1 ;;; nnweb.el --- retrieving articles via web search engines
3 ;; Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
4 ;; 2004, 2005, 2006, 2007, 2008, 2009 Free Software Foundation, Inc.
6 ;; Author: Lars Magne Ingebrigtsen <larsi@gnus.org>
9 ;; This file is part of GNU Emacs.
11 ;; GNU Emacs is free software: you can redistribute it and/or modify
12 ;; it under the terms of the GNU General Public License as published by
13 ;; the Free Software Foundation, either version 3 of the License, or
14 ;; (at your option) any later version.
16 ;; GNU Emacs is distributed in the hope that it will be useful,
17 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
18 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19 ;; GNU General Public License for more details.
21 ;; You should have received a copy of the GNU General Public License
22 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
26 ;; Note: You need to have `w3' installed for some functions to work.
30 (eval-when-compile (require 'cl))
42 (autoload 'w3-parse-buffer "w3-parse")
46 (defvoo nnweb-directory (nnheader-concat gnus-directory "nnweb/")
47 "Where nnweb will save its files.")
49 (defvoo nnweb-type 'google
50 "What search engine type is being used.
51 Valid types include `google', `dejanews', and `gmane'.")
53 (defvar nnweb-type-definition
55 (id . "http://www.google.com/groups?as_umsgid=%s&hl=en&dmode=source")
56 (result . "http://groups.google.com/group/%s/msg/%s?dmode=source")
57 (article . nnweb-google-wash-article)
58 (reference . identity)
59 (map . nnweb-google-create-mapping)
60 (search . nnweb-google-search)
61 (address . "http://groups.google.com/groups")
62 (base . "http://groups.google.com")
63 (identifier . nnweb-google-identity))
64 (dejanews ;; alias of google
65 (id . "http://www.google.com/groups?as_umsgid=%s&hl=en&dmode=source")
66 (result . "http://groups.google.com/group/%s/msg/%s?dmode=source")
67 (article . nnweb-google-wash-article)
68 (reference . identity)
69 (map . nnweb-google-create-mapping)
70 (search . nnweb-google-search)
71 (address . "http://groups.google.com/groups")
72 (base . "http://groups.google.com")
73 (identifier . nnweb-google-identity))
75 (article . nnweb-gmane-wash-article)
76 (id . "http://gmane.org/view.php?group=%s")
77 (reference . identity)
78 (map . nnweb-gmane-create-mapping)
79 (search . nnweb-gmane-search)
80 (address . "http://search.gmane.org/nov.php")
81 (identifier . nnweb-gmane-identity)))
82 "Type-definition alist.")
84 (defvoo nnweb-search nil
85 "Search string to feed to Google.")
87 (defvoo nnweb-max-hits 999
88 "Maximum number of hits to display.")
90 (defvoo nnweb-ephemeral-p nil
91 "Whether this nnweb server is ephemeral.")
93 ;;; Internal variables
95 (defvoo nnweb-articles nil)
96 (defvoo nnweb-buffer nil)
97 (defvoo nnweb-group-alist nil)
98 (defvoo nnweb-group nil)
99 (defvoo nnweb-hashtb nil)
101 ;;; Interface functions
103 (nnoo-define-basics nnweb)
105 (deffoo nnweb-retrieve-headers (articles &optional group server fetch-old)
106 (nnweb-possibly-change-server group server)
108 (set-buffer nntp-server-buffer)
110 (let (article header)
111 (mm-with-unibyte-current-buffer
112 (while (setq article (pop articles))
113 (when (setq header (cadr (assq article nnweb-articles)))
114 (nnheader-insert-nov header))))
117 (deffoo nnweb-request-scan (&optional group server)
118 (nnweb-possibly-change-server group server)
119 (if nnweb-ephemeral-p
120 (setq nnweb-hashtb (gnus-make-hashtable 4095))
121 (unless nnweb-articles
122 (nnweb-read-overview group)))
123 (funcall (nnweb-definition 'map))
124 (unless nnweb-ephemeral-p
126 (nnweb-write-overview group)))
128 (deffoo nnweb-request-group (group &optional server dont-check)
129 (nnweb-possibly-change-server group server)
130 (unless (or nnweb-ephemeral-p
133 (nnweb-read-overview group))
135 ((not nnweb-articles)
136 (nnheader-report 'nnweb "No matching articles"))
138 (let ((active (if nnweb-ephemeral-p
139 (cons (caar nnweb-articles)
140 (caar (last nnweb-articles)))
141 (cadr (assoc group nnweb-group-alist)))))
142 (nnheader-report 'nnweb "Opened group %s" group)
144 "211 %d %d %d %s\n" (length nnweb-articles)
145 (car active) (cdr active) group)))))
147 (deffoo nnweb-close-group (group &optional server)
148 (nnweb-possibly-change-server group server)
149 (when (gnus-buffer-live-p nnweb-buffer)
151 (set-buffer nnweb-buffer)
152 (set-buffer-modified-p nil)
153 (kill-buffer nnweb-buffer)))
156 (deffoo nnweb-request-article (article &optional group server buffer)
157 (nnweb-possibly-change-server group server)
159 (set-buffer (or buffer nntp-server-buffer))
160 (let* ((header (cadr (assq article nnweb-articles)))
161 (url (and header (mail-header-xref header))))
163 (mm-with-unibyte-current-buffer
164 (mm-url-insert url)))
165 (and (stringp article)
166 (nnweb-definition 'id t)
167 (let ((fetch (nnweb-definition 'id))
169 (when (string-match "^<\\(.*\\)>$" article)
170 (setq art (match-string 1 article)))
171 (when (and fetch art)
172 (setq url (format fetch
173 (mm-url-form-encode-xwfu art)))
174 (mm-with-unibyte-current-buffer
176 (if (nnweb-definition 'reference t)
178 (funcall (nnweb-definition
179 'reference) article)))))))
180 (unless nnheader-callback-function
181 (funcall (nnweb-definition 'article)))
182 (nnheader-report 'nnweb "Fetched article %s" article)
183 (cons group (and (numberp article) article))))))
185 (deffoo nnweb-close-server (&optional server)
186 (when (and (nnweb-server-opened server)
187 (gnus-buffer-live-p nnweb-buffer))
189 (set-buffer nnweb-buffer)
190 (set-buffer-modified-p nil)
191 (kill-buffer nnweb-buffer)))
192 (nnoo-close-server 'nnweb server))
194 (deffoo nnweb-request-list (&optional server)
195 (nnweb-possibly-change-server nil server)
197 (set-buffer nntp-server-buffer)
198 (nnmail-generate-active (list (assoc server nnweb-group-alist)))
201 (deffoo nnweb-request-update-info (group info &optional server)
202 (nnweb-possibly-change-server group server))
204 (deffoo nnweb-asynchronous-p ()
207 (deffoo nnweb-request-create-group (group &optional server args)
208 (nnweb-possibly-change-server nil server)
209 (nnweb-request-delete-group group)
210 (push `(,group ,(cons 1 0)) nnweb-group-alist)
214 (deffoo nnweb-request-delete-group (group &optional force server)
215 (nnweb-possibly-change-server group server)
216 (gnus-pull group nnweb-group-alist t)
218 (gnus-delete-file (nnweb-overview-file group))
221 (nnoo-define-skeleton nnweb)
223 ;;; Internal functions
225 (defun nnweb-read-overview (group)
226 "Read the overview of GROUP and build the map."
227 (when (file-exists-p (nnweb-overview-file group))
228 (mm-with-unibyte-buffer
229 (nnheader-insert-file-contents (nnweb-overview-file group))
230 (goto-char (point-min))
233 (setq header (nnheader-parse-nov))
235 (push (list (mail-header-number header)
236 header (mail-header-xref header))
238 (nnweb-set-hashtb header (car nnweb-articles)))))))
240 (defun nnweb-write-overview (group)
241 "Write the overview file for GROUP."
242 (with-temp-file (nnweb-overview-file group)
243 (let ((articles nnweb-articles))
245 (nnheader-insert-nov (cadr (pop articles)))))))
247 (defun nnweb-set-hashtb (header data)
248 (gnus-sethash (nnweb-identifier (mail-header-xref header))
251 (defun nnweb-get-hashtb (url)
252 (gnus-gethash (nnweb-identifier url) nnweb-hashtb))
254 (defun nnweb-identifier (ident)
255 (funcall (nnweb-definition 'identifier) ident))
257 (defun nnweb-overview-file (group)
258 "Return the name of the overview file of GROUP."
259 (nnheader-concat nnweb-directory group ".overview"))
261 (defun nnweb-write-active ()
262 "Save the active file."
263 (gnus-make-directory nnweb-directory)
264 (with-temp-file (nnheader-concat nnweb-directory "active")
265 (prin1 `(setq nnweb-group-alist ',nnweb-group-alist) (current-buffer))))
267 (defun nnweb-read-active ()
268 "Read the active file."
269 (load (nnheader-concat nnweb-directory "active") t t t))
271 (defun nnweb-definition (type &optional noerror)
272 "Return the definition of TYPE."
273 (let ((def (cdr (assq type (assq nnweb-type nnweb-type-definition)))))
276 (error "Undefined definition %s" type))
279 (defun nnweb-possibly-change-server (&optional group server)
281 (unless (nnweb-server-opened server)
282 (nnweb-open-server server))
284 (unless nnweb-group-alist
287 (setq nnweb-hashtb (gnus-make-hashtable 4095)))
289 (setq nnweb-group group)))
291 (defun nnweb-init (server)
292 "Initialize buffers and such."
293 (unless (gnus-buffer-live-p nnweb-buffer)
296 (nnheader-set-temp-buffer
297 (format " *nnweb %s %s %s*"
298 nnweb-type nnweb-search server))
299 (mm-disable-multibyte)
303 ;;; groups.google.com
306 (defun nnweb-google-wash-article ()
307 ;; We have Google's masked e-mail addresses here. :-/
308 (let ((case-fold-search t)
309 (start-re "<pre>[\r\n ]*")
310 (end-re "[\r\n ]*</pre>"))
311 (goto-char (point-min))
313 (or (re-search-forward "The requested message.*could not be found."
315 (not (and (re-search-forward start-re nil t)
316 (re-search-forward end-re nil t)))))
317 ;; FIXME: Don't know how to indicate "not found".
318 ;; Should this function throw an error? --rsteib
320 (gnus-message 3 "Requested article not found")
322 (delete-region (point-min)
323 (re-search-forward start-re))
324 (goto-char (point-min))
325 (delete-region (progn
326 (re-search-forward end-re)
329 (mm-url-decode-entities))))
331 (defun nnweb-google-parse-1 (&optional Message-ID)
332 "Parse search result in current buffer."
335 (active (cadr (assoc nnweb-group nnweb-group-alist)))
336 Subject Score Date Newsgroups From
339 (push (list nnweb-group (setq active (cons 1 0)))
341 ;; Go through all the article hits on this page.
342 (goto-char (point-min))
345 "a +href=\"/group/\\([^>\"]+\\)/browse_thread/[^>]+#\\([0-9a-f]+\\)"
347 (setq Newsgroups (match-string-no-properties 1)
348 ;; Note: Starting with Google Groups 2, `mid' is a Google-internal
349 ;; ID, not a proper Message-ID.
350 mid (match-string-no-properties 2)
352 (nnweb-definition 'result) Newsgroups mid))
353 (narrow-to-region (search-forward ">" nil t)
354 (search-forward "</a>" nil t))
355 (mm-url-remove-markup)
356 (mm-url-decode-entities)
357 (setq Subject (buffer-string))
358 (goto-char (point-max))
360 (narrow-to-region (point)
361 (search-forward "</table" nil t))
363 (mm-url-remove-markup)
364 (mm-url-decode-entities)
365 (goto-char (point-max))
368 "^\\(?:\\(\\w+\\) \\([0-9]+\\)\\|\\S-+\\)\\(?: \\([0-9]\\{4\\}\\)\\)? by ?\\(.*\\)"
370 (setq Date (if (match-string 1)
371 (format "%s %s 00:00:00 %s"
375 (substring (current-time-string) -4)))
376 (current-time-string)))
377 (setq From (match-string 4)))
380 (unless (nnweb-get-hashtb url)
384 (make-full-mail-header
385 (cdr active) (if Newsgroups
386 (concat "(" Newsgroups ") " Subject)
388 From Date (or Message-ID mid)
391 (nnweb-set-hashtb (cadar map) (car map))))
394 (defun nnweb-google-reference (id)
395 (let ((map (nnweb-google-parse-1 id)) header)
397 (nconc nnweb-articles map))
398 (when (setq header (cadar map))
399 (mm-with-unibyte-current-buffer
400 (mm-url-insert (mail-header-xref header)))
403 (defun nnweb-google-create-mapping ()
404 "Perform the search and create a number-to-url alist."
406 (set-buffer nnweb-buffer)
408 (nnheader-message 7 "Searching google...")
409 (when (funcall (nnweb-definition 'search) nnweb-search)
414 (nconc nnweb-articles (nnweb-google-parse-1)))
415 ;; Check if there are more articles to fetch
416 (goto-char (point-min))
418 (if (or (not (re-search-forward
419 "<a [^>]+href=\"\n?\\([^>\" \n\t]+\\)[^<]*<img[^>]+src=[^>]+next"
421 (>= i nnweb-max-hits))
423 ;; Yup, there are more articles
424 (setq more (concat (nnweb-definition 'base) (match-string 1)))
427 (nnheader-message 7 "Searching google...(%d)" i)
428 (mm-url-insert more))))
429 ;; Return the articles in the right order.
430 (nnheader-message 7 "Searching google...done")
432 (sort nnweb-articles 'car-less-than-car))))))
434 (defun nnweb-google-search (search)
437 (nnweb-definition 'address)
439 (mm-url-encode-www-form-urlencoded
441 ("num" . ,(number-to-string
442 (min 100 nnweb-max-hits)))
451 (defun nnweb-google-identity (url)
452 "Return an unique identifier based on URL."
453 (if (string-match "selm=\\([^ &>]+\\)" url)
460 (defun nnweb-gmane-create-mapping ()
461 "Perform the search and create a number-to-url alist."
463 (set-buffer nnweb-buffer)
464 (let ((case-fold-search t)
465 (active (or (cadr (assoc nnweb-group nnweb-group-alist))
469 (nnheader-message 7 "Searching Gmane..." )
470 (when (funcall (nnweb-definition 'search) nnweb-search)
471 (goto-char (point-min))
472 ;; Skip the status line
474 ;; Thanks to Olly Betts we now have NOV lines in our buffer!
476 (unless (or (eolp) (looking-at "\x0d"))
477 (let ((header (nnheader-parse-nov)))
478 (let ((xref (mail-header-xref header))
479 (from (mail-header-from header))
480 (subject (mail-header-subject header))
481 (rfc2047-encoding-type 'mime))
482 (when (string-match " \\([^:]+\\)[:/]\\([0-9]+\\)" xref)
483 (mail-header-set-xref
485 (format "http://article.gmane.org/%s/%s/raw"
486 (match-string 1 xref)
487 (match-string 2 xref))))
489 ;; Add host part to gmane-encrypted addresses
490 (when (string-match "@$" from)
491 (mail-header-set-from header
492 (concat from "public.gmane.org")))
494 (mail-header-set-subject header
495 (rfc2047-encode-string subject))
497 (unless (nnweb-get-hashtb (mail-header-xref header))
498 (mail-header-set-number header (incf (cdr active)))
499 (push (list (mail-header-number header) header) map)
500 (nnweb-set-hashtb (cadar map) (car map))))))
502 (nnheader-message 7 "Searching Gmane...done")
504 (sort (nconc nnweb-articles map) 'car-less-than-car)))))
506 (defun nnweb-gmane-wash-article ()
507 (let ((case-fold-search t))
508 (goto-char (point-min))
509 (when (search-forward "<!--X-Head-of-Message-->" nil t)
510 (delete-region (point-min) (point))
511 (goto-char (point-min))
512 (while (looking-at "^<li><em>\\([^ ]+\\)</em>.*</li>")
513 (replace-match "\\1\\2" t)
515 (mm-url-remove-markup))))
517 (defun nnweb-gmane-search (search)
520 (nnweb-definition 'address)
522 (mm-url-encode-www-form-urlencoded
523 `(("query" . ,search)
524 ("HITSPERPAGE" . ,(number-to-string nnweb-max-hits))
525 ;;("TOPDOC" . "1000")
527 (setq buffer-file-name nil)
528 (set-buffer-multibyte t)
529 (mm-decode-coding-region (point-min) (point-max) 'utf-8)
532 (defun nnweb-gmane-identity (url)
533 "Return a unique identifier based on URL."
534 (if (string-match "group=\\(.+\\)" url)
539 ;;; General web/w3 interface utility functions
542 (defun nnweb-insert-html (parse)
543 "Insert HTML based on a w3 parse tree."
545 ;; We used to call nnheader-string-as-multibyte here, but it cannot
546 ;; be right, so I removed it. If a bug shows up because of this change,
547 ;; please do not blindly revert the change, but help me find the real
548 ;; cause of the bug instead. --Stef
550 (insert "<" (symbol-name (car parse)) " ")
553 (concat (symbol-name (car param)) "="
555 (if (consp (cdr param))
561 (mapc 'nnweb-insert-html (nth 2 parse))
562 (insert "</" (symbol-name (car parse)) ">\n")))
564 (defun nnweb-parse-find (type parse &optional maxdepth)
565 "Find the element of TYPE in PARSE."
567 (nnweb-parse-find-1 type parse maxdepth)))
569 (defun nnweb-parse-find-1 (type contents maxdepth)
570 (when (or (null maxdepth)
571 (not (zerop maxdepth)))
572 (when (consp contents)
573 (when (eq (car contents) type)
574 (throw 'found contents))
575 (when (listp (cdr contents))
576 (dolist (element contents)
577 (when (consp element)
578 (nnweb-parse-find-1 type element
579 (and maxdepth (1- maxdepth)))))))))
581 (defun nnweb-parse-find-all (type parse)
582 "Find all elements of TYPE in PARSE."
584 (nnweb-parse-find-all-1 type parse)))
586 (defun nnweb-parse-find-all-1 (type contents)
588 (when (consp contents)
589 (if (eq (car contents) type)
590 (push contents result)
591 (when (listp (cdr contents))
592 (dolist (element contents)
593 (when (consp element)
595 (nconc result (nnweb-parse-find-all-1 type element))))))))
599 (defun nnweb-text (parse)
600 "Return a list of text contents in PARSE."
601 (let ((nnweb-text nil))
603 (nreverse nnweb-text)))
605 (defun nnweb-text-1 (contents)
606 (dolist (element contents)
607 (if (stringp element)
608 (push element nnweb-text)
609 (when (and (consp element)
610 (listp (cdr element)))
611 (nnweb-text-1 element)))))
615 ;; arch-tag: f59307eb-c90f-479f-b7d2-dbd8bf51b697
616 ;;; nnweb.el ends here