish-e-index-files', which should be
 283 used instead, has been customized."
 284   :type '(file)
 285   :group 'nnir)
 286
 287 (defcustom nnir-swish-e-index-files
 288   (list nnir-swish-e-index-file)
 289   "*List of index files for swish-e.
 290 This could be a server parameter."
 291   :type '(repeat (file))
 292   :group 'nnir)
 293
 294 (defcustom nnir-swish-e-program "swish-e"
 295   "*Name of swish-e search executable.
 296 This cannot be a server parameter."
 297   :type '(string)
 298   :group 'nnir)
 299
 300 (defcustom nnir-swish-e-additional-switches '()
 301   "*A list of strings, to be given as additional arguments to swish-e.
 302
 303 Note that this should be a list.  Ie, do NOT use the following:
 304     (setq nnir-swish-e-additional-switches \"-i -w\") ; wrong
 305 Instead, use this:
 306     (setq nnir-swish-e-additional-switches '(\"-i\" \"-w\"))
 307
 308 This could be a server parameter."
 309   :type '(repeat (string))
 310   :group 'nnir)
 311
 312 (defcustom nnir-swish-e-remove-prefix (concat (getenv "HOME") "/Mail/")
 313   "*The prefix to remove from each file name returned by swish-e
 314 in order to get a group name (albeit with / instead of .).  This is a
 315 regular expression.
 316
 317 This variable is very similar to `nnir-namazu-remove-prefix', except
 318 that it is for swish-e, not Namazu.
 319
 320 This could be a server parameter."
 321   :type '(regexp)
 322   :group 'nnir)
 323
 324 ;; HyREX engine, see <URL:http://ls6-www.cs.uni-dortmund.de/>
 325
 326 (defcustom nnir-hyrex-program "nnir-search"
 327   "*Name of the nnir-search executable."
 328   :type '(string)
 329   :group 'nnir)
 330
 331 (defcustom nnir-hyrex-additional-switches '()
 332   "*A list of strings, to be given as additional arguments for nnir-search.
 333 Note that this should be a list. Ie, do NOT use the following:
 334     (setq nnir-hyrex-additional-switches \"-ddl ddl.xml -c nnir\") ; wrong !
 335 Instead, use this:
 336     (setq nnir-hyrex-additional-switches '(\"-ddl\" \"ddl.xml\" \"-c\" \"nnir\"))"
 337   :type '(repeat (string))
 338   :group 'nnir)
 339
 340 (defcustom nnir-hyrex-index-directory (getenv "HOME")
 341   "*Index directory for HyREX."
 342   :type '(directory)
 343   :group 'nnir)
 344
 345 (defcustom nnir-hyrex-remove-prefix (concat (getenv "HOME") "/Mail/")
 346   "*The prefix to remove from each file name returned by HyREX
 347 in order to get a group name (albeit with / instead of .).
 348
 349 For example, suppose that HyREX returns file names such as
 350 \"/home/john/Mail/mail/misc/42\".  For this example, use the following
 351 setting:  (setq nnir-hyrex-remove-prefix \"/home/john/Mail/\")
 352 Note the trailing slash.  Removing this prefix gives \"mail/misc/42\".
 353 `nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to
 354 arrive at the correct group name, \"mail.misc\"."
 355   :type '(directory)
 356   :group 'nnir)
 357
 358 ;; Namazu engine, see <URL:http://www.namazu.org/>
 359
 360 (defcustom nnir-namazu-program "namazu"
 361   "*Name of Namazu search executable."
 362   :type '(string)
 363   :group 'nnir)
 364
 365 (defcustom nnir-namazu-index-directory (expand-file-name "~/Mail/namazu/")
 366   "*Index directory for Namazu."
 367   :type '(directory)
 368   :group 'nnir)
 369
 370 (defcustom nnir-namazu-additional-switches '()
 371   "*A list of strings, to be given as additional arguments to namazu.
 372 The switches `-q', `-a', and `-s' are always used, very few other switches
 373 make any sense in this context.
 374
 375 Note that this should be a list.  Ie, do NOT use the following:
 376     (setq nnir-namazu-additional-switches \"-i -w\") ; wrong
 377 Instead, use this:
 378     (setq nnir-namazu-additional-switches '(\"-i\" \"-w\"))"
 379   :type '(repeat (string))
 380   :group 'nnir)
 381
 382 (defcustom nnir-namazu-remove-prefix (concat (getenv "HOME") "/Mail/")
 383   "*The prefix to remove from each file name returned by Namazu
 384 in order to get a group name (albeit with / instead of .).
 385
 386 For example, suppose that Namazu returns file names such as
 387 \"/home/john/Mail/mail/misc/42\".  For this example, use the following
 388 setting:  (setq nnir-namazu-remove-prefix \"/home/john/Mail/\")
 389 Note the trailing slash.  Removing this prefix gives \"mail/misc/42\".
 390 `nnir' knows to remove the \"/42\" and to replace \"/\" with \".\" to
 391 arrive at the correct group name, \"mail.misc\"."
 392   :type '(directory)
 393   :group 'nnir)
 394
 395 ;; Imap variables
 396
 397 (defvar nnir-imap-search-arguments
 398   '(("Whole message" . "TEXT")
 399     ("Subject" . "SUBJECT")
 400     ("To" . "TO")
 401     ("From" . "FROM")
 402     ("Imap" . ""))
 403   "Mapping from user readable keys to IMAP search items for use in nnir")
 404
 405 (defvar nnir-imap-search-other "HEADER %S"
 406   "The IMAP search item to use for anything other than
 407   `nnir-imap-search-arguments'. By default this is the name of an
 408   email header field")
 409
 410 (defvar nnir-imap-search-argument-history ()
 411   "The history for querying search options in nnir")
 412
 413 ;;; Developer Extension Variable:
 414
 415 (defvar nnir-engines
 416   `((wais    nnir-run-waissearch
 417              ())
 418     (imap    nnir-run-imap
 419              ((criteria
 420                "Imap Search in"                   ; Prompt
 421                ,(mapcar 'car nnir-imap-search-arguments) ; alist for completing
 422                nil                                ; allow any user input
 423                nil                                ; initial value
 424                nnir-imap-search-argument-history  ; the history to use
 425                ,nnir-imap-default-search-key      ; default
 426                )))
 427     (gmane   nnir-run-gmane
 428              ((author . "Gmane Author: ")))
 429     (swish++ nnir-run-swish++
 430              ((group . "Swish++ Group spec: ")))
 431     (swish-e nnir-run-swish-e
 432              ((group . "Swish-e Group spec: ")))
 433     (namazu  nnir-run-namazu
 434              ())
 435     (hyrex   nnir-run-hyrex
 436              ((group . "Hyrex Group spec: ")))
 437     (find-grep nnir-run-find-grep
 438                ((grep-options . "Grep options: "))))
 439   "Alist of supported search engines.
 440 Each element in the alist is a three-element list (ENGINE FUNCTION ARGS).
 441 ENGINE is a symbol designating the searching engine.  FUNCTION is also
 442 a symbol, giving the function that does the search.  The third element
 443 ARGS is a list of cons pairs (PARAM . PROMPT).  When issuing a query,
 444 the FUNCTION will issue a query for each of the PARAMs, using PROMPT.
 445
 446 The value of `nnir-search-engine' must be one of the ENGINE symbols.
 447 For example, for searching a server using namazu include
 448     (nnir-search-engine namazu)
 449 in the server definition.  Note that you have to set additional
 450 variables for most backends.  For example, the `namazu' backend
 451 needs the variables `nnir-namazu-program',
 452 `nnir-namazu-index-directory' and `nnir-namazu-remove-prefix'.
 453
 454 Add an entry here when adding a new search engine.")
 455
 456 (defvar nnir-get-article-nov-override-function nil
 457   "If non-nil, a function that will be passed each search result.  This
 458 should return a message's headers in NOV format.
 459
 460 If this variable is nil, or if the provided function returns nil for a search
 461 result, `gnus-retrieve-headers' will be called instead.")
 462
 463 ;;; Internal Variables:
 464
 465 (defvar nnir-current-query nil
 466   "Internal: stores current query (= group name).")
 467
 468 (defvar nnir-current-server nil
 469   "Internal: stores current server (does it ever change?).")
 470
 471 (defvar nnir-current-group-marked nil
 472   "Internal: stores current list of process-marked groups.")
 473
 474 (defvar nnir-artlist nil
 475   "Internal: stores search result.")
 476
 477 (defvar nnir-tmp-buffer " *nnir*"
 478   "Internal: temporary buffer.")
 479
 480 (defvar nnir-search-history ()
 481   "Internal: the history for querying search options in nnir")
 482
 483 (defvar nnir-extra-parms nil
 484   "Internal: stores request for extra search parms")
 485
 486 ;;; Code:
 487
 488 ;; Gnus glue.
 489
 490 (defun gnus-group-make-nnir-group (nnir-extra-parms)
 491   "Create an nnir group.  Asks for query."
 492   (interactive "P")
 493   (setq nnir-current-query nil
 494         nnir-current-server nil
 495         nnir-current-group-marked nil
 496         nnir-artlist nil)
 497   (let* ((query (read-string "Query: " nil 'nnir-search-history))
 498          (parms (list (cons 'query query)))
 499          (srv (if (gnus-server-server-name)
 500                   "all" "")))
 501     (add-to-list 'parms (cons 'unique-id (message-unique-id)) t)
 502     (gnus-group-read-ephemeral-group
 503      (concat "nnir:" (prin1-to-string parms)) (list 'nnir srv) t
 504      (cons (current-buffer) gnus-current-window-configuration)
 505      nil)))
 506
 507 ;; Summary mode commands.
 508
 509 (defun gnus-summary-nnir-goto-thread ()
 510   "Only applies to nnir groups.  Go to group this article came from
 511 and show thread that contains this article."
 512   (interactive)
 513   (unless (eq 'nnir (car (gnus-find-method-for-group gnus-newsgroup-name)))
 514     (error "Can't execute this command unless in nnir group"))
 515   (let* ((cur (gnus-summary-article-number))
 516          (group (nnir-artlist-artitem-group nnir-artlist cur))
 517          (backend-number (nnir-artlist-artitem-number nnir-artlist cur))
 518          (id (mail-header-id (gnus-summary-article-header)))
 519          (refs (split-string
 520                 (mail-header-references (gnus-summary-article-header)))))
 521     (if (eq (car (gnus-find-method-for-group group)) 'nnimap)
 522         (progn
 523           (nnimap-possibly-change-group (gnus-group-short-name group) nil)
 524           (with-current-buffer (nnimap-buffer)
 525             (let* ((cmd
 526                     (let ((value
 527                            (format
 528                             "(OR HEADER REFERENCES %s HEADER Message-Id %s)"
 529                             id id)))
 530                       (dolist (refid refs value)
 531                         (setq value
 532                               (format
 533                                "(OR (OR HEADER Message-Id %s HEADER REFERENCES %s) %s)"
 534                                refid refid value)))))
 535                    (result (nnimap-command "UID SEARCH %s" cmd)))
 536               (gnus-summary-read-group-1
 537                group t t gnus-summary-buffer nil
 538                (and (car result)
 539                     (delete 0 (mapcar
 540                                #'string-to-number
 541                                (cdr (assoc "SEARCH" (cdr result))))))))))
 542       (gnus-summary-read-group-1 group t t gnus-summary-buffer
 543                                  nil (list backend-number))
 544       (gnus-summary-limit (list backend-number))
 545       (gnus-summary-refer-thread))))
 546
 547
 548 (if (fboundp 'eval-after-load)
 549     (eval-after-load "gnus-sum"
 550       '(define-key gnus-summary-goto-map
 551          "T" 'gnus-summary-nnir-goto-thread))
 552   (add-hook 'gnus-summary-mode-hook
 553             (function (lambda ()
 554                         (define-key gnus-summary-goto-map
 555                           "T" 'gnus-summary-nnir-goto-thread)))))
 556
 557
 558
 559 ;; Gnus backend interface functions.
 560
 561 (deffoo nnir-open-server (server &optional definitions)
 562   ;; Just set the server variables appropriately.
 563   (nnoo-change-server 'nnir server definitions))
 564
 565 (deffoo nnir-request-group (group &optional server fast info)
 566   "GROUP is the query string."
 567   (nnir-possibly-change-server server)
 568   ;; Check for cache and return that if appropriate.
 569   (if (and (equal group nnir-current-query)
 570            (equal gnus-group-marked nnir-current-group-marked)
 571            (or (null server)
 572                (equal server nnir-current-server)))
 573       nnir-artlist
 574     ;; Cache miss.
 575     (setq nnir-artlist (nnir-run-query group server)))
 576   (with-current-buffer nntp-server-buffer
 577     (setq nnir-current-query group)
 578     (when server (setq nnir-current-server server))
 579     (setq nnir-current-group-marked gnus-group-marked)
 580     (if (zerop (length nnir-artlist))
 581         (nnheader-report 'nnir "Search produced empty results.")
 582       ;; Remember data for cache.
 583       (nnheader-insert "211 %d %d %d %s\n"
 584                        (nnir-artlist-length nnir-artlist) ; total #
 585                        1              ; first #
 586                        (nnir-artlist-length nnir-artlist) ; last #
 587                        group))))      ; group name
 588
 589 (deffoo nnir-retrieve-headers (articles &optional group server fetch-old)
 590   (save-excursion
 591     (let ((artlist (copy-sequence articles))
 592           art artitem artgroup artno artrsv artfullgroup
 593           novitem novdata foo server)
 594       (while (not (null artlist))
 595         (setq art (car artlist))
 596         (or (numberp art)
 597             (nnheader-report
 598              'nnir
 599              "nnir-retrieve-headers doesn't grok message ids: %s"
 600              art))
 601         (setq artitem (nnir-artlist-article nnir-artlist art))
 602         (setq artrsv (nnir-artitem-rsv artitem))
 603         (setq artfullgroup (nnir-artitem-group artitem))
 604         (setq artno (nnir-artitem-number artitem))
 605         (setq artgroup (gnus-group-real-name artfullgroup))
 606         (setq server (gnus-group-server artfullgroup))
 607         ;; retrieve NOV or HEAD data for this article, transform into
 608         ;; NOV data and prepend to `novdata'
 609         (set-buffer nntp-server-buffer)
 610         (nnir-possibly-change-server server)
 611         (let ((gnus-override-method
 612                (gnus-server-to-method server)))
 613           ;; if nnir-get-article-nov-override-function is set, use it
 614           (if nnir-get-article-nov-override-function
 615               (setq novitem (funcall nnir-get-article-nov-override-function
 616                                      artitem))
 617             ;; else, set novitem through nnheader-parse-nov/nnheader-parse-head
 618             (case (setq foo (gnus-retrieve-headers (list artno)
 619                                                    artfullgroup nil))
 620               (nov
 621                (goto-char (point-min))
 622                (setq novitem (nnheader-parse-nov)))
 623               (headers
 624                (goto-char (point-min))
 625                (setq novitem (nnheader-parse-head)))
 626               (t (error "Unknown header type %s while requesting article %s of group %s"
 627                         foo artno artfullgroup)))))
 628         ;; replace article number in original group with article number
 629         ;; in nnir group
 630         (when novitem
 631           (mail-header-set-number novitem art)
 632           (mail-header-set-from novitem
 633                                 (mail-header-from novitem))
 634           (mail-header-set-subject
 635            novitem
 636            (format "[%d: %s/%d] %s"
 637                    artrsv artgroup artno
 638                    (mail-header-subject novitem)))
 639           (push novitem novdata)
 640           (setq artlist (cdr artlist))))
 641       (setq novdata (nreverse novdata))
 642       (set-buffer nntp-server-buffer) (erase-buffer)
 643       (mapc 'nnheader-insert-nov novdata)
 644       'nov)))
 645
 646 (deffoo nnir-request-article (article
 647                               &optional group server to-buffer)
 648   (if (stringp article)
 649       (nnheader-report
 650        'nnir
 651        "nnir-retrieve-headers doesn't grok message ids: %s"
 652        article)
 653     (save-excursion
 654       (let* ((artitem (nnir-artlist-article nnir-artlist
 655                                             article))
 656              (artfullgroup (nnir-artitem-group artitem))
 657              (artno (nnir-artitem-number artitem))
 658              ;; Bug?
 659              ;; Why must we bind nntp-server-buffer here?  It won't
 660              ;; work if `buf' is used, say.  (Of course, the set-buffer
 661              ;; line below must then be updated, too.)
 662              (nntp-server-buffer (or to-buffer nntp-server-buffer)))
 663         (set-buffer nntp-server-buffer)
 664         (erase-buffer)
 665         (message "Requesting article %d from group %s"
 666                  artno artfullgroup)
 667         (gnus-request-article artno artfullgroup nntp-server-buffer)
 668         (cons artfullgroup artno)))))
 669
 670
 671 (nnoo-define-skeleton nnir)
 672
 673
 674 (defmacro nnir-add-result (dirnam artno score prefix server artlist)
 675   "Ask `nnir-compose-result' to construct a result vector,
 676 and if it is non-nil, add it to artlist."
 677   `(let ((result (nnir-compose-result ,dirnam ,artno ,score ,prefix ,server)))
 678      (when (not (null result))
 679        (push result ,artlist))))
 680
 681 (autoload 'nnmaildir-base-name-to-article-number "nnmaildir")
 682
 683 ;; Helper function currently used by the Swish++ and Namazu backends;
 684 ;; perhaps useful for other backends as well
 685 (defun nnir-compose-result (dirnam article score prefix server)
 686   "Extract the group from dirnam, and create a result vector
 687 ready to be added to the list of search results."
 688
 689   ;; remove nnir-*-remove-prefix from beginning of dirnam filename
 690   (when (string-match (concat "^" prefix) dirnam)
 691     (setq dirnam (replace-match "" t t dirnam)))
 692
 693   (when (file-readable-p (concat prefix dirnam article))
 694     ;; remove trailing slash and, for nnmaildir, cur/new/tmp
 695     (setq dirnam
 696           (substring dirnam 0
 697                      (if (string= (gnus-group-server server) "nnmaildir")
 698                          -5 -1)))
 699
 700     ;; Set group to dirnam without any leading dots or slashes,
 701     ;; and with all subsequent slashes replaced by dots
 702     (let ((group (gnus-replace-in-string
 703                  (gnus-replace-in-string dirnam "^[./\\]" "" t)
 704                  "[/\\]" "." t)))
 705
 706     (vector (nnir-group-full-name group server)
 707             (if (string= (gnus-group-server server) "nnmaildir")
 708                 (nnmaildir-base-name-to-article-number
 709                  (substring article 0 (string-match ":" article))
 710                  group nil)
 711               (string-to-number article))
 712             (string-to-number score)))))
 713
 714 ;;; Search Engine Interfaces:
 715
 716 ;; freeWAIS-sf interface.
 717 (defun nnir-run-waissearch (query server &optional group)
 718   "Run given query agains waissearch.  Returns vector of (group name, file name)
 719 pairs (also vectors, actually)."
 720   ;; (when group
 721   ;;   (error "The freeWAIS-sf backend cannot search specific groups"))
 722   (save-excursion
 723     (let ((qstring (cdr (assq 'query query)))
 724           (prefix (nnir-read-server-parm 'nnir-wais-remove-prefix server))
 725           artlist score artno dirnam)
 726       (set-buffer (get-buffer-create nnir-tmp-buffer))
 727       (erase-buffer)
 728       (message "Doing WAIS query %s..." query)
 729       (call-process nnir-wais-program
 730                     nil                 ; input from /dev/null
 731                     t                   ; output to current buffer
 732                     nil                 ; don't redisplay
 733                     "-d" (nnir-read-server-parm 'nnir-wais-database server) ; database to search
 734                     qstring)
 735       (message "Massaging waissearch output...")
 736       ;; remove superfluous lines
 737       (keep-lines "Score:")
 738       ;; extract data from result lines
 739       (goto-char (point-min))
 740       (while (re-search-forward
 741               "Score: +\\([0-9]+\\).*'\\([0-9]+\\) +\\([^']+\\)/'" nil t)
 742         (setq score (match-string 1)
 743               artno (match-string 2)
 744               dirnam (match-string 3))
 745         (unless (string-match prefix dirnam)
 746           (nnheader-report 'nnir "Dir name %s doesn't contain prefix %s"
 747                            dirnam prefix))
 748         (setq group (gnus-replace-in-string
 749                      (replace-match "" t t dirnam) "/" "."))
 750         (push (vector (nnir-group-full-name group server)
 751                       (string-to-number artno)
 752                       (string-to-number score))
 753               artlist))
 754       (message "Massaging waissearch output...done")
 755       (apply 'vector
 756              (sort artlist
 757                    (function (lambda (x y)
 758                                (> (nnir-artitem-rsv x)
 759                                   (nnir-artitem-rsv y)))))))))
 760
 761 ;; imap interface
 762 (defun nnir-run-imap (query srv &optional groups)
 763   "Run a search against an IMAP back-end server.
 764 This uses a custom query language parser; see `nnir-imap-make-query' for
 765 details on the language and supported extensions"
 766   (save-excursion
 767     (let ((qstring (cdr (assq 'query query)))
 768           (server (cadr (gnus-server-to-method srv)))
 769           (defs (caddr (gnus-server-to-method srv)))
 770           (criteria (or (cdr (assq 'criteria query))
 771                         (cdr (assoc nnir-imap-default-search-key
 772                                     nnir-imap-search-arguments))))
 773           (gnus-inhibit-demon t)
 774           (groups (or groups (nnir-get-active srv)))
 775           artlist)
 776       (message "Opening server %s" server)
 777       (apply
 778        'vconcat
 779        (mapcar
 780         (lambda (x)
 781           (let ((group x))
 782             (condition-case ()
 783                 (when (nnimap-possibly-change-group
 784                        (gnus-group-short-name group) server)
 785                   (with-current-buffer (nnimap-buffer)
 786                     (message "Searching %s..." group)
 787                     (let ((arts 0)
 788                           (result (nnimap-command "UID SEARCH %s"
 789                                                   (if (string= criteria "")
 790                                                       qstring
 791                                                     (nnir-imap-make-query
 792                                                      criteria qstring)))))
 793                       (mapc
 794                        (lambda (artnum) (push (vector group artnum 1) artlist)
 795                          (setq arts (1+ arts)))
 796                        (and (car result)
 797                             (delete 0 (mapcar #'string-to-number
 798                                               (cdr (assoc "SEARCH"
 799                                                           (cdr result)))))))
 800                       (message "Searching %s... %d matches" group arts)))
 801                   (message "Searching %s...done" group))
 802               (quit nil))
 803             (reverse artlist)))
 804         groups)))))
 805
 806 (defun nnir-imap-make-query (criteria qstring)
 807   "Parse the query string and criteria into an appropriate IMAP search
 808 expression, returning the string query to make.
 809
 810 This implements a little language designed to return the expected results
 811 to an arbitrary query string to the end user.
 812
 813 The search is always case-insensitive, as defined by RFC2060, and supports
 814 the following features (inspired by the Google search input language):
 815
 816 Automatic \"and\" queries
 817     If you specify multiple words then they will be treated as an \"and\"
 818     expression intended to match all components.
 819
 820 Phrase searches
 821     If you wrap your query in double-quotes then it will be treated as a
 822     literal string.
 823
 824 Negative terms
 825     If you precede a term with \"-\" then it will negate that.
 826
 827 \"OR\" queries
 828     If you include an upper-case \"OR\" in your search it will cause the
 829     term before it and the term after it to be treated as alternatives.
 830
 831 In future the following will be added to the language:
 832  * support for date matches
 833  * support for location of text matching within the query
 834  * from/to/etc headers
 835  * additional search terms
 836  * flag based searching
 837  * anything else that the RFC supports, basically."
 838   ;; Walk through the query and turn it into an IMAP query string.
 839   (nnir-imap-query-to-imap criteria (nnir-imap-parse-query qstring)))
 840
 841
 842 (defun nnir-imap-query-to-imap (criteria query)
 843   "Turn a s-expression format query into IMAP."
 844   (mapconcat
 845    ;; Turn the expressions into IMAP text
 846    (lambda (item)
 847      (nnir-imap-expr-to-imap criteria item))
 848    ;; The query, already in s-expr format.
 849    query
 850    ;; Append a space between each expression
 851    " "))
 852
 853
 854 (defun nnir-imap-expr-to-imap (criteria expr)
 855   "Convert EXPR into an IMAP search expression on CRITERIA"
 856   ;; What sort of expression is this, eh?
 857   (cond
 858    ;; Simple string term
 859    ((stringp expr)
 860     (format "%s %S" criteria expr))
 861    ;; Trivial term: and
 862    ((eq expr 'and) nil)
 863    ;; Composite term: or expression
 864    ((eq (car-safe expr) 'or)
 865     (format "OR %s %s"
 866             (nnir-imap-expr-to-imap criteria (second expr))
 867             (nnir-imap-expr-to-imap criteria (third expr))))
 868    ;; Composite term: just the fax, mam
 869    ((eq (car-safe expr) 'not)
 870     (format "NOT (%s)" (nnir-imap-query-to-imap criteria (rest expr))))
 871    ;; Composite term: just expand it all.
 872    ((and (not (null expr)) (listp expr))
 873     (format "(%s)" (nnir-imap-query-to-imap criteria expr)))
 874    ;; Complex value, give up for now.
 875    (t (error "Unhandled input: %S" expr))))
 876
 877
 878 (defun nnir-imap-parse-query (string)
 879   "Turn STRING into an s-expression based query based on the IMAP
 880 query language as defined in `nnir-imap-make-query'.
 881
 882 This involves turning individual tokens into higher level terms
 883 that the search language can then understand and use."
 884   (with-temp-buffer
 885     ;; Set up the parsing environment.
 886     (insert string)
 887     (goto-char (point-min))
 888     ;; Now, collect the output terms and return them.
 889     (let (out)
 890       (while (not (nnir-imap-end-of-input))
 891         (push (nnir-imap-next-expr) out))
 892       (reverse out))))
 893
 894
 895 (defun nnir-imap-next-expr (&optional count)
 896   "Return the next expression from the current buffer."
 897   (let ((term (nnir-imap-next-term count))
 898         (next (nnir-imap-peek-symbol)))
 899     ;; Are we looking at an 'or' expression?
 900     (cond
 901      ;; Handle 'expr or expr'
 902      ((eq next 'or)
 903       (list 'or term (nnir-imap-next-expr 2)))
 904      ;; Anything else
 905      (t term))))
 906
 907
 908 (defun nnir-imap-next-term (&optional count)
 909   "Return the next TERM from the current buffer."
 910   (let ((term (nnir-imap-next-symbol count)))
 911     ;; What sort of term is this?
 912     (cond
 913      ;; and -- just ignore it
 914      ((eq term 'and) 'and)
 915      ;; negated term
 916      ((eq term 'not) (list 'not (nnir-imap-next-expr)))
 917      ;; generic term
 918      (t term))))
 919
 920
 921 (defun nnir-imap-peek-symbol ()
 922   "Return the next symbol from the current buffer, but don't consume it."
 923   (save-excursion
 924     (nnir-imap-next-symbol)))
 925
 926 (defun nnir-imap-next-symbol (&optional count)
 927   "Return the next symbol from the current buffer, or nil if we are
 928 at the end of the buffer.  If supplied COUNT skips some symbols before
 929 returning the one at the supplied position."
 930   (when (and (numberp count) (> count 1))
 931     (nnir-imap-next-symbol (1- count)))
 932   (let ((case-fold-search t))
 933     ;; end of input stream?
 934     (unless (nnir-imap-end-of-input)
 935       ;; No, return the next symbol from the stream.
 936       (cond
 937        ;; negated expression -- return it and advance one char.
 938        ((looking-at "-") (forward-char 1) 'not)
 939        ;; quoted string
 940        ((looking-at "\"") (nnir-imap-delimited-string "\""))
 941        ;; list expression -- we parse the content and return this as a list.
 942        ((looking-at "(")
 943         (nnir-imap-parse-query (nnir-imap-delimited-string ")")))
 944        ;; keyword input -- return a symbol version
 945        ((looking-at "\\band\\b") (forward-char 3) 'and)
 946        ((looking-at "\\bor\\b")  (forward-char 2) 'or)
 947        ((looking-at "\\bnot\\b") (forward-char 3) 'not)
 948        ;; Simple, boring keyword
 949        (t (let ((start (point))
 950                 (end (if (search-forward-regexp "[[:blank:]]" nil t)
 951                          (prog1
 952                              (match-beginning 0)
 953                            ;; unskip if we hit a non-blank terminal character.
 954                            (when (string-match "[^[:blank:]]" (match-string 0))
 955                              (backward-char 1)))
 956                        (goto-char (point-max)))))
 957             (buffer-substring start end)))))))
 958
 959 (defun nnir-imap-delimited-string (delimiter)
 960   "Return a delimited string from the current buffer."
 961   (let ((start (point)) end)
 962     (forward-char 1)                    ; skip the first delimiter.
 963     (while (not end)
 964       (unless (search-forward delimiter nil t)
 965         (error "Unmatched delimited input with %s in query" delimiter))
 966       (let ((here (point)))
 967         (unless (equal (buffer-substring (- here 2) (- here 1)) "\\")
 968           (setq end (point)))))
 969     (buffer-substring (1+ start) (1- end))))
 970
 971 (defun nnir-imap-end-of-input ()
 972   "Are we at the end of input?"
 973   (skip-chars-forward "[[:blank:]]")
 974   (looking-at "$"))
 975
 976
 977 ;; Swish++ interface.
 978 ;; -cc- Todo
 979 ;; Search by
 980 ;; - group
 981 ;; Sort by
 982 ;; - rank (default)
 983 ;; - article number
 984 ;; - file size
 985 ;; - group
 986 (defun nnir-run-swish++ (query server &optional group)
 987   "Run QUERY against swish++.
 988 Returns a vector of (group name, file name) pairs (also vectors,
 989 actually).
 990
 991 Tested with swish++ 4.7 on GNU/Linux and with swish++ 5.0b2 on
 992 Windows NT 4.0."
 993
 994   ;; (when group
 995   ;;   (error "The swish++ backend cannot search specific groups"))
 996
 997   (save-excursion
 998     (let ( (qstring (cdr (assq 'query query)))
 999            (groupspec (cdr (assq 'group query)))
1000            (prefix (nnir-read-server-parm 'nnir-swish++-remove-prefix server))
1001            artlist
1002            ;; nnml-use-compressed-files might be any string, but probably this
1003            ;; is sufficient.  Note that we can't only use the value of
1004            ;; nnml-use-compressed-files because old articles might have been
1005            ;; saved with a different value.
1006            (article-pattern (if (string= (gnus-group-server server) "nnmaildir")
1007                                 ":[0-9]+"
1008                               "^[0-9]+\\(\\.[a-z0-9]+\\)?$"))
1009            score artno dirnam filenam)
1010
1011       (when (equal "" qstring)
1012         (error "swish++: You didn't enter anything"))
1013
1014       (set-buffer (get-buffer-create nnir-tmp-buffer))
1015       (erase-buffer)
1016
1017       (if groupspec
1018           (message "Doing swish++ query %s on %s..." qstring groupspec)
1019         (message "Doing swish++ query %s..." qstring))
1020
1021       (let* ((cp-list `( ,nnir-swish++-program
1022                          nil            ; input from /dev/null
1023                          t              ; output
1024                          nil            ; don't redisplay
1025                          "--config-file" ,(nnir-read-server-parm 'nnir-swish++-configuration-file server)
1026                          ,@(nnir-read-server-parm 'nnir-swish++-additional-switches server)
1027                          ,qstring       ; the query, in swish++ format
1028                          ))
1029              (exitstatus
1030               (progn
1031                 (message "%s args: %s" nnir-swish++-program
1032                          (mapconcat 'identity (cddddr cp-list) " ")) ;; ???
1033                 (apply 'call-process cp-list))))
1034         (unless (or (null exitstatus)
1035                     (zerop exitstatus))
1036           (nnheader-report 'nnir "Couldn't run swish++: %s" exitstatus)
1037           ;; swish++ failure reason is in this buffer, show it if
1038           ;; the user wants it.
1039           (when (> gnus-verbose 6)
1040             (display-buffer nnir-tmp-buffer))))
1041
1042       ;; The results are output in the format of:
1043       ;; V 4.7 Linux
1044       ;; rank relative-path-name file-size file-title
1045       ;; V 5.0b2:
1046       ;; rank relative-path-name file-size topic??
1047       ;; where rank is an integer from 1 to 100.
1048       (goto-char (point-min))
1049       (while (re-search-forward
1050               "\\(^[0-9]+\\) \\([^ ]+\\) [0-9]+ \\(.*\\)$" nil t)
1051         (setq score (match-string 1)
1052               filenam (match-string 2)
1053               artno (file-name-nondirectory filenam)
1054               dirnam (file-name-directory filenam))
1055
1056         ;; don't match directories
1057         (when (string-match article-pattern artno)
1058           (when (not (null dirnam))
1059
1060             ;; maybe limit results to matching groups.
1061             (when (or (not groupspec)
1062                       (string-match groupspec dirnam))
1063               (nnir-add-result dirnam artno score prefix server artlist)))))
1064
1065       (message "Massaging swish++ output...done")
1066
1067       ;; Sort by score
1068       (apply 'vector
1069              (sort artlist
1070                    (function (lambda (x y)
1071                                (> (nnir-artitem-rsv x)
1072                                   (nnir-artitem-rsv y)))))))))
1073
1074 ;; Swish-E interface.
1075 (defun nnir-run-swish-e (query server &optional group)
1076   "Run given query against swish-e.
1077 Returns a vector of (group name, file name) pairs (also vectors,
1078 actually).
1079
1080 Tested with swish-e-2.0.1 on Windows NT 4.0."
1081
1082   ;; swish-e crashes with empty parameter to "-w" on commandline...
1083   ;; (when group
1084   ;;   (error "The swish-e backend cannot search specific groups"))
1085
1086   (save-excursion
1087     (let ((qstring (cdr (assq 'query query)))
1088           (prefix
1089            (or (nnir-read-server-parm 'nnir-swish-e-remove-prefix server)
1090                (error "Missing parameter `nnir-swish-e-remove-prefix'")))
1091           artlist score artno dirnam group )
1092
1093       (when (equal "" qstring)
1094         (error "swish-e: You didn't enter anything"))
1095
1096       (set-buffer (get-buffer-create nnir-tmp-buffer))
1097       (erase-buffer)
1098
1099       (message "Doing swish-e query %s..." query)
1100       (let* ((index-files
1101               (or (nnir-read-server-parm
1102                    'nnir-swish-e-index-files server)
1103                   (error "Missing parameter `nnir-swish-e-index-files'")))
1104              (additional-switches
1105               (nnir-read-server-parm
1106                'nnir-swish-e-additional-switches server))
1107              (cp-list `(,nnir-swish-e-program
1108                         nil             ; input from /dev/null
1109                         t               ; output
1110                         nil             ; don't redisplay
1111                         "-f" ,@index-files
1112                         ,@additional-switches
1113                         "-w"
1114                         ,qstring        ; the query, in swish-e format
1115                         ))
1116              (exitstatus
1117               (progn
1118                 (message "%s args: %s" nnir-swish-e-program
1119                          (mapconcat 'identity (cddddr cp-list) " "))
1120                 (apply 'call-process cp-list))))
1121         (unless (or (null exitstatus)
1122                     (zerop exitstatus))
1123           (nnheader-report 'nnir "Couldn't run swish-e: %s" exitstatus)
1124           ;; swish-e failure reason is in this buffer, show it if
1125           ;; the user wants it.
1126           (when (> gnus-verbose 6)
1127             (display-buffer nnir-tmp-buffer))))
1128
1129       ;; The results are output in the format of:
1130       ;; rank path-name file-title file-size
1131       (goto-char (point-min))
1132       (while (re-search-forward
1133               "\\(^[0-9]+\\) \\([^ ]+\\) \"\\([^\"]+\\)\" [0-9]+$" nil t)
1134         (setq score (match-string 1)
1135               artno (match-string 3)
1136               dirnam (file-name-directory (match-string 2)))
1137
1138         ;; don't match directories
1139         (when (string-match "^[0-9]+$" artno)
1140           (when (not (null dirnam))
1141
1142             ;; remove nnir-swish-e-remove-prefix from beginning of dirname
1143             (when (string-match (concat "^" prefix) dirnam)
1144               (setq dirnam (replace-match "" t t dirnam)))
1145
1146             (setq dirnam (substring dirnam 0 -1))
1147             ;; eliminate all ".", "/", "\" from beginning. Always matches.
1148             (string-match "^[./\\]*\\(.*\\)$" dirnam)
1149             ;; "/" -> "."
1150             (setq group (gnus-replace-in-string (match-string 1 dirnam) "/" "."))
1151             ;; Windows "\\" -> "."
1152             (setq group (gnus-replace-in-string group "\\\\" "."))
1153
1154             (push (vector (nnir-group-full-name group server)
1155                           (string-to-number artno)
1156                           (string-to-number score))
1157                   artlist))))
1158
1159       (message "Massaging swish-e output...done")
1160
1161       ;; Sort by score
1162       (apply 'vector
1163              (sort artlist
1164                    (function (lambda (x y)
1165                                (> (nnir-artitem-rsv x)
1166                                   (nnir-artitem-rsv y)))))))))
1167
1168 ;; HyREX interface
1169 (defun nnir-run-hyrex (query server &optional group)
1170   (save-excursion
1171     (let ((artlist nil)
1172           (groupspec (cdr (assq 'group query)))
1173           (qstring (cdr (assq 'query query)))
1174           (prefix (nnir-read-server-parm 'nnir-hyrex-remove-prefix server))
1175           score artno dirnam)
1176       (when (and (not groupspec) group)
1177         (setq groupspec
1178               (regexp-opt
1179                (mapcar (lambda (x) (gnus-group-real-name x)) group))))
1180       (set-buffer (get-buffer-create nnir-tmp-buffer))
1181       (erase-buffer)
1182       (message "Doing hyrex-search query %s..." query)
1183       (let* ((cp-list
1184               `( ,nnir-hyrex-program
1185                  nil                    ; input from /dev/null
1186                  t                      ; output
1187                  nil                    ; don't redisplay
1188                  "-i",(nnir-read-server-parm 'nnir-hyrex-index-directory server) ; index directory
1189                  ,@(nnir-read-server-parm 'nnir-hyrex-additional-switches server)
1190                  ,qstring          ; the query, in hyrex-search format
1191                  ))
1192              (exitstatus
1193               (progn
1194                 (message "%s args: %s" nnir-hyrex-program
1195                          (mapconcat 'identity (cddddr cp-list) " "))
1196                 (apply 'call-process cp-list))))
1197         (unless (or (null exitstatus)
1198                     (zerop exitstatus))
1199           (nnheader-report 'nnir "Couldn't run hyrex-search: %s" exitstatus)
1200           ;; nnir-search failure reason is in this buffer, show it if
1201           ;; the user wants it.
1202           (when (> gnus-verbose 6)
1203             (display-buffer nnir-tmp-buffer)))) ;; FIXME: Dont clear buffer !
1204       (message "Doing hyrex-search query \"%s\"...done" qstring)
1205       (sit-for 0)
1206       ;; nnir-search returns:
1207       ;;   for nnml/nnfolder: "filename mailid weigth"
1208       ;;   for nnimap:        "group mailid weigth"
1209       (goto-char (point-min))
1210       (delete-non-matching-lines "^\\S + [0-9]+ [0-9]+$")
1211       ;; HyREX doesn't search directly in groups -- so filter out here.
1212       (when groupspec
1213         (keep-lines groupspec))
1214       ;; extract data from result lines
1215       (goto-char (point-min))
1216       (while (re-search-forward
1217               "\\(\\S +\\) \\([0-9]+\\) \\([0-9]+\\)" nil t)
1218         (setq dirnam (match-string 1)
1219               artno (match-string 2)
1220               score (match-string 3))
1221         (when (string-match prefix dirnam)
1222           (setq dirnam (replace-match "" t t dirnam)))
1223         (push (vector (nnir-group-full-name
1224                        (gnus-replace-in-string dirnam "/" ".") server)
1225                       (string-to-number artno)
1226                       (string-to-number score))
1227               artlist))
1228       (message "Massaging hyrex-search output...done.")
1229       (apply 'vector
1230              (sort artlist
1231                    (function (lambda (x y)
1232                                (if (string-lessp (nnir-artitem-group x)
1233                                                  (nnir-artitem-group y))
1234                                    t
1235                                  (< (nnir-artitem-number x)
1236                                     (nnir-artitem-number y)))))))
1237       )))
1238
1239 ;; Namazu interface
1240 (defun nnir-run-namazu (query server &optional group)
1241   "Run given query against Namazu.  Returns a vector of (group name, file name)
1242 pairs (also vectors, actually).
1243
1244 Tested with Namazu 2.0.6 on a GNU/Linux system."
1245   ;; (when group
1246   ;;   (error "The Namazu backend cannot search specific groups"))
1247   (save-excursion
1248     (let ((article-pattern (if (string= (gnus-group-server server) "nnmaildir")
1249                                ":[0-9]+"
1250                              "^[0-9]+$"))
1251           artlist
1252           (qstring (cdr (assq 'query query)))
1253           (prefix (nnir-read-server-parm 'nnir-namazu-remove-prefix server))
1254           score group article
1255           (process-environment (copy-sequence process-environment)))
1256       (setenv "LC_MESSAGES" "C")
1257       (set-buffer (get-buffer-create nnir-tmp-buffer))
1258       (erase-buffer)
1259       (let* ((cp-list
1260               `( ,nnir-namazu-program
1261                  nil                    ; input from /dev/null
1262                  t                      ; output
1263                  nil                    ; don't redisplay
1264                  "-q"                   ; don't be verbose
1265                  "-a"                   ; show all matches
1266                  "-s"                   ; use short format
1267                  ,@(nnir-read-server-parm 'nnir-namazu-additional-switches server)
1268                  ,qstring               ; the query, in namazu format
1269                  ,(nnir-read-server-parm 'nnir-namazu-index-directory server) ; index directory
1270                  ))
1271              (exitstatus
1272               (progn
1273                 (message "%s args: %s" nnir-namazu-program
1274                          (mapconcat 'identity (cddddr cp-list) " "))
1275                 (apply 'call-process cp-list))))
1276         (unless (or (null exitstatus)
1277                     (zerop exitstatus))
1278           (nnheader-report 'nnir "Couldn't run namazu: %s" exitstatus)
1279           ;; Namazu failure reason is in this buffer, show it if
1280           ;; the user wants it.
1281           (when (> gnus-verbose 6)
1282             (display-buffer nnir-tmp-buffer))))
1283
1284       ;; Namazu output looks something like this:
1285       ;; 2. Re: Gnus agent expire broken (score: 55)
1286       ;; /home/henrik/Mail/mail/sent/1310 (4,138 bytes)
1287
1288       (goto-char (point-min))
1289       (while (re-search-forward
1290               "^\\([0-9]+\\.\\).*\\((score: \\([0-9]+\\)\\))\n\\([^ ]+\\)"
1291               nil t)
1292         (setq score (match-string 3)
1293               group (file-name-directory (match-string 4))
1294               article (file-name-nondirectory (match-string 4)))
1295
1296         ;; make sure article and group is sane
1297         (when (and (string-match article-pattern article)
1298                    (not (null group)))
1299           (nnir-add-result group article score prefix server artlist)))
1300
1301       ;; sort artlist by score
1302       (apply 'vector
1303              (sort artlist
1304                    (function (lambda (x y)
1305                                (> (nnir-artitem-rsv x)
1306                                   (nnir-artitem-rsv y)))))))))
1307
1308 (defun nnir-run-find-grep (query server &optional grouplist)
1309   "Run find and grep to obtain matching articles."
1310   (let* ((method (gnus-server-to-method server))
1311          (sym (intern
1312                (concat (symbol-name (car method)) "-directory")))
1313          (directory (cadr (assoc sym (cddr method))))
1314          (regexp (cdr (assoc 'query query)))
1315          (grep-options (cdr (assoc 'grep-options query)))
1316          artlist)
1317     (unless directory
1318       (error "No directory found in method specification of server %s"
1319              server))
1320     (apply
1321      'vconcat
1322      (mapcar (lambda (x)
1323                (let ((group x))
1324                  (message "Searching %s using find-grep..."
1325                           (or group server))
1326                  (save-window-excursion
1327                    (set-buffer (get-buffer-create nnir-tmp-buffer))
1328                    (erase-buffer)
1329                    (if (> gnus-verbose 6)
1330                        (pop-to-buffer (current-buffer)))
1331                    (cd directory) ; Using relative paths simplifies
1332                                   ; postprocessing.
1333                    (let ((group
1334                           (if (not group)
1335                               "."
1336                             ;; Try accessing the group literally as
1337                             ;; well as interpreting dots as directory
1338                             ;; separators so the engine works with
1339                             ;; plain nnml as well as the Gnus Cache.
1340                             (let ((group (gnus-group-real-name group)))
1341                               ;; Replace cl-func find-if.
1342                               (if (file-directory-p group)
1343                                   group
1344                                 (if (file-directory-p
1345                                      (setq group
1346                                            (gnus-replace-in-string
1347                                             group
1348                                             "\\." "/" t)))
1349                                     group))))))
1350                      (unless group
1351                        (error "Cannot locate directory for group"))
1352                      (save-excursion
1353                        (apply
1354                         'call-process "find" nil t
1355                         "find" group "-type" "f" "-name" "[0-9]*" "-exec"
1356                         "grep"
1357                         `("-l" ,@(and grep-options
1358                                       (split-string grep-options "\\s-" t))
1359                           "-e" ,regexp "{}" "+"))))
1360
1361                    ;; Translate relative paths to group names.
1362                    (while (not (eobp))
1363                      (let* ((path (split-string
1364                                    (buffer-substring
1365                                     (point)
1366                                     (line-end-position)) "/" t))
1367                             (art (string-to-number (car (last path)))))
1368                        (while (string= "." (car path))
1369                          (setq path (cdr path)))
1370                        (let ((group (mapconcat 'identity
1371                                                ;; Replace cl-func:
1372                                                ;; (subseq path 0 -1)
1373                                                (let ((end (1- (length path)))
1374                                                      res)
1375                                                  (while
1376                                                      (>= (setq end (1- end)) 0)
1377                                                    (push (pop path) res))
1378                                                  (nreverse res))
1379                                                ".")))
1380                          (push
1381                           (vector (nnir-group-full-name group server) art 0)
1382                           artlist))
1383                        (forward-line 1)))
1384                    (message "Searching %s using find-grep...done"
1385                             (or group server))
1386                    artlist)))
1387      grouplist))))
1388
1389 (declare-function mm-url-insert "mm-url" (url &optional follow-refresh))
1390 (declare-function mm-url-encode-www-form-urlencoded "mm-url" (pairs))
1391
1392 ;; gmane interface
1393 (defun nnir-run-gmane (query srv &optional groups)
1394   "Run a search against a gmane back-end server."
1395   (if (gnus-string-match-p "gmane" srv)
1396       (let* ((case-fold-search t)
1397              (qstring (cdr (assq 'query query)))
1398              (server (cadr (gnus-server-to-method srv)))
1399              (groupspec (if groups
1400                             (mapconcat
1401                              (function (lambda (x)
1402                                          (format "group:%s"
1403                                                  (gnus-group-short-name x))))
1404                              groups " ") ""))
1405              (authorspec
1406               (if (assq 'author query)
1407                   (format "author:%s" (cdr (assq 'author query))) ""))
1408              (search (format "%s %s %s"
1409                              qstring groupspec authorspec))
1410              artlist)
1411         (require 'mm-url)
1412         (with-current-buffer nntp-server-buffer
1413           (erase-buffer)
1414           (mm-url-insert
1415            (concat
1416             "http://search.gmane.org/nov.php"
1417             "?"
1418             (mm-url-encode-www-form-urlencoded
1419              `(("query" . ,search)
1420                ("HITSPERPAGE" . "999")))))
1421           (unless (featurep 'xemacs) (set-buffer-multibyte t))
1422           (mm-decode-coding-region (point-min) (point-max) 'utf-8)
1423           (goto-char (point-min))
1424           (forward-line 1)
1425           (while (not (eobp))
1426             (unless (or (eolp) (looking-at "\x0d"))
1427               (let ((header (nnheader-parse-nov)))
1428                 (let ((xref (mail-header-xref header))
1429                       (xscore (string-to-number (cdr (assoc 'X-Score
1430                                (mail-header-extra header))))))
1431                   (when (string-match " \\([^:]+\\)[:/]\\([0-9]+\\)" xref)
1432                     (push
1433                      (vector
1434                       (gnus-group-prefixed-name (match-string 1 xref) srv)
1435                       (string-to-number (match-string 2 xref)) xscore)
1436                      artlist)))))
1437             (forward-line 1)))
1438         ;; Sort by score
1439         (apply 'vector
1440                (sort artlist
1441                      (function (lambda (x y)
1442                                  (> (nnir-artitem-rsv x)
1443                                     (nnir-artitem-rsv y)))))))
1444     (message "Can't search non-gmane nntp groups")))
1445
1446 ;;; Util Code:
1447
1448 (defun nnir-read-parms (query nnir-search-engine)
1449   "Reads additional search parameters according to `nnir-engines'."
1450   (let ((parmspec (caddr (assoc nnir-search-engine nnir-engines))))
1451     (nconc query
1452            (mapcar 'nnir-read-parm parmspec))))
1453
1454 (defun nnir-read-parm (parmspec)
1455   "Reads a single search parameter.
1456 `parmspec' is a cons cell, the car is a symbol, the cdr is a prompt."
1457   (let ((sym (car parmspec))
1458         (prompt (cdr parmspec)))
1459     (if (listp prompt)
1460         (let* ((result (apply 'gnus-completing-read prompt))
1461                (mapping (or (assoc result nnir-imap-search-arguments)
1462                             (cons nil nnir-imap-search-other))))
1463           (cons sym (format (cdr mapping) result)))
1464       (cons sym (read-string prompt)))))
1465
1466 (defun nnir-run-query (query nserver)
1467   "Invoke appropriate search engine function (see `nnir-engines').
1468   If some groups were process-marked, run the query for each of the groups
1469   and concat the results."
1470   (let ((q (car (read-from-string query)))
1471         (groups (if (string= "all-ephemeral" nserver)
1472                     (with-current-buffer gnus-server-buffer
1473                       (list (list (gnus-server-server-name))))
1474                   (nnir-sort-groups-by-server
1475                    (or gnus-group-marked (list (gnus-group-group-name)))))))
1476     (apply 'vconcat
1477            (mapcar (lambda (x)
1478                      (let* ((server (car x))
1479                             (nnir-search-engine
1480                              (or (nnir-read-server-parm 'nnir-search-engine
1481                                                         server)
1482                                  (cdr (assoc (car
1483                                               (gnus-server-to-method server))
1484                                              nnir-method-default-engines))))
1485                             search-func)
1486                        (setq search-func (cadr
1487                                           (assoc nnir-search-engine
1488                                                  nnir-engines)))
1489                        (if search-func
1490                            (funcall search-func
1491                                     (if nnir-extra-parms
1492                                         (nnir-read-parms q nnir-search-engine)
1493                                       q)
1494                                     server (cdr x))
1495                          nil)))
1496                    groups))))
1497
1498 (defun nnir-read-server-parm (key server)
1499   "Returns the parameter value of key for the given server, where
1500 server is of form 'backend:name'."
1501   (let ((method (gnus-server-to-method server)))
1502     (cond ((and method (assq key (cddr method)))
1503            (nth 1 (assq key (cddr method))))
1504           (t nil))))
1505
1506 (defun nnir-group-full-name (shortname server)
1507   "For the given group name, return a full Gnus group name.
1508 The Gnus backend/server information is added."
1509   (gnus-group-prefixed-name shortname (gnus-server-to-method server)))
1510
1511 (defun nnir-possibly-change-server (server)
1512   (unless (and server (nnir-server-opened server))
1513     (nnir-open-server server)))
1514
1515
1516 ;; Data type article list.
1517
1518 (defun nnir-artlist-length (artlist)
1519   "Returns number of articles in artlist."
1520   (length artlist))
1521
1522 (defun nnir-artlist-article (artlist n)
1523   "Returns from ARTLIST the Nth artitem (counting starting at 1)."
1524   (elt artlist (1- n)))
1525
1526 (defun nnir-artitem-group (artitem)
1527   "Returns the group from the ARTITEM."
1528   (elt artitem 0))
1529
1530 (defun nnir-artlist-artitem-group (artlist n)
1531   "Returns from ARTLIST the group of the Nth artitem (counting from 1)."
1532   (nnir-artitem-group (nnir-artlist-article artlist n)))
1533
1534 (defun nnir-artitem-number (artitem)
1535   "Returns the number from the ARTITEM."
1536   (elt artitem 1))
1537
1538 (defun nnir-artlist-artitem-number (artlist n)
1539   "Returns from ARTLIST the number of the Nth artitem (counting from 1)."
1540   (nnir-artitem-number (nnir-artlist-article artlist n)))
1541
1542 (defun nnir-artitem-rsv (artitem)
1543   "Returns the Retrieval Status Value (RSV, score) from the ARTITEM."
1544   (elt artitem 2))
1545
1546 (defun nnir-artlist-artitem-rsv (artlist n)
1547   "Returns from ARTLIST the Retrieval Status Value of the Nth
1548 artitem (counting from 1)."
1549   (nnir-artitem-rsv (nnir-artlist-article artlist n)))
1550
1551 ;; unused?
1552 (defun nnir-artlist-groups (artlist)
1553   "Returns a list of all groups in the given ARTLIST."
1554   (let ((res nil)
1555         (with-dups nil))
1556     ;; from each artitem, extract group component
1557     (setq with-dups (mapcar 'nnir-artitem-group artlist))
1558     ;; remove duplicates from above
1559     (mapc (function (lambda (x) (add-to-list 'res x)))
1560             with-dups)
1561     res))
1562
1563 (defun nnir-sort-groups-by-server (groups)
1564   "sorts a list of groups into an alist keyed by server"
1565 (if (car groups)
1566   (let (value)
1567     (dolist (var groups value)
1568       (let ((server (gnus-group-server var)))
1569         (if (assoc server value)
1570             (nconc (cdr (assoc server value)) (list var))
1571           (push (cons (gnus-group-server var) (list var)) value))))
1572     value)
1573   nil))
1574
1575 (defun nnir-get-active (srv)
1576   (let ((method (gnus-server-to-method srv))
1577         groups)
1578     (gnus-request-list method)
1579     (with-current-buffer nntp-server-buffer
1580       (let ((cur (current-buffer))
1581             name)
1582         (goto-char (point-min))
1583         (unless (string= gnus-ignored-newsgroups "")
1584           (delete-matching-lines gnus-ignored-newsgroups))
1585         ;; We treat NNTP as a special case to avoid problems with
1586         ;; garbage group names like `"foo' that appear in some badly
1587         ;; managed active files. -jh.
1588         (if (eq (car method) 'nntp)
1589             (while (not (eobp))
1590               (ignore-errors
1591                 (push (cons
1592                        (mm-string-as-unibyte
1593                         (buffer-substring
1594                          (point)
1595                          (progn
1596                            (skip-chars-forward "^ \t")
1597                            (point))))
1598                        (let ((last (read cur)))
1599                          (cons (read cur) last)))
1600                       groups))
1601               (forward-line))
1602           (while (not (eobp))
1603             (ignore-errors
1604               (push (mm-string-as-unibyte
1605                      (let ((p (point)))
1606                        (skip-chars-forward "^ \t\\\\")
1607                        (setq name (buffer-substring (+ p 1) (- (point) 1)))
1608                        (gnus-group-full-name name method)))
1609                     groups))
1610             (forward-line)))))
1611     groups))
1612
1613 ;; The end.
1614 (provide 'nnir)
1615
1616 ;;; nnir.el ends here