1 ;;; nnweb.el --- retrieving articles via web search engines
2 ;; Copyright (C) 1996,97 Free Software Foundation, Inc.
4 ;; Author: Lars Magne Ingebrigtsen <larsi@ifi.uio.no>
7 ;; This file is part of GNU Emacs.
9 ;; GNU Emacs is free software; you can redistribute it and/or modify
10 ;; it under the terms of the GNU General Public License as published by
11 ;; the Free Software Foundation; either version 2, or (at your option)
14 ;; GNU Emacs is distributed in the hope that it will be useful,
15 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
16 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 ;; GNU General Public License for more details.
19 ;; You should have received a copy of the GNU General Public License
20 ;; along with GNU Emacs; see the file COPYING. If not, write to the
21 ;; Free Software Foundation, Inc., 59 Temple Place - Suite 330,
22 ;; Boston, MA 02111-1307, USA.
26 ;; Note: You need to have `url' and `w3' installed for this
31 (eval-when-compile (require 'cl))
45 (defvoo nnweb-directory (nnheader-concat gnus-directory "nnweb/")
46 "Where nnweb will save its files.")
48 (defvoo nnweb-type 'dejanews
49 "What search engine type is being used.")
51 (defvar nnweb-type-definition
53 (article . nnweb-dejanews-wash-article)
54 (map . nnweb-dejanews-create-mapping)
55 (search . nnweb-dejanews-search)
56 (address . "http://xp9.dejanews.com/dnquery.xp")
57 (identifier . nnweb-dejanews-identity))
59 (article . nnweb-reference-wash-article)
60 (map . nnweb-reference-create-mapping)
61 (search . nnweb-reference-search)
62 (address . "http://www.reference.com/cgi-bin/pn/go")
63 (identifier . identity))
65 (article . nnweb-altavista-wash-article)
66 (map . nnweb-altavista-create-mapping)
67 (search . nnweb-altavista-search)
68 (address . "http://www.altavista.digital.com/cgi-bin/query")
69 (id . "/cgi-bin/news?id@%s")
70 (identifier . identity)))
71 "Type-definition alist.")
73 (defvoo nnweb-search nil
74 "Search string to feed to DejaNews.")
76 (defvoo nnweb-max-hits 100
77 "Maximum number of hits to display.")
79 (defvoo nnweb-ephemeral-p nil
80 "Whether this nnweb server is ephemeral.")
82 ;;; Internal variables
84 (defvoo nnweb-articles nil)
85 (defvoo nnweb-buffer nil)
86 (defvoo nnweb-group-alist nil)
87 (defvoo nnweb-group nil)
88 (defvoo nnweb-hashtb nil)
90 ;;; Interface functions
92 (nnoo-define-basics nnweb)
94 (deffoo nnweb-retrieve-headers (articles &optional group server fetch-old)
95 (nnweb-possibly-change-server group server)
97 (set-buffer nntp-server-buffer)
100 (while (setq article (pop articles))
101 (when (setq header (cadr (assq article nnweb-articles)))
102 (nnheader-insert-nov header)))
105 (deffoo nnweb-request-scan (&optional group server)
106 (nnweb-possibly-change-server group server)
107 (setq nnweb-hashtb (gnus-make-hashtable 4095))
108 (funcall (nnweb-definition 'map))
109 (unless nnweb-ephemeral-p
111 (nnweb-write-overview group)))
113 (deffoo nnweb-request-group (group &optional server dont-check)
114 (nnweb-possibly-change-server nil server)
116 (not (equal group nnweb-group))
117 (not nnweb-ephemeral-p))
118 (let ((info (assoc group nnweb-group-alist)))
119 (setq nnweb-group group)
120 (setq nnweb-type (nth 2 info))
121 (setq nnweb-search (nth 3 info))
123 (nnweb-read-overview group))))
125 ((not nnweb-articles)
126 (nnheader-report 'nnweb "No matching articles"))
128 (let ((active (if nnweb-ephemeral-p
129 (cons (caar nnweb-articles)
130 (caar (last nnweb-articles)))
131 (cadr (assoc group nnweb-group-alist)))))
132 (nnheader-report 'nnweb "Opened group %s" group)
134 "211 %d %d %d %s\n" (length nnweb-articles)
135 (car active) (cdr active) group)))))
137 (deffoo nnweb-close-group (group &optional server)
138 (nnweb-possibly-change-server group server)
139 (when (gnus-buffer-live-p nnweb-buffer)
141 (set-buffer nnweb-buffer)
142 (set-buffer-modified-p nil)
143 (kill-buffer nnweb-buffer)))
146 (deffoo nnweb-request-article (article &optional group server buffer)
147 (nnweb-possibly-change-server group server)
149 (set-buffer (or buffer nntp-server-buffer))
150 (let* ((header (cadr (assq article nnweb-articles)))
151 (url (and header (mail-header-xref header))))
153 (nnweb-fetch-url url))
154 (and (stringp article)
155 (nnweb-definition 'id t)
156 (let ((fetch (nnweb-definition 'id))
158 (when (string-match "^<\\(.*\\)>$" article)
159 (setq art (match-string 1 article)))
163 (format fetch article))))))
164 (unless nnheader-callback-function
165 (funcall (nnweb-definition 'article))
166 (nnweb-decode-entities))
167 (nnheader-report 'nnweb "Fetched article %s" article)
170 (deffoo nnweb-close-server (&optional server)
171 (when (and (nnweb-server-opened server)
172 (gnus-buffer-live-p nnweb-buffer))
174 (set-buffer nnweb-buffer)
175 (set-buffer-modified-p nil)
176 (kill-buffer nnweb-buffer)))
177 (nnoo-close-server 'nnweb server))
179 (deffoo nnweb-request-list (&optional server)
180 (nnweb-possibly-change-server nil server)
182 (set-buffer nntp-server-buffer)
183 (nnmail-generate-active nnweb-group-alist)
186 (deffoo nnweb-request-update-info (group info &optional server)
187 (nnweb-possibly-change-server group server)
188 ;;(setcar (cddr info) nil)
191 (deffoo nnweb-asynchronous-p ()
194 (deffoo nnweb-request-create-group (group &optional server args)
195 (nnweb-possibly-change-server nil server)
196 (nnweb-request-delete-group group)
197 (push `(,group ,(cons 1 0) ,@args) nnweb-group-alist)
201 (deffoo nnweb-request-delete-group (group &optional force server)
202 (nnweb-possibly-change-server group server)
203 (gnus-delete-assoc group nnweb-group-alist)
204 (gnus-delete-file (nnweb-overview-file group))
207 (nnoo-define-skeleton nnweb)
209 ;;; Internal functions
211 (defun nnweb-read-overview (group)
212 "Read the overview of GROUP and build the map."
213 (when (file-exists-p (nnweb-overview-file group))
214 (nnheader-temp-write nil
215 (nnheader-insert-file-contents (nnweb-overview-file group))
216 (goto-char (point-min))
219 (setq header (nnheader-parse-nov))
221 (push (list (mail-header-number header)
222 header (mail-header-xref header))
224 (nnweb-set-hashtb header (car nnweb-articles)))))))
226 (defun nnweb-write-overview (group)
227 "Write the overview file for GROUP."
228 (nnheader-temp-write (nnweb-overview-file group)
229 (let ((articles nnweb-articles))
231 (nnheader-insert-nov (cadr (pop articles)))))))
233 (defun nnweb-set-hashtb (header data)
234 (gnus-sethash (nnweb-identifier (mail-header-xref header))
237 (defun nnweb-get-hashtb (url)
238 (gnus-gethash (nnweb-identifier url) nnweb-hashtb))
240 (defun nnweb-identifier (ident)
241 (funcall (nnweb-definition 'identifier) ident))
243 (defun nnweb-overview-file (group)
244 "Return the name of the overview file of GROUP."
245 (nnheader-concat nnweb-directory group ".overview"))
247 (defun nnweb-write-active ()
248 "Save the active file."
249 (nnheader-temp-write (nnheader-concat nnweb-directory "active")
250 (prin1 `(setq nnweb-group-alist ',nnweb-group-alist) (current-buffer))))
252 (defun nnweb-read-active ()
253 "Read the active file."
254 (load (nnheader-concat nnweb-directory "active") t t t))
256 (defun nnweb-definition (type &optional noerror)
257 "Return the definition of TYPE."
258 (let ((def (cdr (assq type (assq nnweb-type nnweb-type-definition)))))
261 (error "Undefined definition %s" type))
264 (defun nnweb-possibly-change-server (&optional group server)
267 (unless (nnweb-server-opened server)
268 (nnweb-open-server server)))
269 (unless nnweb-group-alist
272 (when (and (not nnweb-ephemeral-p)
273 (not (equal group nnweb-group)))
274 (nnweb-request-group group nil t))))
276 (defun nnweb-init (server)
277 "Initialize buffers and such."
278 (unless (gnus-buffer-live-p nnweb-buffer)
281 (nnheader-set-temp-buffer
282 (format " *nnweb %s %s %s*" nnweb-type nnweb-search server))))))
284 (defun nnweb-fetch-url (url)
286 (if (not nnheader-callback-function)
287 (let ((buf (current-buffer)))
289 (set-buffer nnweb-buffer)
291 (url-insert-file-contents url)
292 (copy-to-buffer buf (point-min) (point-max))
294 (nnweb-url-retrieve-asynch
295 url 'nnweb-callback (current-buffer) nnheader-callback-function)
298 (defun nnweb-callback (buffer callback)
299 (when (gnus-buffer-live-p url-working-buffer)
301 (set-buffer url-working-buffer)
302 (funcall (nnweb-definition 'article))
303 (nnweb-decode-entities)
305 (goto-char (point-max))
306 (insert-buffer-substring url-working-buffer))
308 (gnus-kill-buffer url-working-buffer)))
310 (defun nnweb-url-retrieve-asynch (url callback &rest data)
311 (let ((url-request-method "GET")
312 (old-asynch url-be-asynchronous)
313 (url-request-data nil)
314 (url-request-extra-headers nil)
315 (url-working-buffer (generate-new-buffer-name " *nnweb*")))
316 (setq-default url-be-asynchronous t)
318 (set-buffer (get-buffer-create url-working-buffer))
319 (setq url-current-callback-data data
320 url-be-asynchronous t
321 url-current-callback-func callback)
323 (setq-default url-be-asynchronous old-asynch)))
325 (defun nnweb-encode-www-form-urlencoded (pairs)
326 "Return PAIRS encoded for forms."
330 (concat (w3-form-encode-xwfu (car data)) "="
331 (w3-form-encode-xwfu (cdr data)))))
334 (defun nnweb-fetch-form (url pairs)
335 (let ((url-request-data (nnweb-encode-www-form-urlencoded pairs))
336 (url-request-method "POST")
337 (url-request-extra-headers
338 '(("Content-type" . "application/x-www-form-urlencoded"))))
339 (url-insert-file-contents url)
340 (setq buffer-file-name nil))
343 (defun nnweb-decode-entities ()
344 (goto-char (point-min))
345 (while (re-search-forward "&\\([a-z]+\\);" nil t)
346 (replace-match (char-to-string (or (cdr (assq (intern (match-string 1))
351 (defun nnweb-remove-markup ()
352 (goto-char (point-min))
353 (while (search-forward "<!--" nil t)
354 (delete-region (match-beginning 0)
355 (or (search-forward "-->" nil t)
357 (goto-char (point-min))
358 (while (re-search-forward "<[^>]+>" nil t)
359 (replace-match "" t t)))
362 ;;; DejaNews functions.
365 (defun nnweb-dejanews-create-mapping ()
366 "Perform the search and create an number-to-url alist."
368 (set-buffer nnweb-buffer)
370 (when (funcall (nnweb-definition 'search) nnweb-search)
374 (active (or (cadr (assoc nnweb-group nnweb-group-alist))
376 Subject Score Date Newsgroup Author
379 ;; Go through all the article hits on this page.
380 (goto-char (point-min))
381 (nnweb-decode-entities)
382 (goto-char (point-min))
383 (while (re-search-forward "^ +[0-9]+\\." nil t)
386 (cond ((re-search-forward "^ +[0-9]+\\." nil t)
388 ((search-forward "\n\n" nil t)
392 (goto-char (point-min))
393 (when (looking-at ".*HREF=\"\\([^\"]+\\)\"")
394 (setq url (match-string 1)))
395 (nnweb-remove-markup)
396 (goto-char (point-min))
397 (while (search-forward "\t" nil t)
399 (goto-char (point-min))
400 (while (re-search-forward "^ +\\([^:]+\\): +\\(.*\\)$" nil t)
401 (set (intern (match-string 1)) (match-string 2)))
403 (when (string-match "#[0-9]+/[0-9]+ *$" Subject)
404 (setq Subject (substring Subject 0 (match-beginning 0))))
406 (unless (nnweb-get-hashtb url)
410 (make-full-mail-header
411 (cdr active) (concat "(" Newsgroup ") " Subject) Author Date
412 (concat "<" (nnweb-identifier url) "@dejanews>")
413 nil 0 (string-to-int Score) url))
415 (nnweb-set-hashtb (cadar map) (car map))))
416 ;; See whether there is a "Get next 20 hits" button here.
417 (if (or (not (re-search-forward
418 "HREF=\"\\([^\"]+\\)\">Get next" nil t))
419 (>= i nnweb-max-hits))
422 (setq more (match-string 1))
424 (url-insert-file-contents more)))
425 ;; Return the articles in the right order.
427 (sort (nconc nnweb-articles map) 'car-less-than-car))))))
429 (defun nnweb-dejanews-wash-article ()
430 (let ((case-fold-search t))
431 (goto-char (point-min))
432 (re-search-forward "<PRE>" nil t)
433 (delete-region (point-min) (point))
434 (re-search-forward "</PRE>" nil t)
435 (delete-region (point) (point-max))
436 (nnweb-remove-markup)
437 (goto-char (point-min))
438 (while (and (looking-at " *$")
441 (while (looking-at "\\(^[^ ]+:\\) *")
442 (replace-match "\\1 " t)
444 (when (re-search-forward "\n\n+" nil t)
445 (replace-match "\n" t t))
446 (goto-char (point-min))
447 (when (search-forward "[More Headers]" nil t)
448 (replace-match "" t t))))
450 (defun nnweb-dejanews-search (search)
452 (nnweb-definition 'address)
453 `(("query" . ,search)
454 ("defaultOp" . "AND")
455 ("svcclass" . "dncurrent")
457 ("format" . "verbose")
459 ("showsort" . "score")
461 ("ageweight" . "1")))
464 (defun nnweb-dejanews-identity (url)
465 "Return an unique identifier based on URL."
466 (if (string-match "recnum=\\([0-9]+\\)" url)
474 (defun nnweb-reference-create-mapping ()
475 "Perform the search and create an number-to-url alist."
477 (set-buffer nnweb-buffer)
479 (when (funcall (nnweb-definition 'search) nnweb-search)
483 (active (or (cadr (assoc nnweb-group nnweb-group-alist))
485 Subject Score Date Newsgroups From Message-ID
488 ;; Go through all the article hits on this page.
489 (goto-char (point-min))
490 (search-forward "</pre><hr>" nil t)
491 (delete-region (point-min) (point))
492 ;(nnweb-decode-entities)
493 (goto-char (point-min))
494 (while (re-search-forward "^ +[0-9]+\\." nil t)
497 (if (re-search-forward "^$" nil t)
500 (goto-char (point-min))
501 (when (looking-at ".*href=\"\\([^\"]+\\)\"")
502 (setq url (match-string 1)))
503 (nnweb-remove-markup)
504 (goto-char (point-min))
505 (while (search-forward "\t" nil t)
507 (goto-char (point-min))
508 (while (re-search-forward "^\\([^:]+\\): \\(.*\\)$" nil t)
509 (set (intern (match-string 1)) (match-string 2)))
511 (search-forward "</pre>" nil t)
513 (unless (nnweb-get-hashtb url)
517 (make-full-mail-header
518 (cdr active) (concat "(" Newsgroups ") " Subject) From Date
520 nil 0 (string-to-int Score) url))
522 (nnweb-set-hashtb (cadar map) (car map))))
524 ;; Return the articles in the right order.
526 (sort (nconc nnweb-articles map) 'car-less-than-car))))))
528 (defun nnweb-reference-wash-article ()
529 (let ((case-fold-search t))
530 (goto-char (point-min))
531 (re-search-forward "^</center><hr>" nil t)
532 (delete-region (point-min) (point))
533 (search-forward "<pre>" nil t)
535 (let ((body (point-marker)))
536 (search-forward "</pre>" nil t)
537 (delete-region (point) (point-max))
538 (nnweb-remove-markup)
539 (goto-char (point-min))
540 (while (looking-at " *$")
542 (narrow-to-region (point-min) body)
543 (while (and (re-search-forward "^$" nil t)
546 (goto-char (point-min))
547 (while (looking-at "\\(^[^ ]+:\\) *")
548 (replace-match "\\1 " t)
550 (goto-char (point-min))
551 (when (re-search-forward "^References:" nil t)
553 (point) (if (re-search-forward "^$\\|^[^:]+:" nil t)
556 (goto-char (point-min))
558 (unless (looking-at "References")
561 (goto-char (point-min))
562 (while (search-forward "," nil t)
563 (replace-match " " t t)))
565 (set-marker body nil))))
567 (defun nnweb-reference-search (search)
568 (url-insert-file-contents
570 (nnweb-definition 'address)
572 (nnweb-encode-www-form-urlencoded
573 `(("search" . "advanced")
574 ("querytext" . ,search)
579 ("organization" . "")
582 ("choice" . "Search")
583 ("startmonth" . "Jul")
585 ("startyear" . "1996")
590 ("verbosity" . "Verbose")
591 ("ranking" . "Relevance")
595 (setq buffer-file-name nil)
602 (defun nnweb-altavista-create-mapping ()
603 "Perform the search and create an number-to-url alist."
605 (set-buffer nnweb-buffer)
608 (when (funcall (nnweb-definition 'search) nnweb-search part)
612 (active (or (cadr (assoc nnweb-group nnweb-group-alist))
614 subject date from id group
617 ;; Go through all the article hits on this page.
618 (goto-char (point-min))
619 (search-forward "<dt>" nil t)
620 (delete-region (point-min) (match-beginning 0))
621 (goto-char (point-min))
622 (while (search-forward "<dt>" nil t)
623 (replace-match "\n<blubb>"))
624 (nnweb-decode-entities)
625 (goto-char (point-min))
626 (while (re-search-forward "<blubb>.*href=\"\\([^\"]+\\)\"><strong>\\([^>]*\\)</strong></a><dd>\\([^-]+\\)- <b>\\([^<]+\\)<.*href=\"news:\\([^\"]+\\)\">.*\">\\(.+\\)</a><P>"
628 (setq url (match-string 1)
629 subject (match-string 2)
630 date (match-string 3)
631 group (match-string 4)
632 id (concat "<" (match-string 5) ">")
633 from (match-string 6))
635 (unless (nnweb-get-hashtb url)
639 (make-full-mail-header
640 (cdr active) (concat "(" group ") " subject) from date
643 (nnweb-set-hashtb (cadar map) (car map))))
644 ;; See if we want more.
645 (when (or (not nnweb-articles)
646 (>= i nnweb-max-hits)
647 (not (funcall (nnweb-definition 'search)
648 nnweb-search (incf part))))
650 ;; Return the articles in the right order.
652 (sort (nconc nnweb-articles map) 'car-less-than-car)))))))
654 (defun nnweb-altavista-wash-article ()
655 (goto-char (point-min))
656 (let ((case-fold-search t))
657 (when (re-search-forward "^<strong>" nil t)
658 (delete-region (point-min) (match-beginning 0)))
659 (goto-char (point-min))
660 (while (looking-at "<strong>\\([^ ]+\\) +</strong> +\\(.*\\)$")
661 (replace-match "\\1: \\2" t)
663 (when (re-search-backward "^References:" nil t)
664 (narrow-to-region (point) (progn (forward-line 1) (point)))
665 (goto-char (point-min))
666 (while (re-search-forward "<A.*\\?id@\\([^\"]+\\)\">[0-9]+</A>" nil t)
667 (replace-match "<\\1> " t)))
669 (nnweb-remove-markup)))
671 (defun nnweb-altavista-search (search &optional part)
672 (url-insert-file-contents
674 (nnweb-definition 'address)
676 (nnweb-encode-www-form-urlencoded
679 ,@(when part `(("stq" . ,(int-to-string (* part 30)))))
685 (setq buffer-file-name nil)
690 ;;; nnweb.el ends here